PDA

View Full Version : Mjpeg vs avchd codec tests- the good, the bad and ugly



PappasArts
06-24-2010, 04:16 AM
For Updates: MIKOS ARTS on Twitter
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://MIKOSarts.wordpress.com/page-2-of-blog
http://mikosarts2.wordpress.com/blog-page-2
http://mikosarts2.wordpress.com
http://MIKOSarts.COM
http://MIKOSarts.net
http://MIKOS.info


Been doing some codec comparison tests with MJPEG 70+ Mbits 1080p and AVCHD 1080P.
Both codecs are amazing and have their rightful place for certain shots and projects.
However- MJPEG is a better codec, and if it can get tuned near perfect,
that would be awesome for certain projects.

The posterization tests below ( AKA Banding )- The GH1 smooths out the AVCHD recording somewhere in the
chain, and that make's it a breading ground for that artifact.

STANDARD
-2 contrast
-2 detail
color 0
NR- -2

http://i33.photobucket.com/albums/d94/PappasArts/codectestv2.jpg


In my Posterization tests, I have concluded that Nostalgic for some
reason produces less of this artifact. Trust me I prefer to use standard
setting, however nostalgic appears to be doing something to the signal.

- 2 on all settings for the posterization test.
3200K
35mm Nikkor at f2.8

http://i33.photobucket.com/albums/d94/PappasArts/POSTERIZATION.jpg


Here's 24PN vs 24P over 60i
Shot in "C" settings
http://i33.photobucket.com/albums/d94/PappasArts/24PNVS60i.jpg


MIKOS ARTS
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://MIKOSarts.wordpress.com/page-2-of-blog
http://mikosarts2.wordpress.com/blog-page-2
http://mikosarts2.wordpress.com
http://MIKOSarts.com
http://MIKOSart.net
http://MIKOS.info


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

Ian-T
06-24-2010, 05:39 AM
The MJPEG looks almost like a wavelet codec compared to the AVCHD. It’s much cleaner. Also, my problem with the GH-1 was always what seems like in-cam over sharpening or rough edges. The MJPEG does not display this…at least to my eyes. I said it a year ago….but the MJPEG looks better than the AVCHD but folks kept saying that it is noisier and less efficient. Some of us thought that higher bit rates on MJPEG would solve that…and it seems that thought was correct. Thanks for those shots.

Is the MJPEG still 30p only or is there a “true” 24p patch yet?

PDR
06-24-2010, 06:20 AM
what software are you using to decode the videos for the screenshots? (some of them, like premiere/ae, have a bug decoding native avchd and h.264 chroma interpretation)

PappasArts
06-24-2010, 12:31 PM
what software are you using to decode the videos for the screenshots? (some of them, like premiere/ae, have a bug decoding native avchd and h.264 chroma interpretation)

Cineform NeoScene for conversion- which is one of the best. Then taken in to After Effects so there is no AVCHD anymore.

MIKOS ARTS
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://MIKOSarts.wordpress.com/page-2-of-blog
http://mikosarts2.wordpress.com/blog-page-2
http://mikosarts2.wordpress.com
http://MIKOSarts.com
http://MIKOSarts.net
http://MIKOS.info


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

adammada
06-24-2010, 12:39 PM
thanks, this corresponds with my tests: MJpeg is the No. 1

PappasArts
06-24-2010, 01:02 PM
The MJPEG looks almost like a wavelet codec compared to the AVCHD. It’s much cleaner. Also, my problem with the GH-1 was always what seems like in-cam over sharpening or rough edges. The MJPEG does not display this…at least to my eyes. I said it a year ago….but the MJPEG looks better than the AVCHD but folks kept saying that it is noisier and less efficient. Some of us thought that higher bit rates on MJPEG would solve that…and it seems that thought was correct. Thanks for those shots.

Is the MJPEG still 30p only or is there a “true” 24p patch yet?

Hey Ian-T,

I've liked MJPEG way back too; however once the codec got super charged, it's performance is exceptional. It does look like a wavelet codec compared to AVCHD. The 720 Hi-bit rate has even more bits thrown at it since it's a smaller frame.

There is more room to tweak the MJPEG codec since we still have never had it explained to us; the relationship between the 4 Quality settings and what the individual numbers do; the Table settings- and their relations with all one another.

Knowing that- would help tremendously in fine tuning it.

