PDA

View Full Version : GH1 firmware research volunteers required



Pages : 1 2 3 4 5 6 7 [8] 9 10 11 12 13

Simon Höfer
05-28-2010, 12:23 PM
Car3o, the money I sent via paypal just came back today. The message said you didn't log into paypal or registered completely. Now that I look at it again. The email ends with "...@gmail.comm". Was it not set properly?

tyampel
05-28-2010, 12:30 PM
This was corrected a while ago.
Please resend it again to the right address.
This first page link works properly now.

Vitaliy Kiselev
05-28-2010, 12:46 PM
Car3o, the money I sent via paypal just came back today. The message said you didn't log into paypal or registered completely. Now that I look at it again. The email ends with "...@gmail.comm". Was it not set properly?

Yes, it was long time ago for few days. Fixed. Paypal is just slow, as usual :-)

Vitaliy Kiselev
05-28-2010, 12:49 PM
Just made charger for GH1 batteries from chinese charger for DU07/DU21 batteries (seller send me it by mistake long time ago :-) ).
Hope it'll work ok, I placed third-party battery and it looks like it is charging (as electrically contacts are the same and both kinds are 7.2v batteries).
Anyone used chinese chargers on original batteries?

sunsurfh2o
05-28-2010, 12:50 PM
i don't think this is accurate, as i had just stills in my thumbnails w/my footage and it froze on me twice. maybe he has a different arrangement of clips. it's 100% going to freeze your camera the second you go through thumbnails and your 1080p 24p footage is there, it will freeze.

To clarify, you only need to make sure that the last item recorded IS NOT 1080p (with 24p patch) if you want to scroll thumbnails. Example: take a 720p or still image immediately following 1080p footage. Once you've done that you can press play button and zoom out to thumbnail view. From here you can scroll through thumbnails BUT the instant you click on a 1080p file to view zoomed in, you freeze your camera.

Vitaliy Kiselev
05-28-2010, 12:55 PM
To clarify, you only need to make sure that the last item recorded IS NOT 1080p (with 24p patch) if you want to scroll thumbnails. Example: take a 720p or still image immediately following 1080p footage. Once you've done that you can press play button and zoom out to thumbnail view. From here you can scroll through thumbnails BUT the instant you click on a 1080p file to view zoomed in, you freeze your camera.

Thanks for detailed description.

rundavids
05-28-2010, 02:20 PM
Hi Tester. Sorry to ask, but do you think it's fixable hack to playback 1080p footage in camera?

Vitaliy Kiselev
05-28-2010, 02:24 PM
Hi Tester. Sorry to ask, but do you think it's fixable hack to playback 1080p footage in camera?

I can't say at this time.

Vitaliy Kiselev
05-29-2010, 12:52 AM
Good news is that self-made chinese charger is working.
I'll be using Transcent Ultimate 8Gb Class 10 and other Class 6 cards.
Did some tests, and I am not able to archive declared 16Mb/s speed on my readers.
Real is about 9-10Mb/s write (same as cheap A-data card :-) )
Anyway, it is enough for our task.

Ben_B
05-29-2010, 12:58 AM
Did you mean MB or Mb? Because 9-10Mb/s doesn't seem like enough. 10MB on the other hand is plenty.

Vitaliy Kiselev
05-29-2010, 01:10 AM
Did you mean MB or Mb? Because 9-10Mb/s doesn't seem like enough. 10MB on the other hand is plenty.

10Megabytes/second :-)
10Mbit/s SDHC card is hard to find.

Ben_B
05-29-2010, 01:48 AM
Big MB. Just checking :D

g.l
05-29-2010, 05:01 AM
The reader might be the bottleneck, many can't break 8-10MB/s.

IanB
05-29-2010, 05:58 AM
Donated my 25$...This is gods work....

panystac
05-29-2010, 07:44 AM
tester13, I was wondering if there's any way to look at the G2 firmware to see if it's encrypted??

I guess you can't look at it until they release a firmware update??

The G2 on Amazon is around $799 USD with the single lens zoom kit, but in Japan, you can find it for around the equivalent of $580 USD!!!!!!!!!!!!!!!!!


Any response to this question from a couple of days back?

Thanks!!! :)

Oh, and the G2 with kit lens is now down to the equivalent of around $530 USD

rotrier
05-29-2010, 09:15 AM
Another $50 donation for incredible work! Keep it up guys!

TurboTwo
05-29-2010, 09:24 AM
Hi Car3o ,How`s the fund going for what Tester needs next ?

sunsurfh2o
05-29-2010, 10:06 AM
I've PM'd Tester already but in the event he is on a holiday is there anyone else here that have run into trouble with upgrading FW and found it was UNRELATED to usual missteps like 1) file naming missing double underscore 2) version set/incrementing while using pTool 3) following Panny's install instructions (reformat, 2 pics, fresh battery or AC???

I've done this a dozen times and had no problems but suddenly I can't continue with my testing, or rolling back to original FW. Only thing that seems different this time is in the last FW upgrade I broke through the 1.3 to 1.4 in the version nomenclature.

Any IMMEDIATE help would be appreciated as I have some work to do with camera tomorrow and need to roll back to original FW for it.

Thanks!

Vitaliy Kiselev
05-29-2010, 11:30 AM
Any response to this question from a couple of days back?

Oh, and the G2 with kit lens is now down to the equivalent of around $530 USD

We'll look at G2/G10 encryption later.
GH1 and porting patches to G1 and GF1 are priority.

Vitaliy Kiselev
05-29-2010, 11:31 AM
Any IMMEDIATE help would be appreciated as I have some work to do with camera tomorrow and need to roll back to original FW for it.


See PM.

Vitaliy Kiselev
05-29-2010, 11:37 AM
The reader might be the bottleneck, many can't break 8-10MB/s.

Yes, reader can be bottleneck.
I'll try to get Sandisk Extreme III to check this.
Unfortunately I don't see any good readers in the vicinity.

Vitaliy Kiselev
05-29-2010, 04:25 PM
PTool 3.30 released.

MJPEG patches added

Note
We are slowly moving to changing MJPEG bitrate.
It is easy to change it to lower values, but not so to increase it.

Notes to testers are on the first page - second post.

Please note that I need testers to make it work, as it is always better to have many eyes and test more values to understand how it all works.

andyjar
05-30-2010, 02:53 PM
TESTER13

Hi. I am shooting explosion stock with your PAL> NTSC conversion. The recording is stopping by itself, at a random point in, either just as the explosion happens or a second or two in. Same in 1080 or 720. It's not overheating, as I left the cam inside until i needed it and it was only outside for a few minutes. MTS files available.

NURBS
05-30-2010, 03:05 PM
Hello Guys'

I was little busy. But tonight I testing ptool again. ;-)

MJPEG tests - Part 1.

MJPEG Quality: 1024
MJPEG Table: 1

Video

Bit rate : 28.2 Mbps
Bits/(Pixel*Frame) : 1.020

MJPEG Quality: 1024
MJPEG Table: 4

Video

Bit rate : 28.3 Mbps
Bits/(Pixel*Frame) : 1.023

MJPEG Quality: 1024
MJPEG Table: 8

Video

Bit rate : 27.9 Mbps
Bits/(Pixel*Frame) : 1.010

MJPEG Quality: 1024
MJPEG Table: 16

Video

Bit rate : 27.4 Mbps
Bits/(Pixel*Frame) : 0.990

MJPEG Quality: 1024
MJPEG Table: 32

Video

Bit rate : 25.4 Mbps
Bits/(Pixel*Frame) : 0.919

MJPEG Quality: 1024
MJPEG Table: 64

Video

Bit rate : 23.9 Mbps
Bits/(Pixel*Frame) : 0.865

MJPEG Quality: 1024
MJPEG Table: 128

Video

Bit rate : 23.3 Mbps
Bits/(Pixel*Frame) : 0.842

MJPEG Quality: 1024
MJPEG Table: 1024

Video

Bit rate : 21.3 Mbps
Bits/(Pixel*Frame) : 0.771

All tests working well.

MJPEG Quality: 2048
MJPEG Table: 1024

Can't working well, camera show message about card speed and freezing.

MJPEG Quality: can't more ~1600 (1700 don't working)

More tests will be soon.

Regards
NURBS

Vitaliy Kiselev
05-30-2010, 03:09 PM
Did some tests on GH1.

Changing MJPEG quality from 0x80 to 0x95 results in bitrate increase from 30.5Mbit/sec to 35.7Mbit/sec on same scene (exactly about 17% increase).

Also few AVCHD test had been done.

Checked sensitivity of GH1 using low lit scene.
GH1 with F3.5 (used this lens to get similar FOV), ISO1600, 1/50 have much darker picture compared to DX1 F1.8/15Db gain,1/60. Exposure check gives almost 2.5 stops underexposure. DX1 will be noisier, of course, but this is CCD noise, and not GH1 patterns.
1/60 sec shutter on GH1 results in flickering and bands under even incandescent 50Hz lights.

Vitaliy Kiselev
05-30-2010, 03:14 PM
TESTER13
Hi. I am shooting explosion stock with your PAL> NTSC conversion. The recording is stopping by itself, at a random point in, either just as the explosion happens or a second or two in.

You need to test it without PAL->NTSC patch. And get real NTSC camera for test.
One of the reasons can be audio level error or similar things. Your explosion can cause some hardware error.

Vitaliy Kiselev
05-30-2010, 03:18 PM
Hello Guys'

I was little busy. But tonight I testing ptool again. ;-)

MJPEG tests - Part 1.

MJPEG Quality: 1024
MJPEG Table: 1

Video

Bit rate : 28.2 Mbps
Bits/(Pixel*Frame) : 1.020



Sorry, but I do not see any point in such tests.
Look at the used values.
Why 1024 for Quality???
You must only slightly increase it over 0x80, look at my results.
Do not touch MJPEG Table at first and adjust only if you have problems with quality setting (use 1 to 1024 for this value, change by large value, like to 1, 250, 500, 600, 700, 860, 1000).
Also. card must be real good, as we are at >35Mbits, so good Class 10 or speedy Class 6 card required to move forward. You can get cheap Transcent Class 10 cards or test few Class 6 cards and find fastest.

spamrakuen
05-30-2010, 03:32 PM
Changing MJPEG quality from 0x80 to 0x95 results in bitrate increase from 30.5Mbit/sec to 35.7Mbit/sec on same scene (exactly about 17% increase).

Wow! Amazing job! :)

Vitaliy Kiselev
05-30-2010, 03:39 PM
Unfortunately GH1 update is quite slow process.
So, it'll be best to write small firmware portion to change values using file on SD card.
Unfortunately we don't have assembler available.
So, our next target to obtain must be Panasonic C compiler and assembler.

jhero
05-30-2010, 04:02 PM
Quick question,

Has anyone noticed a large difference going from stock audio of 192/32 to 384-448/48 ?

Vitaliy Kiselev
05-30-2010, 04:12 PM
Quick question,

Has anyone noticed a large difference going from stock audio of 192/32 to 384-448/48 ?

Please turn back few pages, and you'll find varous answers to your question.
And stock audio is 192/48.

cowpunk52
05-30-2010, 04:43 PM
I apologize if this has been addressed, but I couldn't find any mention of it in my search. I've updated my firmware to the native 24p hack, and when I try to play back in camera, it doesn't freeze. Actually, I just get a message saying "this motion image cannot be played." I then tap the shutter, then go back to shooting like normal!

Just an observation on my part, since I noticed most people saying that the camera would freeze and require a battery pull.

Vitaliy Kiselev
05-30-2010, 05:26 PM
PTool 3.30 updated without version change.

Main change is AVCHD Compression patch, using test results this patch is targeted towards allowing better behaviour while increasing bitrate using other patches.
Note - this patch is untested, as I don't have time to test it carefully (I added it to public version to save us some time), but it is based on test results of other internal patches.

Note to testers updated on first page.

Sean Jeong
05-30-2010, 06:15 PM
Tester just wondering, are you working on hacking the auto gain feature of the GH1? Just wanna know if i have to invest in a decent recorder.

andyjar
05-30-2010, 06:26 PM
You need to test it without PAL->NTSC patch. And get real NTSC camera for test.
One of the reasons can be audio level error or similar things. Your explosion can cause some hardware error.

It's not a noise issue as there is almost no sound (fast conflagrations, not explosions.)

I can't simply get a real NTSC camera...

But what I was shooting was often very complicated,

http://imgur.com/1T9hd.jpg

...and I did have some AVCHD bitrate changes besides the PAL>NTSC hack and I thought I might have broken the codec with the complexity. I'll try without your hacked firmware later.

Vitaliy Kiselev
05-30-2010, 06:50 PM
It's not a noise issue as there is almost no sound (fast conflagrations, not explosions.)
I can't simply get a real NTSC camera...
...and I did have some AVCHD bitrate changes besides the PAL>NTSC hack and I thought I might have broken the codec with the complexity. I'll try without your hacked firmware later.

Remove all other stuff, use only PAL to NTSC.
May be you are having some interference or SD card have issues.
Your problems are not related to PAL->NTSC hack.

Ozpeter
05-30-2010, 10:07 PM
Tester just wondering, are you working on hacking the auto gain feature of the GH1? Just wanna know if i have to invest in a decent recorder.

http://www.dvxuser.com/V6/search.php?searchid=5640570 should bring up many of the previous discussions in this thread on the subject (or search the thread for "limiter").

At risk if going over the old ground again, if you disable the GH1 limiter, then instead of limited audio you'll get distorted audio. To avoid that, you need some kind of level control in the displays, with associated metering, etc etc, and the chances of even the highly gifted Tester13 coming up with that anytime soon are somewhat - er - limited, IMHO.

Unless you are just doing family stuff, you are likely to need a decent recorder anyway.

Vitaliy Kiselev
05-30-2010, 11:53 PM
At risk if going over the old ground again, if you disable the GH1 limiter, then instead of limited audio you'll get distorted audio. To avoid that, you need some kind of level control in the displays, with associated metering, etc etc, and the chances of even the highly gifted Tester13 coming up with that anytime soon are somewhat - er - limited, IMHO.

I made small research and we have some parts already implemented - we can use playback audio level as control to set our level and we also have necessary small icons that can be taken from audio dubbing screen.

Vitaliy Kiselev
05-31-2010, 01:18 AM
New testing results.

Bad news.
I tried AVCHD compression patch with VB-33.500,MaxB-24.000,AltB-22.000.
It freezes camera.
With all parameters at defauls, no freezing with this patch.
So, this needs more work.

Good news.
Shot some MJPEG test footage using quality = 256.
Amazingly looking 57,5MBit result (bitrate is 2x from original).
Such bitrate is archived on high detail scenes.
No problems so far, but high speed card is required!

rawfa
05-31-2010, 02:03 AM
New testing results.

Bad news.
I tried AVCHD compression patch with VB-33.500,MaxB-24.000,AltB-22.000.
It freezes camera.
With all parameters at defauls, no freezing with this patch.
So, this needs more work.

Good news.
Shot some MJPEG test footage using quality = 256.
Amazingly looking 57,5MBit result (bitrate is 2x from original).
Such bitrate is archived on high detail scenes.
No problems so far, but high speed card is required!

Jesus Christ! 57,5mbit!!! That is incredible! Is this in full hd? Can we see some samples?

Vitaliy Kiselev
05-31-2010, 02:25 AM
Jesus Christ! 57,5mbit!!! That is incredible! Is this in full hd? Can we see some samples?

Please, try to read before posting!
This is MJPEG 1280x720p30.

As for example footage - I don't have time for this, sorry.
Ask testers.

Plus we are on too high level, as quality setting behaviour is quite weird.
For example 256 works, but 384 (another 50% increase) show that maximum clip length is decreased accordingly, but actual bitrate is back to about 27Mbit.

spamrakuen
05-31-2010, 02:38 AM
Good news.
Shot some MJPEG test footage using quality = 256.
Amazingly looking 57,5MBit result (bitrate is 2x from original).
Such bitrate is archived on high detail scenes.
No problems so far, but high speed card is required!

OMG! That's... that's... I've no words! :shocked:

Vitaliy Kiselev
05-31-2010, 02:42 AM
Bitrate peaks are at about 62-63Mbits/s (grass, leaves, etc), I also don't see frame skipping.
But anyway MJPEG requires much more work to understand encoder in more detail.

rawfa
05-31-2010, 02:59 AM
Please, try to read before posting!
This is MJPEG 1280x720p30.

As for example footage - I don't have time for this, sorry.
Ask testers.

Plus we are on too high level, as quality setting behaviour is quite weird.
For example 256 works, but 384 (another 50% increase) show that maximum clip length is decreased accordingly, but actual bitrate is back to about 27Mbit.į

When I asked for example footage I was not asking you specifically...just anyone.
1280*720p at this bitrate should be fantastic...if it could be 24p o 25p instead of 30 it would be even more. Great work, guys!!

spamrakuen
05-31-2010, 03:20 AM
Bitrate peaks are at about 62-63Mbits/s (grass, leaves, etc), I also don't see frame skipping.
But anyway MJPEG requires much more work to understand encoder in more detail.

Are you able to playback the video on camera or it freezes?

Vitaliy Kiselev
05-31-2010, 03:33 AM
Are you able to playback the video on camera or it freezes?

I did not made many tests.
For few one I got "Read error. Check the card."
I believe camera make check for bitrate.
As low detail scene with same settings plays fine.
On computer all files play fine.

TrueIndigo
05-31-2010, 04:38 AM
If these elevated MJPEG bitrates can eventually be used at 24-25 fps (without frame skipping) it would be so satisfying! And if it were also possible to change MJPEG frame to size to 1080 it would make it even more usable. Very encouraging, thanks.

Vitaliy Kiselev
05-31-2010, 04:48 AM
If these elevated MJPEG bitrates can eventually be used at 24-25 fps (without frame skipping) it would be so satisfying! And if it were also possible to change MJPEG frame to size to 1080 it would make it even more usable. Very encouraging, thanks.

Yeah, fantasy is good.

Vitaliy Kiselev
05-31-2010, 06:44 AM
Small tests update.

Generally 1080p MJPEG is possible.
1920p30 stops after few seconds.
1920p24 runs as long as necessary (same bitrate as 1280).
File is pretty ok and playable.
But picture is awful green-magenta nightmare with artefacts.
So, MJPEG encoder works ok.
But all sensor setup and may be buffers are problematic.

Rabot
05-31-2010, 07:23 AM
Hello Tester13, any idea if a windowed mode would be possible?

Vitaliy Kiselev
05-31-2010, 07:50 AM
Hello Tester13, any idea if a windowed mode would be possible?

I don't know at this time.
Most probably answer is no.
I need sensor manual and firmware source to tell exactly.

NURBS
05-31-2010, 08:49 AM
My tests show that result:

MJPEG tests - Part 2. (static picture)

MJPEG Quality: 256
MJPEG Table: 266

Duration : 10s 0ms
Bit rate : 41.9 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.517

If don't turn on MJPEG table, bitrate been less.

Regards
NuRBS

Vitaliy Kiselev
05-31-2010, 09:09 AM
My tests show that result:

MJPEG tests - Part 2. (static picture)

MJPEG Quality: 256
MJPEG Table: 266

Duration : 10s 0ms
Bit rate : 41.9 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.517

If don't turn on MJPEG table, bitrate been less.



Thanks.
I believe that table index (note - this is index in table of values!) drives estimator behaviour.
And quality sets maximum bitrate.
Shoot some detailed scene, I got jump from 30Mbps static simple scene to high detailed trees scene at 57Mbps with some fragments at 63Mbps.

videoguy69
05-31-2010, 09:28 AM
Tester13,

Is it possible for you to disable certain features in the camera like internal sound to reserve processing power for other features? Like improving the codec?

Vitaliy Kiselev
05-31-2010, 09:35 AM
Tester13,
Is it possible for you to disable certain features in the camera like internal sound to reserve processing power for other features? Like improving the codec?

:-)
Almost all features are implemented in hardware.
CPU in LSI only sets certain parameters of other hardware parts. :-)
So, it won't help.

sirk
05-31-2010, 10:02 AM
I made small research and we have some parts already implemented - we can use playback audio level as control to set our level and we also have necessary small icons that can be taken from audio dubbing screen.

What are playback audio levels? What do you intend to do with the icons? Could you please be more precise on this?

Thanks
btw: i begin to love your mjpeg

Vitaliy Kiselev
05-31-2010, 10:10 AM
What are playback audio levels?

SETUP->VOLUME.


What do you intend to do with the icons?

Look at PLAYBACK MENU -> AUDIO DUB.
See small progress indicator under microphone image.
It is perfect for audio level display (black and yellow colors are also good).

NURBS
05-31-2010, 10:26 AM
Very interesting thing.

Setting 'MJPEG Table:' have direct influence for MJPEG thumbnail picture.

MJPEG Table: 1 -> 500kb
MJPEG Table: 128 -> 146kb
MJPEG Table: 1024 -> 60kb

;)

Regards
NURBS

Vitaliy Kiselev
05-31-2010, 10:29 AM
Very interesting thing.

Setting 'MJPEG Table:' have direct influence for MJPEG thumbnail picture.

MJPEG Table: 1 -> 500kb
MJPEG Table: 128 -> 146kb
MJPEG Table: 1024 -> 60kb


Interesting.
Quality also affects thumbnail size btw.
Top table have only 1/5 compression ratio.

NURBS
05-31-2010, 10:34 AM
DVX forum script compress 500kb image to 80kb :-) but 146kb and 64kb is original.

Vitaliy Kiselev
05-31-2010, 10:38 AM
DVX forum script compress 500kb image to 80kb :-) but 146kb and 64kb is original.

Do not use attachments as they are invisible to non-members.
Use member upload center (red icon to the right) and insert provided link in message text.
It also won't touch your images.

sirk
05-31-2010, 11:07 AM
crazy difference in image quality! are those uploads screengrabs from the video or thumbnails of it?

Vitaliy Kiselev
05-31-2010, 12:49 PM
PTool 3.31 Released
MJPEG compression patches added.
MJPEG 1920 patch added, just for fun (very weird picture!).

This release allow you to precisely set MJPEG bitrate, for high detailed scenes, for average and for low detailed. Maximum normal working bitrate is at least 63Mbit/sec.
But you need Sandisk or good Class 10 card.

Note to testers is on first page - second post.
Table values can be also found at the end for second post.

P.S. We are now most popular thread in DSLR section and second or third on whole forum.

e-steve
05-31-2010, 12:57 PM
I hope you (Tester13) have some Google Analytics on your own website to see how often Panny is actually checking on your progress :)

Keep up the good work!!

It's hard for me to tell how promising the MJPEG is looking -- can someone who understands what we might be realistically hoping for shed some perspective?

In other words, is tester13 close to some major breakthrough with the MJPEG progress?

This is the only thread I read every day (5x-10x)...

Nice work guys!

Vitaliy Kiselev
05-31-2010, 01:22 PM
It's hard for me to tell how promising the MJPEG is looking -- can someone who understands what we might be realistically hoping for shed some perspective?


MJPEG patches are already pretty usable.
For example, checking only E1 Quality and set it to 256 give you about 57Mbps bitrate at high detailed scenes (and I don't see any JPEG artefacts).

Here is my last test quality settings, working ok (average low detail bitrate was about 35Mbit):
Quality settings (E1 to E4) - 0x160, 0xDC, 0xC8, 0xB8
Table settings (E1 to E4) - 70, 40, 133, 120,

But we need many testers to understand how estimator works.

NURBS
05-31-2010, 01:52 PM
For same 'tester13' sttings I have: (static scene in all video same).

Bit rate : 43.1 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.558

Next test:

Quality settings (E1 to E4) - 256, 220, 200, 184
Table settings (E1 to E4) - 233, 292, 300, 310

Bit rate : 13.0 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 0.471

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 70, 70, 80, 70

Bit rate : 29.4 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.063

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 70, 32, 32, 32

Bit rate : 45.5 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.646

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 32, 24, 24, 24

Bit rate : 54.0 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.955

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 12, 12, 12

Bit rate : 44.6 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.613

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

Bit rate : 54.5 Mbps - Wow!
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.971

I can't testing 1920 mode setting, because don't have fast card.

Vitaliy Kiselev
05-31-2010, 01:59 PM
I can't testing 1920 mode setting, because don't have fast card.

Go and by good fast card. Transcent Class 10 is quite cheap and looks good.

1920 mode do not require fast card (with defailt MJPEG settings, you can further decrease quality, btw), but must be used with 30-24fps patch.
Image is really weird magenta-green and distored.
Whole patch purpose was to prove that MJPEG encoder is capable to produce 1080p24.