Someone mentioned that MJPEG was noisier in their city low light tests. Well that is the fact that MJPEG unlike AVCHD is not having some form of NR going on in the signal path IMO. The MJPEG seems to be straight forward- and doesn't apply such processes IMO- which is good. So if your going to shoot above 800iso in MJPEG, don't be surprised that there's going to be more noise ( anymore then shooting +18db on a camcorder ). The 800iso is over a 1100 ( from the dxomark.com ) lab tests.

I wish more people would try to figure out the formulas, instead of waiting on the sidelines for it to come to them from the few others trying to perfect it. The more people that work on this, the more ground we can cover and improve it even more possibly.

MIKOS ARTS
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://MIKOSarts.wordpress.com/page-2-of-blog
http://mikosarts2.wordpress.com/blog-page-2
http://mikosarts2.wordpress.com
http://MIKOSarts.COM
http://MIKOS.info


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

Ian-T
06-24-2010, 01:34 PM
I wish more people would try to figure out the formulas, instead of waiting on the sidelines for it to come to them from the few others trying to perfect it. The more people that work on this, the more ground we can cover and improve it even more possibly.

PappasWell, I should be part of that group of testers starting tomorrow when I finally receive my GH-1. :happy:

But once Mjpeg becomes fully stable on this cam then I won't shoot without it. The image looks much smoother and less harsh on the edges. I always thought that this codec would shine if a higher data rate was thrown at it (isn't that what the Pentax cameras are doing?). Much easier to edit also.

PappasArts
06-24-2010, 04:21 PM
Well, I should be part of that group of testers starting tomorrow when I finally receive my GH-1. :happy:

But once Mjpeg becomes fully stable on this cam then I won't shoot without it. The image looks much smoother and less harsh on the edges. I always thought that this codec would shine if a higher data rate was thrown at it (isn't that what the Pentax cameras are doing?). Much easier to edit also.


Some of those artifacts are caused by it not being 24PN and rather 24 over 60i. Barry can confirm this. However as I understand it; 24PN gets more bits thrown at it since there's 36 less frames to compress, and the 60i process takes it's own whack at it as well.

I'm shooting some 24PN c settings right now- for comparison.

MIKOS ARTS
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://MIKOSarts.wordpress.com/page-2-of-blog
http://mikosarts2.wordpress.com/blog-page-2
http://mikosarts2.wordpress.com
http://MIKOSarts.COM
http://MIKOS.info


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

wturber
06-24-2010, 05:32 PM
Well, I should be part of that group of testers starting tomorrow when I finally receive my GH-1. :happy:

But once Mjpeg becomes fully stable on this cam then I won't shoot without it. The image looks much smoother and less harsh on the edges. I always thought that this codec would shine if a higher data rate was thrown at it (isn't that what the Pentax cameras are doing?). Much easier to edit also.

The problem with shooting with MJPEG that has much more than a 38Mbs data rate is that you can't review it in-camera. I hope to do some good objective testing this weekend, but it also seems to me that 1080p24 AVCHD holds more detail than the larger MJPEG options.

So in the end, the key is going to be to understand the limitations of the two options and to employ each accordingly. I can easily see myself carrying around a small collection of small capacity SD cards - each carrying its own unique patch combination.

PappasArts
06-24-2010, 05:33 PM
Just added blow up photos above of 24P NATIVE vs 24P over 60i Shot with "C" settings.

MIKOS ARTS
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://MIKOSarts.wordpress.com/page-2-of-blog
http://mikosarts2.wordpress.com/blog-page-2
http://mikosarts2.wordpress.com
http://MIKOSarts.COM
http://MIKOS.info


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

John Caballero
06-24-2010, 07:17 PM
I would be happy with a stable mjpeg 720p24 at some healthy mbits.

greymog
06-24-2010, 07:25 PM
I would be happy with a stable mjpeg 720p24 at some healthy mbits.

+1 and 25p.

PappasArts
06-24-2010, 08:24 PM
I would be happy with a stable mjpeg 720p24 at some healthy mbits.


That would be awesome to have 24P as well for Mjpeg! I wish we could have a better understanding on what the tables and Quality settings do, and how they work together..

For Updates: MIKOS • MLPappas on Twitter
http://PAPPASARTS.WORDPRESS.COM
http://TWITTER.COM/PAPPASARTS
http://MIKOSarts.wordpress.com
http://mikosarts2.wordpress.com/blog-page-2
http://mikosarts2.wordpress.com
http://MIKOSarts.COM

Deere
06-24-2010, 08:54 PM
+1 and 25p.