And look at actual table values (see my second post at first page).

NURBS
05-31-2010, 02:17 PM
Okey, thank you for note about digits table. ;-)

Vitaliy Kiselev
05-31-2010, 02:26 PM
Okey, thank you for note about digits table. ;-)

Testing must be the following:
1) Shooting very low details image (white paper, etc).
2) Shooting medium detail, out of focus image.
3) Focused normal interior image.
4) Glass, trees, very high detail images with good lightling

For each of four scenarios you must make separate clips. And one clip with all this together.
So it'll be possible to understand how it works.

NURBS
05-31-2010, 02:49 PM
Very interesting deal.

SD video DV compression is 25Mbit/s - Mjpeg
HD 720p Mjpeg 50Mbit/s it's very near same compression level.

My Sandisk can write 9Mb/s (70 Mbps) :-) I need new card..

;-) very nice result!

Isaac_Brody
05-31-2010, 04:09 PM
Wait, if I'm reading correctly 1080P mjpeg works? Is this with pulldown or native 24P?

sirk
05-31-2010, 04:30 PM
I don't get this:
the quality settings are "for 1080p30 only" but you provide an extra 1920 patch?
did you change the standard mjpeg 720p setting to 1080p? why the extra 1920 patch?
thanks

sirk
05-31-2010, 04:47 PM
Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

Low complexity (white sheet)
QF (like Gordian Knot)............................0.831
Video Stream Length...............................3s 500ms
Video Stream BitRate..............................23.0 Mbps
Video Stream BitRate Mode.........................VBR
Video Stream Size.................................9.58 MiB (98%)

Medium Quality (out of focus)
QF (like Gordian Knot)............................1.828
Video Stream Length...............................2s 500ms
Video Stream BitRate..............................50.5 Mbps
Video Stream BitRate Mode.........................VBR
Video Stream Size.................................15.1 MiB (99%)

Complex (Leaves and movement)
Display Aspect Ratio..............................16:9
QF (like Gordian Knot)............................1.989
Video Stream Length...............................4s 500ms
Video Stream BitRate..............................55.0 Mbps
Video Stream BitRate Mode.........................VBR
Video Stream Size.................................29.5 MiB (99%)

Alltogether in one shot
QF (like Gordian Knot)............................1.637
Video Stream Length...............................15s 500ms
Video Stream BitRate..............................45.3 Mbps
Video Stream BitRate Mode.........................VBR
Video Stream Size.................................83.6 MiB (99%)

Dont know what QF means (first use of this tool called MediaInfo Mac). Hope it helps.
PS: its 720p. Why does it say "only for 1020p" in the description in ptool?

John Caballero
05-31-2010, 06:23 PM
The mjpeg developments are great news. If 24 FPS could be achieved with 720p and maybe even 1080p with the better bit rate, etc. it would be fabulous. I like the stability, quality and easier post production of mjpeg. tester13 YOU ARE GOOD AT YOUR GAME!

Barry_Green
05-31-2010, 06:26 PM
If "real" 24fps mjpg were possible in 720p mode, with enhanced bitrate that would be delicious. Absolutely mud-free, great-looking hd at 24p...

John Caballero
05-31-2010, 06:29 PM
If "real" 24fps mjpg were possible in 720p mode, with enhanced bitrate that would be delicious. Absolutely mud-free, great-looking hd at 24p...


I am praying for that. The GH 1 with that ability would be awesome.

e-steve
05-31-2010, 06:31 PM
So it doesn't seem that it is clear yet, whether or not the 24p MJPEG is still skipping frames or not...right?

John Caballero
05-31-2010, 06:45 PM
So it doesn't seem that it is clear yet, whether or not the 24p MJPEG is still skipping frames or not...right?


Latest report it was but tester13 was concentrating more on bitrate and and stuff at this moment. Barry, if the Nikon D90 does 720p24 you think it could be possible to achieve it flawlessly on the GH1 with a hack? If anybody can do it tester13 would be the one!

JoeJITSU
05-31-2010, 07:34 PM
Ok noob question here. Might be the dumbest questions but....What advantages will I have if I use the latest firmware? what disadvantages? Is there a how to on the pages here? I bought my Gh1 last feb and did not do any updates to it...should I have done so? I still don't have a pulldown program.....Should I buy a pulldown program or should I wait till a video editing program puts the pulldown into their product? If I do have to buy one....what program would you guys recommend? I know these are allot of questions but I know there is some of you out there that can answer all of them. Thanks

Isaac_Brody
05-31-2010, 08:58 PM
Ok noob question here. Might be the dumbest questions but....What advantages will I have if I use the latest firmware? what disadvantages? Is there a how to on the pages here? I bought my Gh1 last feb and did not do any updates to it...should I have done so? I still don't have a pulldown program.....Should I buy a pulldown program or should I wait till a video editing program puts the pulldown into their product? If I do have to buy one....what program would you guys recommend? I know these are allot of questions but I know there is some of you out there that can answer all of them. Thanks

Best to start a new thread for these questions.

Barry_Green
05-31-2010, 10:30 PM
Barry, if the Nikon D90 does 720p24 you think it could be possible to achieve it flawlessly on the GH1 with a hack?
Well, let's separate a few things out -- just because the Nikon can do it, bears no resemblance to what capabilities may or may not be in the GH1. Different hardware from a different manufacturer, so I don't think any conclusions can be drawn about what the GH1 could theoretically be capable of, based on some other manufacturer's product.

However -- if I were to guess (and this is only a guess) I would think it's certainly possible. We already know that the hardware is capable of scanning at 24fps, because it does so in 1080 mode. And we know that he's been able to get the codec to store only 24fps already. To me, it seems like he's really maybe only one small step away, which is to trigger the existing 24fps codec situation, and simultaneously put the camera in 24fps scanning mode. But, of course, it could be more complex than that.

And, the footage should look a lot better than the D90, because IINM the D90 operates at about 14 megabits, and tester13's already got the GH1 operating at 4x that bandwidth.

John Caballero
05-31-2010, 10:39 PM
Thanks Barry. The GH1 is a much better video camera. Let's hope it can be done. Mjpeg is a very capable codec.

dracore
05-31-2010, 11:10 PM
hey tester,

been busy lately. but I managed to make an update to wingraph32 for you (hurray!). As requested from before, made changes for insensitive text search from either starting symbol or as a substring. Also added address lookup in IDA.

Important changes to mouse-click operations:
- Right-click will now select your root node
- Double left-click will attempt to lookup the node in IDA (if you have it running in background).


Tried assuming "blue" colour to be root node but it doesn't work. Reason being that (in primer example), depth search begins at root and travels AWAY from node (note the arrow directions along edges). The blue root in primer, arrows all point TO the root... so even if I made that root automatically, setting depth of 1, 2, 3, etc... will still only show that one node and nobody else. I guess the only way around this is to do a reverse depth search.


Here is the link for updated build:

http://rapidshare.com/files/393903229/wingraph32.rar.html

John Caballero
05-31-2010, 11:28 PM
Tried assuming "blue" colour to be root node but it doesn't work. Reason being that (in primer example), depth search begins at root and travels AWAY from node (note the arrow directions along edges). The blue root in primer, arrows all point TO the root... so even if I made that root automatically, setting depth of 1, 2, 3, etc... will still only show that one node and nobody else. I guess the only way around this is to do a reverse depth search.


Hey! This is an ENGLISH SPOKEN site! No foreign languages please! LOL! Keep up the great work guys.

dracore
06-01-2010, 12:01 AM
Hey! This is an ENGLISH SPOKEN site! No foreign languages please! LOL! Keep up the great work guys.

LOL wait until you hear us speaking binary :grin:

John Caballero
06-01-2010, 12:13 AM
LOL wait until you hear us speaking binary

Oh no! Binary? No! My head will explode! Can you get me some mjpeg 720p24 please?

NURBS
06-01-2010, 12:16 AM
Little test: (very high detail images with good lightling).

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

Bit rate : 56.6 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 2.047

:bath:

Isaac_Brody
06-01-2010, 12:18 AM
That looks great! Bye bye mud.

Kholi
06-01-2010, 12:22 AM
That's a lot of detail. Can we get some moving footage of that?

NURBS
06-01-2010, 12:31 AM
Okey, video link for this test. (same scene, new shot Bitrate: 51 Mbps)

Link: http://rapidshare.com/files/393926937/P1010043.MOV

Regards
NURBS

NURBS
06-01-2010, 12:44 AM
Next test: (Shooting medium detail, out of focus image).

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

Bit rate : 43.7 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.580

:bath:

NURBS
06-01-2010, 12:46 AM
Same settings:

Bit rate : 44.8 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.621

spamrakuen
06-01-2010, 12:50 AM
Okey, video link for this test. (same scene, new shot Bitrate: 51 Mbps)

Link: http://rapidshare.com/files/393926937/P1010043.MOV

Regards
NURBS

Thank you very much NURBS! Nice sample ;)

A low light, simple scene sample would also be nice, since MJPEG compression is quite agressive in these scenario.

NURBS
06-01-2010, 12:53 AM
Next test: (Shooting very low details image (white paper)).

Bit rate : 33.9 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 1.226

NURBS
06-01-2010, 01:02 AM
Low light scene was in old post's ;)

Video link here: http://rapidshare.com/files/393935130/P1010040.MOV

:bath:

spamrakuen
06-01-2010, 01:13 AM
Low light scene was in old post's ;)

Video link here: http://rapidshare.com/files/393935130/P1010040.MOV


Well, I meant extreme low light. Something like a room in complete darkness lighted by a candle :engel017:

NURBS
06-01-2010, 01:22 AM
Okey ;) in the night time I'll di it.

Regards
NURBS

spamrakuen
06-01-2010, 01:23 AM
Okey ;) in the night time I'll di it.


Thanks! Thanks! Thanks! :bath:

jobless
06-01-2010, 02:09 AM
Quality settings (E1 to E4) - 512, 352, 220, 200
Table settings (E1 to E4) - 24, 24, 24, 24

SCENE 1:

Overall bit rate : 79.5 Mbps
Bit rate : 79.0 Mbps
Bits/(Pixel*Frame) : 2.857
Frame rate : 30.000 fps

download link:
http://www.nikolicnemanja.com/scene1.rar

SCENE 2:

Overall bit rate : 79.9 Mbps
Bit rate : 79.4 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 2.871


download link:
http://www.nikolicnemanja.com/scene2.rar

In rar files I included Both 720 and 1080 AVCHD (23,976 FPS) of the same scene for comparison
Mjpeg outperforming 720 Avchd in terms of sharpness and it's close to 1080 when upscaled...
I get aprx 2 Min of recording with this settings...

Tameside
06-01-2010, 02:36 AM
That is very high bit rate, the camera stops at the sustained write time or the memory card glitches? That's huge huge bit rate, max of class10 card. So hardware is capable of a lot?

jobless
06-01-2010, 02:43 AM
No, I have time indicator of 2 min...
I didn't try to record all 2 min....

Now I tried with even higher settings:
Quality settings (E1 to E4) - 754, 512, 352, 220
Table settings (E1 to E4) - 24, 24, 24, 24

Camera records; Time indicator 1min 23sec
But bitrate didn't increase at all:
Overall bit rate : 56.5 Mbps
Bit rate : 55.9 Mbps
Bits/(Pixel*Frame) : 2.023

spamrakuen
06-01-2010, 02:44 AM
Overall bit rate : 79.5 Mbps
Bit rate : 79.0 Mbps
Bits/(Pixel*Frame) : 2.857
Frame rate : 30.000 fps


This is insane :shocked:
Amazing!
Thanks for the samples.

RandyQ
06-01-2010, 02:47 AM
Little test: (very high detail images with good lightling).

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

Bit rate : 56.6 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 2.047

:bath:

Good grief, this is amazing! Does this test mean that the hacked MJPEG is ready for use? Please send us more samples of this stuff with more motion in it. I'd like to see this codec pushed to the limit!

Great work guys!

slyn4ice
06-01-2010, 02:57 AM
No, I have time indicator of 2 min...
I didn't try to record all 2 min....

Now I tried with even higher settings:
Quality settings (E1 to E4) - 754, 512, 352, 220
Table settings (E1 to E4) - 24, 24, 24, 24

Camera records; Time indicator 1min 23sec
But bitrate didn't increase at all:
Overall bit rate : 56.5 Mbps
Bit rate : 55.9 Mbps
Bits/(Pixel*Frame) : 2.023

Wait, how big is your memory card? 8GB? 16GB?

spamrakuen
06-01-2010, 03:02 AM
Wait, how big is your memory card? 8GB? 16GB?

If I'm not wrong MJPEG have a 2GB file size limit or something like that.

jobless
06-01-2010, 03:04 AM
Wait, how big is your memory card? 8GB? 16GB?

16Gb

slyn4ice
06-01-2010, 03:12 AM
If I'm not wrong MJPEG have a 2GB file size limit or something like that.

Silly me - completely forgot about the size limit. I wonder if there is a way to remove this limit - increased bitrate and removed limit - that would rock!

Cavemandude
06-01-2010, 04:29 AM
Quality settings (E1 to E4) - 512, 352, 220, 200
Table settings (E1 to E4) - 24, 24, 24, 24

SCENE 1:

Overall bit rate : 79.5 Mbps
Bit rate : 79.0 Mbps
Bits/(Pixel*Frame) : 2.857
Frame rate : 30.000 fps

download link:
http://www.nikolicnemanja.com/scene1.rar



When comparing the MJPEG file to the AVCHD files from the SCENE 1 download link, the MJPEG video has crushed black levels and a noisier image than the AVCHD videos even at that super high bitrate (79 Mbps) for the MJPEG file. The AVCHD videos were at 14 Mbps.

Add to that the disadvantage of 30p video, when most people want 24p, and 16K instead of 48K audio.

Update:

Using the VLC Player and Splash Lite Player the black levels were crushed playing the MJPEG video. Using QuickTime or Windows Media Player the black levels looked fine and comparable to the AVCHD videos. Not as noticeable noise difference either when using QuickTime Player of Windows Media Player.

Actually looking at it again with Windows Media Player, the MJPEG video is slightly darker when compared to the AVCHD video but not as drastic when I use the VLC or Splash players.

Randy

jobless
06-01-2010, 04:32 AM
Quality settings (E1 to E4) - 720, 412, 352, 300
Table settings (E1 to E4) - 24, 24, 24, 24

Overall bit rate : 80.7 Mbps
Bit rate : 80.2 Mbps
Bits/(Pixel*Frame) : 2.901

jobless
06-01-2010, 04:37 AM
Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24

Overall bit rate : 101 Mbps
Bit rate : 100 Mbps
Bits/(Pixel*Frame) : 3.630

Camera stops recording after 5 sec due to Card speed...

spamrakuen
06-01-2010, 04:40 AM
Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24

Overall bit rate : 101 Mbps
Bit rate : 100 Mbps
Bits/(Pixel*Frame) : 3.630

Camera stops recording after 5 sec due to Card speed...

:shocked:

Class 10?

jobless
06-01-2010, 04:44 AM
FILM PRO SDHC Class 6

jobless
06-01-2010, 04:50 AM
Quality settings (E1 to E4) - 720, 512, 428, 350
Table settings (E1 to E4) - 24, 24, 24, 24

Overall bit rate : 90.8 Mbps
Bit rate : 90.3 Mbps
Bits/(Pixel*Frame) : 3.266

CAMERa stops recording after 10sec...

thrice
06-01-2010, 04:54 AM
I will be donating shortly, I just have to order my GF1 first :p

jobless
06-01-2010, 04:57 AM
When comparing the MJPEG file to the AVCHD files from the SCENE 1 download link, the MJPEG video has crushed black levels and a noisier image than the AVCHD videos even at that super high bitrate (79 Mbps) for the MJPEG file. The AVCHD videos were at 14 Mbps.

Add to that the disadvantage of 30p video, when most people want 24p, and 16K instead of 48K audio.

Update:

Using the VLC Player and Splash Lite Player the black levels were crushed playing the MJPEG video. Using QuickTime or Windows Media Player the black levels looked fine and comparable to the AVCHD videos. Not as noticeable noise difference either when using QuickTime Player of Windows Media Player.

Actually looking at it again with Windows Media Player, the MJPEG video is slightly darker when compared to the AVCHD video but not as drastic when I use the VLC or Splash players.

Randy

In AE looks the same...
I applied Sharpness at 50 % and mpjeg do not show artefacts ( beside horrible picture with 50% sharpness)

Cavemandude
06-01-2010, 05:04 AM
In AE looks the same...
I applied Sharpness at 50 % and mpjeg do not show artefacts ( beside horrible picture with 50% sharpness)

Ya, I figured that might be the case but there are going to be people playing the files using a video player like I did and just wanted it to be known that this might happen.

Update:

I just dropped the files into Edius Neo Booster and the MJPEG and AVCHD files both have the same black and comparable noise levels when viewing them in the timeline.

One thing I noticed in the timeline is the audio waveform on the MJPEG file has no gap at the beginning of the video, the AVCHD files have a 1/2 second delay before the audio waveform starts. This was both on the 720P and 1080P AVCHD files.

Also, there is much more detail in the MJPEG file compared to the AVCHD file. There is a cement walkway in this video and I can see the detail and texture of the cement in the MJPEG file whereas the AVCHD file this is blurred/smoothed out more.

Randy

spamrakuen
06-01-2010, 05:56 AM
Also, there is much more detail in the MJPEG file compared to the AVCHD file. There is a cement walkway in this video and I can see the detail and texture of the cement in the MJPEG file whereas the AVCHD file this is blurred/smoothed out more.


Yeees.... MJPEG rules. Also, MJPEG is easier to edit.
I hope tester13 will achieve 24 o 25 fps. :Drogar-Happy(DBG):
And at these insane bitrates, plus after lastest tests, we can even dream on 1080p MJPEG (just dreaming, right now...)

androoow
06-01-2010, 06:05 AM
i get the feeling we are about to reach Warp speed with these MJPEG developments .. a major breakthrough is almost here?? quite exciting :)

bumkicho
06-01-2010, 06:22 AM
After looking at NURBS, I couldn't help it. I had to apply hacked patch with the same setting he used in order to see for myself. AND I SEE A LOT MORE DETAILS IN MPEG HD MODE COMPARED TO AVCHD SH MODE.

Setting I used is below.

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

* NOTE: I USE TOSHIBA SDHC 16G CLASS 4

MPEG result - around 43 Mbps
12200419

Here is AVCHD SH result
12199736

From Mov 30fps to MP4 24fps test
12200887

tyampel
06-01-2010, 06:23 AM
One thing I noticed in the timeline is the audio waveform on the MJPEG file has no gap at the beginning of the video, the AVCHD files have a 1/2 second delay before the audio waveform starts. This was both on the 720P and 1080P AVCHD files.

Randy[/quote]

The audio gap may have been caused by pressing the "Record" button a bit too long. The manual warns about this.
Don't remember if it only affects AVCHD recordings.

Tameside
06-01-2010, 06:37 AM
To be honest I'd be happy with 1080p30 or 25p, mud free at 30Mb/s.

Ozpeter
06-01-2010, 06:43 AM
There is always an audio gap in GH1 AVCHD recordings (I mean, in the "stock" camera).

Cavemandude
06-01-2010, 07:04 AM
There is always an audio gap in GH1 AVCHD recordings (I mean, in the "stock" camera).

Yep, but it looks like MJPEG files with the GH1 don't have that "gap" problem, which is a good thing, if that's what you want to record with instead of AVCHD.

Randy

Slice
06-01-2010, 07:13 AM
Is it possible to hack a higher MJPEG color depth of 4:2:2 or 4:4:4?

alignment1
06-01-2010, 07:27 AM
After looking at NURBS, I couldn't help it. I had to apply hacked patch with the same setting he used in order to see for myself. AND I SEE A LOT MORE DETAILS IN MPEG HD MODE COMPARED TO AVCHD SH MODE.

Setting I used is below.

Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

* NOTE: I USE TOSHIBA SDHC 16G CLASS 4

MPEG result - around 43 Mbps
12200419

Here is AVCHD SH result
12199736

From Mov 30fps to MP4 24fps test
12200887
Yep leavs in trees look better, but still see artifacting on the fence as well as the lamp post, I'm pretty sure it's the sprinklers messing w/ the codec. And in the Mov 30fps to MP4 24fps test-- it looks like some chromatic aberration on the wet parts of grass.

getting there:smile:

NURBS
06-01-2010, 08:04 AM
My test for this settings:

Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24

Bit rate : 77 Mbps <- Nice result
Bits/(Pixel*Frame) : 3.000
Duration: 10.00 sec.

Camera don't freezing for all record time. (Sandisk, Class-4, 9MB/s)

bumkicho
06-01-2010, 08:20 AM
My test for this settings:

Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24

Bit rate : 77 Mbps <- Nice result
Bits/(Pixel*Frame) : 3.000
Duration: 10.00 sec.

Camera don't freezing for all record time. (Sandisk, Class-4, 9MB/s)

Wow!

How long of recording time do you get with this setting, NURBS?

Isaac_Brody
06-01-2010, 08:42 AM
This is pretty promising, to my eyes the 720P Mjpeg looks as good as the 1080 avchd. Man you guys are close to cracking this. :)

tyampel
06-01-2010, 08:49 AM
In the email the test duration is stated to be 10.00 seconds.
Perhaps it should be tested fpr a bit longer, as freezes seem to occur after 10 seconds.
How about testing for 30 seconds just to see if there are no freezes?
For posting the files can be cut to clips of 10 seconds.

bumkicho
06-01-2010, 08:53 AM
My test footage is longer than 30 sec. I also captured about 3 min long footage with no problem.

jobless
06-01-2010, 08:57 AM
Quality settings (E1 to E4) - 720, 412, 352, 300
Table settings (E1 to E4) - 24, 24, 24, 24

Time test:

Static high detail scene... 1 min 25 sec and after that camera stops recording ( not freezing)
file size 766MB
Bit rate : 74.7 Mbps
Bits/(Pixel*Frame) : 2.701

Same settings, Camera in motion... 2min 14 sec and I stopped the camera... File size 1,12GB
Bit rate : 71.9 Mbps
Bits/(Pixel*Frame) : 2.600

Barry_Green
06-01-2010, 09:27 AM
Hey guys, thanks so much for all the detailed testing! Getting the MJPG mode over 100mbps... that's starting to feel like a good old intraframe mud-free no-long-GoP happy solution there! And if tester can find a way to enable the sensor scanning at 24fps during MJPG mode... that would be fantastic. And the data rate would drop to 80mbps with no loss in quality, so it should work easily on any Class 10 card.

tester13, what is your opinion on these 720p mjpg breakthroughs -- do you think you could find whatever setting it is that triggers the sensor to scan at 24fps instead of 60? I know you're interested in making 1080p MJPG too, but it seems like an incredibly useful modification is so very close here; I know I'd be quite happy with 720p 24p MJPG @ 80mbps until the miraculous 1080p MJPG is ready! :)

You guys all rock, thanks for the most useful DSLR-related thread in all of internet history!

tyampel
06-01-2010, 09:33 AM
The first setting results in 9 MB per second SD usage.
At this rate the 2GB limit based on the FAT16 format is reached in 222.2 seconds ot about 3.7 minutes.

The second case results:
8.35 MB per second
Maximum recording time for a 2GB file 240 seconds or 4 minutes.
The interesting thing to see is how the camera is handling opening a new file when the limit is reached. I am afraid it will lose some video and audio during this split operation.

NURBS
06-01-2010, 10:16 AM
My test for this settings: (lot of detail in scene)

Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24

Format : M-JPEG
Duration : 5s 500ms
Bit rate mode : Variable
Bit rate : 105 Mbps
Frame rate : 30.000 fps
Bits/(Pixel*Frame) : 3.784
Stream size : 68.6 MiB (100%)

Camera recording stop at 5s 500ms, for low detail scene it's not actual. (Sandisk, Class-4, 9MB/s)

Link: http://rapidshare.com/files/394102026/P1010042.MOV

jobless
06-01-2010, 10:29 AM
I tried to rich 2gb limit with this settings:
Quality settings (E1 to E4) - 720, 412, 352, 300
Table settings (E1 to E4) - 24, 24, 24, 24
Indoor shooting
Bit rate : 74.5 Mbps
Bits/(Pixel*Frame) : 2.695

Time indicator was at 1 min and 24 sec when I started shooting. However clip is 3min and 31sec and 1,84 GB. It was all normal during the recording and the camera didn't show any errors...

sunboy
06-01-2010, 10:43 AM
How about the audio quality in MJPEG mode?

Vitaliy Kiselev
06-01-2010, 11:21 AM
Thanks for testing

If fact main target was determinating top speed of SD writing speed for GH1.
Speed is clearly enough.
I'll change MJPEG patches with time, but today they are already useful for someone shooting short scenes and loving 30fps.

One thing I can't understand is why we don't have all this GH1 features in official firmware.
I don't talk about AVCHD here, as we don't know it's real limits yet.
But 1280p30 with high bitrate and 48Khz sound and 1920p24 with good sound also must be implemented officially, because we need file spanning. And current firmware needs small changes to implement both 1920 and spanning (if any Panasonic programmer is reading this - you just need to make enough files at the start with header part completed, then change file handle as one file size is near 2Gb, etc, after this just delete all unused files, fast, dirty and working).
I believe most indie makers will be ok with 1920p24 MJPEG at 100Mbit?

Also tried HDMI playback. Looks very good at fullhd resolution.
Why GH1 blocks live monitoring?
I believe we have one reason - native buffer is 1920p24 or 1920p25.
And HDMI output is 60i. So to monitor it must do AVCHD encoding->AVCHD decoding and only then show at HDMI output.
We''ll try to look at this later.

spamrakuen
06-01-2010, 11:28 AM
I believe most indie makers will be ok with 1920p24 MJPEG at 100Mbit?

720p@24/25 would be great, 1080p@24/25 would be awesome. :shocked:

bumkicho
06-01-2010, 11:41 AM
Thanks for testing

Thank you, tester13, for your work.


today they are already useful for someone shooting short scenes and loving 30fps.

Yes!!


I believe most indie makers will be ok with 1920p24 MJPEG at 100Mbit?

It would be a dream come true. I would be thrilled as much with AVCHD 24Mbps in FHD mode or SH mode. Tester13, any progress on AVCHD side?

Barry_Green
06-01-2010, 11:45 AM
Speaking strictly for myself (unless anyone else wants to join in)... while we would all love 1080/24p MJPG at 100 megabits, we'd be almost as happy with 720/24p MJPG at 100 megabits. So I guess the question is: which one is easier to implement? Is it easier to get the MJPG system to understand and properly compress 1080p data, or is it easier to get the chip to scan at 24fps when in 720p mode?

Either solution would be nigh unto the holy grail.

I'm ready to forget the AVCHD option if we could get functional MJPG.

Excellent work all the way around!

Isaac_Brody
06-01-2010, 11:49 AM
Speaking strictly for myself (unless anyone else wants to join in)... while we would all love 1080/24p MJPG at 100 megabits, we'd be almost as happy with 720/24p MJPG at 100 megabits.
Either solution would be nigh unto the holy grail.

I'm ready to forget the AVCHD option if we could get functional MJPG.

Excellent work all the way around!

+1. Either one would be great as long as there's 24P. I could care less about AVCHD. With the choice of them both I would shoot strictly mjpeg. File spanning would be amazing.

Vitaliy Kiselev
06-01-2010, 11:51 AM
while we would all love 1080/24p MJPG at 100 megabits, we'd be almost as happy with 720/24p MJPG at 100 megabits. So I guess the question is: which one is easier to implement? Is it easier to get the MJPG system to understand and properly compress 1080p data, or is it easier to get the chip to scan at 24fps when in 720p mode?

We don't know.
At first we must understand how 848 wide MJPEG is implemented.
As for MJPEG encoder 1080p patch proves that encoder is up to the task, we just need to feed proper image.

NURBS
06-01-2010, 11:53 AM
Live HDMI output it's very important our destination, ;-) it change the game! :bath:
'Live TV out' in record time too important thing (for steadycam operator) ;-)

Regards
NURBS

John Caballero
06-01-2010, 11:54 AM
Originally Posted by Barry_Green
Speaking strictly for myself (unless anyone else wants to join in)... while we would all love 1080/24p MJPG at 100 megabits, we'd be almost as happy with 720/24p MJPG at 100 megabits.
Either solution would be nigh unto the holy grail.

I'm ready to forget the AVCHD option if we could get functional MJPG.

Excellent work all the way around!

+1! MJPG 720p24 or if possible 1080p24 would make me forget AVCHD as well!

philiplipetz
06-01-2010, 12:01 PM
Thanks for testing
I believe most indie makers will be ok with 1920p24 MJPEG at 100Mbit?


yes, and more yes

andyjar
06-01-2010, 12:11 PM
Whats the advantage of MJPEG over AVCHD?

TrueIndigo
06-01-2010, 12:14 PM
Wonder if the eventual Panasonic AF-100 will deliver images as good as a mature version of this mjpeg hack...?

philiplipetz
06-01-2010, 12:21 PM
Wonder if the eventual Panasonic AF-100 will deliver images as good as a mature version of this mjpeg hack...?

Bet the AF-100 has better anti-aliasing filter, optimized for video not stills.

tyampel
06-01-2010, 12:44 PM
Whats the advantage of MJPEG over AVCHD?

With MJPEG each frame is a JPEG image. With AVCHD only key frames are carrying full frame information. These are then followed by frames that carry the changes.
In fast moving situations the CODEC or its implementation is not good enough to catch and encode those frame to frame changes, resulting in "mud".
This is a simplified version, but should answer your question.

tsupka
06-01-2010, 12:46 PM
Live HDMI output it's very important our destination, ;-) it change the game! :bath:
'Live TV out' in record time too important thing (for steadycam operator) ;-)

Regards
NURBS

and for more complicated production situations - for client's monitor support for example, for cranes, crashcams, focuspullers and so on

will be great feature even in no perfect quality

but thanx tester13 anyway))

bumkicho
06-01-2010, 12:48 PM
I'm ready to forget the AVCHD option if we could get functional MJPG.

I am not sure if I can say the same. I picked GH1 over 7D because it allows, in FHD and SH mode, me to shoot as long as SDHC card can hold. 720/24p at 100Mbps would be great! But it will have less than 2 minute limit. With the current patch, 720/30p at around 55Mbps is only for 3 minutes. It is great, but I am not sure if I want to give up AVCHD/24Mbps with unlimited length.

Jasketti
06-01-2010, 01:33 PM
and for more complicated production situations - for client's monitor support for example, for cranes, crashcams, focuspullers and so on

will be great feature even in no perfect quality

but thanx tester13 anyway))

Plus the ability to use an external recorder like nanoflash, 4:2:2 I-frame only 280 Mbps!

This is fantastic work. You guys are awesome! Hope you can find the way to open up the HDMI live with as good quality as possible. It would solve all the problems for.

Kellar42
06-01-2010, 01:37 PM
+1 for 1080 MJPG. This is amazing work.

spamrakuen
06-01-2010, 02:52 PM
Okey ;) in the night time I'll di it.


:engel017:

Barry_Green
06-01-2010, 02:53 PM
Wonder if the eventual Panasonic AF-100 will deliver images as good as a mature version of this mjpeg hack...?
I would expect that no matter how much progress can be made on the GH1, it's still a $600 camera body. The AF100 is 10x more expensive and should have better everything -- better dsp, better compression, better OLPF, and on and on.

sirk
06-01-2010, 03:21 PM
:engel017:

Low light test MJPEG vs AVCHD
MJPEG:
Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

with this settings top bitrate is up to 55Mbps.

This clip here is 16.7 Mbps
MJPEG
http://rapidshare.com/files/394191609/P1040169.MOV.zip.html

AVCHD
http://rapidshare.com/files/394191366/00109.MTS.zip.html

Sirk

Ozpeter
06-01-2010, 03:23 PM
Some of the recent post speak of file spanning as if it's a missing feature, but it works fine here on original firmware AVCHD. Well, almost fine, as there seems to be a tiny discontinuity at the join between two files, though that could be a function of the NLE's file handling. But maybe I'm misunderstanding what's been said.

Lpowell
06-01-2010, 03:34 PM
I studied the first two posts in this thread, and the increased quality results for MJPEG mode look promising. However it's not clear to me how the camera makes use of the four MJPEG E1 - E4 Quality settings, the four MJPEG E1 - E4 Tables, and the two Bitrate Adjustment settings.

It would be helpful if someone could post a summary (either here or in the first two posts) of min, max and default values for these Quality and Table settings, and how to gauge the bitrate settings to accomodate increased Quality and Table settings.

Car3o
06-01-2010, 03:35 PM
48p would be pretty sweet for mjpeg. i know the sensor readout is 24p, but if it's capable of doing 48p, then that gives it the ability to do slow-mo and easily drop it into a 24p timeline no problem. that's how avatar was filmed at least.

sirk
06-01-2010, 03:37 PM
that's how avatar was filmed at least.

could you also please make skin tones blue?

IanB
06-01-2010, 03:55 PM
Low light test MJPEG vs AVCHD
MJPEG:
Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

with this settings top bitrate is up to 55Mbps.

This clip here is 16.7 Mbps
MJPEG
http://rapidshare.com/files/394191609/P1040169.MOV.zip.html

AVCHD
http://rapidshare.com/files/394191366/00109.MTS.zip.html

Sirk

Assuming you didn't do an CC to these shots, that is just an astounding differance. HD for REAL...I might export these to blue-ray to see how they play on the plasma, wife is going to be really worried when she sees me watching someone play with a floodlight on the plasma...but, she is used to it by now..

spamrakuen
06-01-2010, 04:00 PM
Low light test MJPEG vs AVCHD
MJPEG:
Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

with this settings top bitrate is up to 55Mbps.

This clip here is 16.7 Mbps
MJPEG
http://rapidshare.com/files/394191609/P1040169.MOV.zip.html

AVCHD
http://rapidshare.com/files/394191366/00109.MTS.zip.html

Sirk

Thank you, both look quite good.

JerryB
06-01-2010, 04:04 PM
Great to see that the mjpeg option is picking up steam. Will definitely be donating more if I see progress toward 720/24p and 1080/24p! Keep it up Tester13 and all the other testers, your doing great work for low-budget indie-filmmakers. Is there going to be a recording limit on the mjpeg or is that in the works of being hacked?


PS. Figuring out the audio limiter situation to have control of audio levels would be a MAJOR plus for run-and-gun regardless of quality. Not the priority right now of course, but I do hope it's still on the list!

Vitaliy Kiselev
06-01-2010, 04:20 PM
As for 720/p24, most probably it won't happen. But who knows.
1080p24 is probable candidate.
But reality can be tough.

Current progress.

I can change 1920p24 sensor settings and make weird picture for AVCHD.
Problem is that it is very risky part, any error can lead to freezing during setup of liveview, and I won't be able to change firmware. We'll leave that for GF1.

JerryB
06-01-2010, 04:42 PM
Although it may have been said somewhere in the 193 pages of this forum, could you please clarify...

Why do think 720/24p Mjpeg will probably not be possible?

Thanks!

spamrakuen
06-01-2010, 04:47 PM
As for 720/p24, most probably it won't happen. But who knows.


Do you think MJPEG 720p@50fps is doable?

Vitaliy Kiselev
06-01-2010, 05:05 PM
been busy lately. but I managed to make an update to wingraph32 for you (hurray!). As requested from before, made changes for insensitive text search from either starting symbol or as a substring. Also added address lookup in IDA.

Important changes to mouse-click operations:
- Right-click will now select your root node
- Double left-click will attempt to lookup the node in IDA (if you have it running in background).



Address lookup don't work.
Even do not switch to IDA window.

Right click seems not a good solution - as we have context menu or right click.
Also hiding all nodes completely is not good ideas as I now understand, just make them much lighter, so they won't clutter view.


Tried assuming "blue" colour to be root node but it doesn't work. Reason being that (in primer example), depth search begins at root and travels AWAY from node (note the arrow directions along edges). The blue root in primer, arrows all point TO the root... so even if I made that root automatically, setting depth of 1, 2, 3, etc... will still only show that one node and nobody else. I guess the only way around this is to do a reverse depth search.

Yes, as this example is make as calling xref to current identifier (and yes, in this case reverse search seem appropreate).
In case of "calling from" graphs all are ok.

Vitaliy Kiselev
06-01-2010, 05:14 PM
Although it may have been said somewhere in the 193 pages of this forum, could you please clarify...
Why do think 720/24p Mjpeg will probably not be possible?


We have encoder setting (here we can do that we want).
And we have sensor settings. This is much tougher.
Sensor seem to have constant read speed without interruptions.
Each setting have many parameters. And without sensor documentation it is almost impossible to change something to get useful picture.
And preset settings allow only resolutions that we already have.
May be scaling down from 1920 to 1280 can be made.
This is why it is interesting to know how 848 is implemented.

dracore
06-01-2010, 06:44 PM
Address lookup don't work.
Even do not switch to IDA window.
.

Can you try this test version:

http://rapidshare.com/files/394239568/wingraph32.rar.html

Double click on a node. Most likely it won't switch to IDA. Post me the contents of the wingraph32.log file that it creates

GrgurMG
06-01-2010, 07:00 PM
Howdy all. I have been quietly following this thread for quite awhile now and am very interested in what I've seen. I definately plan to join the tester ranks in addition to donating, in the near future. The dilemma I'm faced with before I do so is whether or not I might exchange my newly purchased GH1 for another one within the next week or so.

I have the GH1 video banding issue staring me at the face and with all the talks on other threads/forums of people with a higher degree of issue with this then others... and on the flip side people whos GH1s supposedly suffer very little from this... I'm trying to decide which of these catagories mine falls into and if there might be merit to attempting to exchange it for a new one. Thus, I obviously don't want to mess with firmware changes until I'm sure I will not be returning the camera or otherwise I get the replacement.

In relation to the breakdown of the cameras functions discussed in this thread, I notice that video noise and banding is largely supressed by some sort of post camera filtering that doesn't initiate until after the record button is pressed. I wonder how much this process might task the cameras processors and whether or not that had any bearing on there being no HDMI preview on the device. Whether it be a lack of processing resources or just fear of tasking and overheating the system.

Car3o
06-01-2010, 07:09 PM
banding has been discussed in at least a dozen threads. it's a hit a miss really. i've dealt with two different gh1'1, one american version and another japanese and neither had banding unless i was pushing the iso or shooting in dim light scenes. others have said it went away after it "burned" in, but I personally think that's a bunch of farce. in regards to your camera, it's kind of up to you, no clear answer.

Vitaliy Kiselev
06-01-2010, 07:16 PM
Banding is really nature of Panasonic LiveMOS chips (CMOS also have similar thing).
In photo they use smart algorithms with dark frame (viewefinder goes dark for this reason?) or something similar.
My Panasonic camcorder have same banding.
Other camcorder is 3CCD and do not have such issues.
In fantazy land they could allow user to make camera profile using gray cards and running through various resolutions and sensor temperatures. Such thing and simple processing can reduce banding very much.

GrgurMG
06-01-2010, 07:21 PM
That pretty much what I figured.. it's really a hard thing to substantiate.. and I've been through ALL the the threads. It's not so much that I'm horrified with the camera performance or anything.. it's just the nagging possibility that an exchange might produce a better working model for me.

Should anyone have time to spare... at max (+3) exposure comp, shutter priority (1/30) and 1600 ISO the banding is clearly visible while recording with the lens cap on. If anyone cares to tell me if they get the same, less, or more visible/recorded banding at that setting, here's a quick clip of what I get.....

http://www.jerseyforged.com/sampleclips/banding_1600iso_maxexpcomp.zip

Also, on still mode "s" at the same settings the banding is extreme (again with lens cap on) and red colored until the record button is hit (the previously mentioned filtering). Those are obviously crap settings, but it's the best way I can think of to easily compare my banding levels to someone elses.

In any case, I have another week to return the camera and if I don't I'll begin running any needed tests next week.

Vitaliy Kiselev
06-01-2010, 07:22 PM
Double click on a node. Most likely it won't switch to IDA. Post me the contents of the wingraph32.log file that it creates

Switch working now (but going to function not workingm I believe it must be ok at this testing stage).
If you can identify window by calling process (process that calls WinGraph) it'll be better, as I have many IDA windows sometimes.

Vitaliy Kiselev
06-01-2010, 07:23 PM
That pretty much what I figured.. it's really a hard thing to substantiate.. and I've been through ALL the the threads. It's not so much that I'm horrified with the camera performance or anything.. it's just the nagging possibility that an exchange might produce a better working model for me.

Please, make other thread for this.

GrgurMG
06-01-2010, 07:26 PM
Well, I figured since I plan to donate my time and money to this effort.. and we already have very thorough testers on this very active thread... I'd bring it up here. But I relent.

dracore
06-01-2010, 08:20 PM
Switch working now (but going to function not workingm I believe it must be ok at this testing stage).
If you can identify window by calling process (process that calls WinGraph) it'll be better, as I have many IDA windows sometimes.

Can you post the wingraph32.log file? I want to see specifically why the going to function is stopping.

Vitaliy Kiselev
06-01-2010, 08:49 PM
Can you post the wingraph32.log file? I want to see specifically why the going to function is stopping.

Where I can find it?
I don't see it IDA directory.

dracore
06-01-2010, 08:58 PM
Where I can find it?
I don't see it IDA directory.

You're running the test version right? Should be in the same folder as the exe. If not, do a hard drive search :)

Vitaliy Kiselev
06-01-2010, 09:04 PM
You're running the test version right? Should be in the same folder as the exe. If not, do a hard drive search :)

May be you use any folder specific to your computer as I can't find log file.
Just use massagebox or something like it :-)

dracore
06-01-2010, 09:40 PM
May be you use any folder specific to your computer as I can't find log file.
Just use massagebox or something like it :-)


Ok here's another test version:

http://rapidshare.com/files/394277280/wingraph32.rar.html

Vitaliy Kiselev
06-01-2010, 09:54 PM
Shows box with "hproc:11732966"

dracore
06-01-2010, 10:14 PM
hmm.. that's it?

just curious... but is there a .log file? :)

Vitaliy Kiselev
06-01-2010, 10:24 PM
hmm.. that's it?

just curious... but is there a .log file? :)

It is up to you to know.
I don't see any log file.
I really suggest you to see examples on the web, as task is quite typical (plus as far as I know you need separate DLL to send keystrockes to application).

dracore
06-01-2010, 10:31 PM
Yes, I was just surprised that it stopped there. Didn't even get to send any keyboard events...

OK, I need a screenshot in wingraph32 of the node that you are double clicking on

Vitaliy Kiselev
06-01-2010, 10:41 PM
Yes, I was just surprised that it stopped there. Didn't even get to send any keyboard events...

OK, I need a screenshot in wingraph32 of the node that you are double clicking on

It happens on any node :-)
I tried many-many various graphs.
You better contact me by PM or email, af this have no interest for others.

Vitaliy Kiselev
06-02-2010, 02:00 AM
The more I work on this project the more I see GH1 potential.
Today we are like blind kittens.
In reality we need either debug firmware build or object files (they won't provide sources anyway). Speed will be much faster and results will be better.

We must build small team with sole purpose to reach necessary people in Panasonic Lunix division development department.
Under this I mean information digging, many calls via cheap international Skype, etc. to get necessary persons and talk to them.
I do not mean any useless petitions, etc
Contact me by email and PM if you want to become member of such team.
I specially invite people from Japan.

TurboTwo
06-02-2010, 03:53 AM
The more I work on this project the more I see GH1 potential.



Hi Tester,That`s a pretty interesting comment!..what have you discovered/seen that makes you say this,in other words what`s the potential of GH1 in relation to this statement that you hadn`t seen before? :-)

Elias
06-02-2010, 06:19 AM
43rumors mentioned this interview:
http://translate.google.com/translate?js=y&prev=_t&hl=it&ie=UTF-8&layout=1&eotf=1&u=http%3A%2F%2Funwire.hk%2F2010%2F05%2F18%2Fpanaso nic-interview%2F&sl=auto&tl=en
"market has a lot of users would like to join the GH system update function and shooting professional HD format, we will respond to their demands, we will be adding more professional-grade HD video feature inside."

Sounds like firmware update is coming???

schakal.
06-02-2010, 06:28 AM
Or GH2 ? :)

stoneinapond
06-02-2010, 06:34 AM
I believe they were referring to the GH2

e-steve
06-02-2010, 06:44 AM
The more I work on this project the more I see GH1 potential.
Today we are like blind kittens.
In reality we need either debug firmware build or object files (they won't provide sources anyway). Speed will be much faster and results will be better.

We must build small team with sole purpose to reach necessary people in Panasonic Lunix division development department.
Under this I mean information digging, many calls via cheap international Skype, etc. to get necessary persons and talk to them.
I do not mean any useless petitions, etc
Contact me by email and PM if you want to become member of such team.
I specially invite people from Japan.

Hope you can get this off the ground and that Panasonic can see how this could put the GH1 into legendary status and sell so many more GH1s!

Vitaliy Kiselev
06-02-2010, 11:33 AM
Hope you can get this off the ground and that

I can't do all myself.
Problem is very little number of people understand it.

I want to emphasize that this project can stop without information and tools.

And we really need Panasonic compilers (they are provided with any AM3 based controller).
Plus first hands information.
I really mean very many calls and collecting all contacts is special database.
We can start from: http://panasonic.net/corporate/segments/sc/

If you passively read this thread, it is time to join!
Contact me at info dot pentax-hack dot info.

NPrincen
06-02-2010, 12:57 PM
And we really need Panasonic compilers (they are provided with any AM3 based controller).
Plus first hands information.

Maybe this should be our next buy item. Seems more important than getting a G-1 or GF-1 body. Did a little research and found this:

http://www.object.co.jp/panamc/panamc_eng/prodcut-info.html

I have not been able to find the price of the AM3 development board, but the AM1 board can be bought here:

http://search.digikey.com/scripts/DkSearch/dksus.dll?Detail&name=MMK01-C78-ND

This is available for $358.50. It includes the C compiler and the debugger, but may be locked to the specific MN101 processor. I will continue to try to search for a MN103 development board.

Thanks Tester13 for your efforts!

Vitaliy Kiselev
06-02-2010, 01:08 PM
Maybe this should be our next buy item. Seems more important than getting a G-1 or GF-1 body.


We already have these bodies, and don't have any need to buy AM3 board.


Did a little research and found this:
http://www.object.co.jp/panamc/panamc_eng/prodcut-info.html


I know about this guys.
You can ask them for evaluation versions of compiler.


I have not been able to find the price of the AM3 development board, but the AM1 board can be bought

First, we don't need AM3 board.
All we need is code provided with this board.
With htis code I'll be able to download compilers myself.

As I specially outline - C and ASM compilers themselfs are not enough, we need to establish good contacts and get as much information as possible.
And you don't need to be programmer to help in this areas.

NPrincen
06-02-2010, 01:29 PM
First, we don't need AM3 board.
All we need is code provided with this board.
With htis code I'll be able to download compilers myself.

I understand that you only need the code to activate the compiler. The only legal way to get the code is to buy the board. I have experience with this for another type of microprocessor. Full versions of the compilers for professional use are very expensive (sometimes $10,000). But, when you buy a devlopment board, they throw in a version of the compiler for "educational use." There hope is that you end up deciding to make a product with the microprocessor and later need the full compiler version. You get a simplified version of the compiler for next to nothing ($358). Only way to get this compiler for little money is to buy the board.

Any other way of obtaining the compiler activation code is software piracy and I thought we were trying to keep this project clean from a legal perspective.

Vitaliy Kiselev
06-02-2010, 01:44 PM
Any other way of obtaining the compiler activation code is software piracy and I thought we were trying to keep this project clean from a legal perspective.

No, it is not piracy, as DebugFactory can be distributed freely according to license and evaluation compiler versions can be also provided to any third-parties without problems. This is exactly that we are doing.

PappasArts
06-02-2010, 04:13 PM
Been testing NURBS settings of:

[ Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24

Bit rate : 77 Mbps <- Nice result ]



Amazing mjpeg performance. However my Kingston 4gig class 6 wasn't fast enough for high detailed shots. I got a message informing that my card wasn't fast enough.

What Class 10's are good?

For Updates: MIKOS • MLPappas on Twitter
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://MIKOSarts.COM

Car3o
06-02-2010, 04:59 PM
those settings are pushing the limits of class 10 cards. 77mbit reaches upward pas 9MB/sec. I'd look at Sandisk Ultra's II or III's to reach those writing speeds.

PappasArts
06-02-2010, 05:16 PM
those settings are pushing the limits of class 10 cards. 77mbit reaches upward pas 9MB/sec. I'd look at Sandisk Ultra's II or III's to reach those writing speeds.

Thanks.. Is there a difference between the Ultras and class 10, as in their speed?

BTW the quality is phenomenal at that speed. I shot some quickies with the anamorphic, which softens the image a little, and missed the focus a tad. However no codec artifacts at all. Amazing!

Here's are a few frames from the 1920x812 upscale from the 720.

For Updates: MIKOS • MLPappas on Twitter
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://MIKOSarts.COM

BlackMagic Digital Camera Info Thread: http://www.bmcuser.com/showthread.php?28-BlackMagic-Cinema-Camera-Technology-Announcements-amp-Latest-Info

GrgurMG
06-02-2010, 06:12 PM
On the issue of SD cards needed for these write speeds...

Has SanDisk dropped the numbered levels of their Extreme cards? I notice tester13 is shooting to get a SanDisk Extreme III 30MB/S 32GB SDHC... but upon looking it appears the newest sold models are simply SanDisk Extreme 30 MB/S. http://www.sandisk.com/products/dslr/sandisk-extreme-sdhc-cards--

Furthermore, I found a website or two that displayed the "SanDisk Extreme III 30MB/s" specifically.. and I noticed a "Class 6" rating for some reason. http://www.mydigitallife.info/2008/08/29/sandisk-extreme-iii-sdhc-30mbs-edition-cards/

Was the "Extreme III" cards simply last years branding or is there still any quantifiable difference between "Extreme" and "Extreme III"?


Thanks.. Is there a difference between the Ultras and class 10, as in their speed?.

I think SanDisk Ultra is specifically their branding for their midrange Class 4 (as appose to standard Class 2) cards... which SanDisk selfrates at 15/MBs. Their newest Class 10 Extemes are rated for 30MB/s... with older models I believe rated for 20MB/s and some labeled Class 6.

Car3o
06-02-2010, 06:19 PM
Ive juggled with how the Ultra rate their speeds before and debated it, but I think it was simply before the industry "standardized" how class ratings were rated. Ultra's seemed to be ahead of the game with their II's and III's rated at 15mb and 30mb, but that's also read/write speeds not necessarily write. If there's a difference, yeah, I think so. Their class 10 and 6 are faster than industries class 10's and 6's. In reality Ultra II are like class 15 and Ultra III are like class 30.

GrgurMG
06-02-2010, 06:28 PM
There's alot of constant rebranding going on it, for instance their SanDisk VideoHD cards used to be branded Class 6 then they dropped it to Class 4. In fact many websites still have the picture of the old packaging, such as Newegg here http://www.newegg.com/Product/Product.aspx?Item=N82E16820171384&cm_re=VideoHD-_-20-171-384-_-Product

... but you'll get the Class4 marked cards in the mail.

sirk
06-02-2010, 06:44 PM
I use a Panasonic SD-HC Class 4 card. It works perfectly with 55Mbps, maybe higher

savagess
06-02-2010, 06:54 PM
Get the bench test program, ATTO Disk Bench mark, then you can be sure.
I bought an el-cheapo Class 6 card.
In the bench test, the read and write speeds for 32kb upwards to 8 mb were above the 10 MB / Sec...

PappasArts
06-02-2010, 07:08 PM
I tried the Sandisk Extreme 8gig 30MB/s one : http://www.sandisk.com/products/imaging/sandisk-extreme-sdhc-cards-

Shots of high detail like 106.09 mbits/s stopped after a 5 seconds.
shots below 70mbits seem not to stop.

For Updates: MIKOS • MLPappas on Twitter
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://MIKOSarts.COM

Car3o
06-02-2010, 07:30 PM
I tried the Sandisk Extreme 8gig 30MB/s one : http://www.sandisk.com/products/imaging/sandisk-extreme-sdhc-cards-

Shots of high detail like 106.09 mbits/s stopped after a 5 seconds.
shots below 70mbits seem not to stop.

how awesome will it be if we can get 24p out of the 720p. or better yet 1080. What were your 70mbit settings?

PappasArts
06-02-2010, 07:47 PM
I use a Panasonic SD-HC Class 4 card. It works perfectly with 55Mbps, maybe higher

Sirk, Was this your settings?

[ Low light test MJPEG vs AVCHD
MJPEG:
Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

with this settings top bitrate is up to 55Mbps. ]


The settings I'm using are topping at 106bps. Way to much for even the sandisk. I would like it to top at 70mbs

Rayzrsharp
06-02-2010, 07:47 PM
"You guys all rock, thanks for the most useful DSLR-related thread in all of internet history!"

Absolutely!! This is INCREDIBLE!!

PappasArts
06-02-2010, 07:51 PM
how awesome will it be if we can get 24p out of the 720p. or better yet 1080. What were your 70mbit settings?

24 would be sweeeeet.....

I'm using NURBS settings of:

[ Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24

However it's topping at 106 if there is a lot of detail in the shot. This even stops my Sandisk Extreme 10 card.

GrgurMG
06-02-2010, 08:15 PM
Pardon if I missed this in a post somewhere, but do we know exactly how the GH1 writes to the card.. that is is it in certain sized (KB/MB) bursts from a buffer or anything? Can we say for sure what's lower, the cameras -continuous- write limit or the best Class 10 cards?

Here's some SanDisk benchmarks for anyone whose interested http://www.legitreviews.com/article/1172/2/.

Vitaliy Kiselev
06-02-2010, 09:06 PM
Pardon if I missed this in a post somewhere, but do we know exactly how the GH1 writes to the card.. that is is it in certain sized (KB/MB) bursts from a buffer or anything? Can we say for sure what's lower, the cameras -continuous- write limit or the best Class 10 cards?

Of course Sandisk must have more potential than camera.
Camera top limit is from 72Mbit to 110Mbit according to all tests.
One of the goals for latest patches was to approximetely find this limit.

Isaac_Brody
06-02-2010, 09:24 PM
Does the 1080 mjpeg patch work? At a lower data rate? Or is it still buggy?

Car3o
06-02-2010, 10:03 PM
Does the 1080 mjpeg patch work? At a lower data rate? Or is it still buggy?

last i read it was still buggy. gonna give it a shot tomorrow see what's up.

JackBayer
06-02-2010, 10:21 PM
Hello!

Iīve been reading along now since something like 60 pages or so... and I think it is very much time to point something out:

This very thread (already, even though tester13 has obviously much more in mind) is one of the reasons why forums and communties do and should exist: Few people of same interests get together and find a way to get something impossible done. Even if it means to collect some money to buy a piece of hardware and deliver it to the one guy who can help them out! Itīs insane and incredible and it could only work out (which it already has) because there are people here spending time and effort and have been doing so since long before I knew what a hvx200 (my beginnings) or dvx or whatever was!

I used to present a TV show in Germany where this whole community thing was a first-timer and we succeeded in quite a few ways, but this here is really new. For me and everyone involved that I know.

Thank you tester13! Iīd guess that yes, you most likely could have antecipiated this kind of feedback but (at least for me) the thought of getting people to buy a cam for you, which I think wasnīt even your idea in the first place, came kind of unexpected? Itīs just awesome!

I bought my GH1 pretty much right about when it became available for the first time in Germany. I remember the guy on the phone of the WEX-camera store saying "yes, we have 2 in stock" a year ago. then it was said that despite everything the GH1 coped better with rolling shutter issues than Canonīs and thatīs when I went for it. Well, today Iīm just starting to think that I am able to use it properly and what do I have to read: "yeah, 1080p mjpeg seems not so impossible..." :)
Come on... :)

Right from indiecam-heaven...

Patrick

Oleg2d
06-03-2010, 12:17 AM
My test for this settings:

Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24


Little test with the NURBS settings:
(Long focus and camera fast motion)
http://rapidshare.com/files/394703133/P1140391.MOV.html
Duration : 5 s 500 ms.
Bit rate : 106 Mbit/sec
Bits/(Pixel*Frame) : 3.817
Stream size : 69,2 Mb (100%)

Camera stops (not freezing) in 5-10 sec. (card - SanDisc Extreme III 30Mb/s 32Gb)

AVCHD shot like this:
(patches not enabled, pal>ntsc conversion only)
http://rapidshare.com/files/394708076/00000.MTS.html
File size : 26,6 МиБ
Duration : 14 с.
Overall bit rate : 15,4 Mbit/sec
Maximum Overall bit rate : 18,0 Mbit/sec

sirk
06-03-2010, 12:27 AM
Sirk, Was this your settings?

[ Low light test MJPEG vs AVCHD
MJPEG:
Quality settings (E1 to E4) - 352, 220, 200, 184
Table settings (E1 to E4) - 24, 24, 24, 24

with this settings top bitrate is up to 55Mbps. ]


The settings I'm using are topping at 106bps. Way to much for even the sandisk. I would like it to top at 70mbs

Yes, those are my settings. I won't want to go up to 106 unless the 2GB limit is removed. Your files would be limited to 2 minutes or something less.
I maybe will also settle to 70Mbps, but for now 55MBps looks sooo sweet.

I guess we are collecting many settings with many different detail levels. Maybe we could implement something like a user interface to chose the Quality tables.

I mean, tester13, maybe, you could add a patch to remove those MJPEG Low quality modes in XVGA or whatever they are. At least for me, they are kinda useless. Instead we could put some rock ass HD Modes in them.

NURBS
06-03-2010, 12:39 AM
Hey Guys'

If you turn on MJPEG 30p -> 24p pach you can record for this settings. ;-) (80 Mbit, no stoping and freezing.)

Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24

Interesting thing -> 720p25 MJPEG mode is posible? (no drob frame).

Regards
NURBS

PappasArts
06-03-2010, 01:31 AM
Hey Guys'

If you turn on MJPEG 30p -> 24p pach you can record for this settings. ;-) (80 Mbit, no stoping and freezing.)

Quality settings (E1 to E4) - 720, 512, 428, 400
Table settings (E1 to E4) - 24, 24, 24, 24

Interesting thing -> 720p25 MJPEG mode is posible? (no drob frame).

Regards
NURBS


The 24P patch- Doesn't that have some duplicate frame issue? Or can this duplicate frame be removed with compressor etc.

Also, has anyone come up with a 70mbit top bitrate 30P?

Michael Pappas
http://www.pbase.com/Arrfilms
Anamorphic GH1 Frames
http://www.dvxuser.com/V6/showpost.php?p=2006938&postcount=1956

Vitaliy Kiselev
06-03-2010, 01:36 AM
Interesting thing -> 720p25 MJPEG mode is posible? (no drob frame).


I don't know.
As I already said we can tune encoder settings, but I am still not sure how buffer things work etc.
Plus we have sensor settings.
Currently I don't know how sensor seting is set from MJPEG settings.
This is due to GH1 firmware implementetion, as they love to use codes, translate one codes to other, copy structures or pointers to structures, etc.

PappasArts
06-03-2010, 01:41 AM
Yes, those are my settings. I won't want to go up to 106 unless the 2GB limit is removed. Your files would be limited to 2 minutes or something less.
I maybe will also settle to 70Mbps, but for now 55MBps looks sooo sweet.

I guess we are collecting many settings with many different detail levels. Maybe we could implement something like a user interface to chose the Quality tables.

I mean, tester13, maybe, you could add a patch to remove those MJPEG Low quality modes in XVGA or whatever they are. At least for me, they are kinda useless. Instead we could put some rock ass HD Modes in them.

The sandisks cant' do 106 without getting the data rate card error/stop recording. That's what I have been experiencing. We need cards that are a solid 15MB sustained write data with a headroom above that to insure 100+ mbit. Class 10 Extremes seem to be the most reliable if the data rate doesn't exceed 70mbit or 10MB. The cards are the limiting factor.

I need to try your 55MBit code. If there is a code that tops out at 70, that would be nice too!


I can't help think how the Zacuto test for the GH1 would have been much different with this superior codec performance were getting been available months ago. The GH1 is getting better; how amazing is that.
.

Isaac_Brody
06-03-2010, 01:45 AM
It's not even close when comparing the mjpeg to the avchd. Mjpeg is wiping the floor with the avchd.

NURBS, do you have a 25P mjpeg with movement so we can look for dropped frames. 25P is so close. :)

Vitaliy Kiselev
06-03-2010, 01:57 AM
It's not even close when comparing the mjpeg to the avchd. Mjpeg is wiping the floor with the avchd.

Generally even intraframe AVCHD will outperform MJPEG.


NURBS, do you have a 25P mjpeg with movement so we can look for dropped frames. 25P is so close. :)

No, he don't have it.
May be 1280p24 (and p25) will be possible, but may be not.
As we have sensor setup for both 1920 modes.
Plus we have scaler (as I don't see any 848 and lower modes).
But many things seems to be hardcoded to 848 and 1280, so it can be just waste of time.

bumkicho
06-03-2010, 02:13 AM
Can 2G size limit for MPEG be removed? I think someone asked this before but no answer yet.

Tester13, I think I read some pages ago that you were wanting to translate some japanese in firmware. I can ask my Japanese coworker to see if she could help. (* She is not a technical person, so her translation may not be in technical terms either) Would this help?

Vitaliy Kiselev
06-03-2010, 02:19 AM
Can 2G size limit for MPEG be removed? I think someone asked this before but no answer yet.

No, if you mean firmware hack, and quite easy by normal firmware modification.
This limit is not MJPEG related it is FAT file system related.
Due to weird licensing issues flash cards still use FAT system instead of normal ones.


Tester13, I think I read some pages ago that you were wanting to translate some japanese in firmware. I can ask my Japanese coworker to see if she could help. (* She is not a technical person, so her translation may not be in technical terms either) Would this help?

Better ask her to look for Panasonic contacts and make few calls.
As we really need some contacts, and under this I do not mean sales and marketing.

bumkicho
06-03-2010, 02:34 AM
Due to weird licensing issues flash cards still use FAT system instead of normal ones.

Tester13, do you by any chance know if GH1 will accept NTFS formatted SDHC card?

Well, I just tried it with no luck. GH1 displays card error and ask if I like to format it.

Chris Light
06-03-2010, 02:50 AM
There's alot of constant rebranding going on it, for instance their SanDisk VideoHD cards used to be branded Class 6 then they dropped it to Class 4. In fact many websites still have the picture of the old packaging, such as Newegg here http://www.newegg.com/Product/Product.aspx?Item=N82E16820171384&cm_re=VideoHD-_-20-171-384-_-Product

... but you'll get the Class4 marked cards in the mail.


wow, i bought this card in-store at Radio Shack http://www.radioshack.com/product/index.jsp?productId=3156591 6 months ago, labeled as a class 6 for $24.99. why is it now labeled as Class 4 for $54.99? that's a rip!

sorry...that was completely off topic.

Cavemandude
06-03-2010, 02:58 AM
Why is the FAT file system an issue with MJPEG and not AVCHD videos?

I thought the GH1 could record long videos and would create multiple 2GB files to achieve this. So for example, using a 16GB card you can shoot a continuous 2-hour video with no gaps or frames missing between each 2GB file it creates.

Randy

PappasArts
06-03-2010, 03:08 AM
It's not even close when comparing the mjpeg to the avchd. Mjpeg is wiping the floor with the avchd.

NURBS, do you have a 25P mjpeg with movement so we can look for dropped frames. 25P is so close. :)


Issac, did you post some time ago that you processed the Mjpeg file in a higher bit file format in AE etc, and got much more image data out of it; Or is that my imagination?

Vitaliy Kiselev
06-03-2010, 03:11 AM
Why is the FAT file system an issue with MJPEG and not AVCHD videos?

Try to read you GH1 manual.
MJPEG always had 2GB limit (and this is same of every camera).
2GB simit is because of file size limit on FAT formatted drives (and AVCHD also splits files).
No spanning in MOV is just because simple logic - no standard player will play such files without gaps or stops. And film makers with NLE knowledge are not primary buyers. So by introducing spanning you'll have one big problem with consumer complains and returns.
AVCHD can be played without problems by many players, including hardware ones without any gaps.

Cavemandude
06-03-2010, 03:32 AM
Ok, got it and thus why Canon DSLR's have a 12 minute limit on their MOV videos, correct? And if Canon/Magic Lantern couldn't come up with a solution/firmware upgrade to get beyond this limit with MOV files does it look unlikely to solve this with the GH1?

Randy

Cavemandude
06-03-2010, 03:41 AM
Little test with the NURBS settings:
(Long focus and camera fast motion)
http://rapidshare.com/files/394703133/P1140391.MOV.html
Duration : 5 s 500 ms.
Bit rate : 106 Mbit/sec
Bits/(Pixel*Frame) : 3.817
Stream size : 69,2 Mb (100%)

Camera stops (not freezing) in 5-10 sec. (card - SanDisc Extreme III 30Mb/s 32Gb)

AVCHD shot like this:
(patches not enabled, pal>ntsc conversion only)
http://rapidshare.com/files/394708076/00000.MTS.html
File size : 26,6 МиБ
Duration : 14 с.
Overall bit rate : 15,4 Mbit/sec
Maximum Overall bit rate : 18,0 Mbit/sec

I attached a frame grab from the MJPEG and AVCHD clip while the GH1 was panning across the grass field in each video. I'm sure you will be able to tell which is which.

Got mud?

Randy

PappasArts
06-03-2010, 03:50 AM
I attached a frame grab from the MJPEG and AVCHD clip while the GH1 was panning across the grass field in each video. I'm sure you will be able to tell which is which.

Got mud?

Randy


Just looking at those two in comparison is frightening.....

Vitaliy Kiselev
06-03-2010, 03:52 AM
And if Canon/Magic Lantern couldn't come up with a solution/firmware upgrade to get beyond this limit with MOV files does this look unlikely to solve this with the GH1?

Just forget about this.
Gapless MJPEG is easy to do, but this requires firmware source.
In this case it can be done in a day or two.

In reality, as I already said, Panasonic must quickly came up with new GH1 firmware. And include special film-maker mode that allow to shoot at very high bitrates and shoot at 1920p24 and 1280p24 gapless MJPEG at any rate from 25 to 80Mbits. I don't talk about Canon-like release, no.
I talk about quick, hard, with large marketing budget.
Yes, they have upcoming new cameras, but this won't make GH1 competitive.
As casual people don't want 70Mbit MJPEG or very high bitrate AVCHD.

sirk
06-03-2010, 04:17 AM
AFAIK FAT File size limit is 4GB - Why must the MOV clips be 2 GB sized?

And to the guy who asked for NTFS support: HAHAAHAHAHAHAHAAHAHAHAHAHAHAHAH (sorry, but man....)

Vitaliy Kiselev
06-03-2010, 04:19 AM
AFAIK FAT File size limit is 4GB - Why must the MOV clips be 2 GB sized?

Use Google, please.
http://www.google.com/search?q=2GB%20FAT%20limit&ie=UTF8&oe=UTF8

sirk
06-03-2010, 04:24 AM
del

sirk
06-03-2010, 04:25 AM
tester13: what do we test today?

Vitaliy Kiselev
06-03-2010, 04:36 AM
tester13: what do we test today?

May be our brain power? :-)

As for 4Gb and 2Gb limit.
2Gb limit is remains of FAT16 system.
And as GH1 code have some parts that are very ancient, it still have 2Gb limit even for FAT32. Just for safety (as anyone can insert FAT16 card).

Cavemandude
06-03-2010, 04:42 AM
I attached a camera comparison chart to put the bitrates being achieved with MJPEG here into perspective.

Randy

spamrakuen
06-03-2010, 04:42 AM
In reality, as I already said, Panasonic must quickly came up with new GH1 firmware. And include special film-maker mode that allow to shoot at very high bitrates and shoot at 1920p24 and 1280p24 gapless MJPEG at any rate from 25 to 80Mbits. I don't talk about Canon-like release, no.
I talk about quick, hard, with large marketing budget.


Hope Pana will listen us!



Yes, they have upcoming new cameras, but this won't make GH1 competitive.
As casual people don't want 70Mbit MJPEG or very high bitrate AVCHD.

Sure. That's why I fear that the GH2 will have wifi, GPS, touch screen... but no high bitrate video mode. 24mpbs AVCHD only, I guess.

About MJPEG and framerates; you told us that 24 o 25 fps seems right now a bit complicated, but what about 50fps? Could that framerate be easier to do?

Thank you for your work

Vitaliy Kiselev
06-03-2010, 04:44 AM
I attached a camera comparison chart to put the bitrates being achieved with MJPEG here into perspective.


Never compare bitrates of different codecs or same codec from different manufacturers or even same codec from same manufacturer but with different implementations details.

Paul Shields
06-03-2010, 04:49 AM
I attached a frame grab from the MJPEG and AVCHD clip while the GH1 was panning across the grass field in each video. I'm sure you will be able to tell which is which.

Got mud?

Randy

Bloody hell !

Essami
06-03-2010, 05:49 AM
Hey,

Is it possible to create new picture profile settings allowing further tweaks on them? Mainly Im interested in lowering contrast and saturation as the -2 settings seem to be not enough and maybe this would give us a little bit more latitude.

Sami

Adventsam
06-03-2010, 06:13 AM
May be our brain power? :-)

As for 4Gb and 2Gb limit.
2Gb limit is remains of FAT16 system.
And as GH1 code have some parts that are very ancient, it still have 2Gb limit even for FAT32. Just for safety (as anyone can insert FAT16 card).

Any plans by Pana to support SDXC, exFAT 64ZB? at 100Mb/sec should store about 100hours! LOL

New G10 and G2 are exFAT compatiable.

http://panasonic.net/avc/sdcard/information/SDXC.html

See the G2 capability here

http://www.imaging-resource.com/PRODS/DMCG2/DMCG2VIDEO.HTM

bumkicho
06-03-2010, 06:39 AM
And to the guy who asked for NTFS support: HAHAAHAHAHAHAHAAHAHAHAHAHAHAHAH (sorry, but man....)

I was wondering if GH1 would accept NTFS formatted SDHC card, so I tried it. And it didn't work. No one asked for NTFS support.

alignment1
06-03-2010, 06:52 AM
Hey,

Is it possible to create new picture profile settings allowing further tweaks on them? Mainly Im interested in lowering contrast and saturation as the -2 settings seem to be not enough and maybe this would give us a little bit more latitude.

Sami

+1 Good thought!

abasfly
06-03-2010, 07:49 AM
Hey,

Is it possible to create new picture profile settings allowing further tweaks on them? Mainly Im interested in lowering contrast and saturation as the -2 settings seem to be not enough and maybe this would give us a little bit more latitude.

Sami


+1 on this.
I find it very important to lower the sharpness and contras and sat say to -4

GrgurMG
06-03-2010, 08:55 AM
uh why?

To give a proper answer... I think tester13 means that due to each camera being a different codec (closest being both Sony's since HDV is flavor of mpeg2), it's hard to extrapolate how certain specs pertain to actual shooting quality... and thus being of limited relativity. There's alot of factors that go into the nuances of encoder settings.. particularly if there's any variable bitrate or variations in framerate involved. For instance, H.264 at 42mbps at 1080_30p should by all accounts blow away our best 70mbps MJPEG results in all situations except maybe when there's crazy motion... but even then that may not always be the case depending on how well or how crappy that H.264 realtime encoding process is. Panasonics TM700 has 28mbps 1080p -60P!- footage that looks amazing in terms of compression... motion seems pretty good from what I've seen too.

I believe, if I'm not mistaken, Tester13 has infered the GH1s AVCHD issues go a bit beyond a lack of bitrate (although more bitrate certainly helps).. there's GOP factors amongst other things. For instance, if the TM700 gets great 1080p_60p footage at 28mbps... it's reasonable to think that if the GH1 had the exact same encoding implementation that it should have decent 1080p_24p since it uses more then half that bitrate and less then half the framerate... but apparently the implementation is different and this is not that case. Then again you also have to factor in that 24p should have more changes frame to frame then 60p and when your dealing with interframe compression that's also a big factor.

Also, theoretically things like banding and noise level differences could have an impact on the overall quality of encodes on different cameras, since cleaner footage has less uneccesary detail thus it should compress/encode better. I've found alot (though perhaps not all) of recent Panasonic models to be consistently cleaner then many Canons in terms of video noise... particularly at low lux.

It's reasonable to believe the GH2, even if it doesn't have higher bitrates (which it probably will), will still have better encoding algorithms for AVCHD... similar to other Panasonic cameras like the TM700 (which is a consumer camcorder btw).

Isaac_Brody
06-03-2010, 09:02 AM
Issac, did you post some time ago that you processed the Mjpeg file in a higher bit file format in AE etc, and got much more image data out of it; Or is that my imagination?

I was processing footage in AE in a higher colorspace. I didn't try it with mjpeg, but I did do it with avchd files and it helped to counteract some of the banding in gradients. I haven't tried it with mjpeg yet.