+1 (including the 25p :D)

zcream
06-25-2010, 05:55 AM
Hi Pappas. I posted it before to spark a discussion but nothing happened. Here it is again. You can notice AVCHD has more processing than MJPEG. So far your tests indicates more NR in AVCHD.
Your MJPEG vs AVCHD snaps showed a softer MJPEG image.
Why are we not considering that the MJPEG is closer to the unprocessed raw sensor signal and the sharpness in AVCHD is simply extra processing ?
I didnt see anything in your tests that could not be explained by a sharpen filter applied to MJPEG.

So, I am suggesting that AVCHD = Sensor + NR + Sharpen
while
MJPEG = Sensor signal.

Another reason for this is that upscaling does take a fair bit of time. And there is no reason for the sensor to upscale in 1080p MJPEG mode. Since its firmware, its a software RTOS implementation.
If there would be an upscale it would be for both AVCHD and MJPEG. Just an upscale for MJPEG makes absolutely no sense.
The other alternative where MJPEG is less processed does make more sense as Panny never intended MJPEG to be used at 1080p.

Ian-T
06-25-2010, 06:47 AM
Just added blow up photos above of 24P NATIVE vs 24P over 60i Shot with "C" settings.

Pappas
http://www.pbase.com/Arrfilms
Wow...look how much noisier the 60i is to the native 24p shot.

sammysammy
06-25-2010, 08:16 AM
After testing the hacks for about 2 weeks now, to me the mjpeg with 422 overall is much more "smother" to the eye and easier to grade..now not saying that the avchd has not improved, it has, but comparing the 2 mjpeg to me is more pleasing to the eye(much easier on my i7 with sony vegas to edit than the avchd)..but the issue still is that we dont have 24p for mjpeg, and i just love the film feel the 24p avchd gives, even after converting 30p to 24p (it just does not look as good as the real 24p avchd)..so for me, im shooting a music video next week, and im playing all this in my head, and still not sure what to use, its killing me..like everyone else if we had 24p mjpeg the i wont touch the avchd ..

Nitsuj
06-25-2010, 10:29 AM
After testing the hacks for about 2 weeks now, to me the mjpeg with 422 overall is much more "smother" to the eye and easier to grade..now not saying that the avchd has not improved, it has, but comparing the 2 mjpeg to me is more pleasing to the eye(much easier on my i7 with sony vegas to edit than the avchd)..but the issue still is that we dont have 24p for mjpeg, and i just love the film feel the 24p avchd gives, even after converting 30p to 24p (it just does not look as good as the real 24p avchd)..so for me, im shooting a music video next week, and im playing all this in my head, and still not sure what to use, its killing me..like everyone else if we had 24p mjpeg the i wont touch the avchd ..

Why not shoot them both? Take 1 - AVCHD - Take 2 - MJPEG that is if you got the time and it doesn't cost extra.

PhilJackson
06-25-2010, 11:40 AM
After testing the hacks for about 2 weeks now, to me the mjpeg with 422 overall is much more "smother" to the eye and easier to grade..now not saying that the avchd has not improved, it has, but comparing the 2 mjpeg to me is more pleasing to the eye(much easier on my i7 with sony vegas to edit than the avchd)..but the issue still is that we dont have 24p for mjpeg, and i just love the film feel the 24p avchd gives, even after converting 30p to 24p (it just does not look as good as the real 24p avchd)..so for me, im shooting a music video next week, and im playing all this in my head, and still not sure what to use, its killing me..like everyone else if we had 24p mjpeg the i wont touch the avchd ..

I'd go with MJPEG if you love the color you get with the new color space. To the uninformed the 24p won't be as big a difference I think than the better colors. But it also depends on the video. Is it going to be color intensive? Are you going to be doing any keying? Or is it more a short film that happens to make a music video? If that was the case I can see the 24p winning out.

sammysammy
06-25-2010, 01:44 PM
Nitsuj i wish i can shoot it twice, but i rather not..phil,i may be leaning toward avchd for the 24p reason ...

sammysammy
06-28-2010, 08:43 AM
By the way guys..any known progress with regarding 24p in mjpeg? i know Vitaliy is working hard, but im not sure if 24p mjpeg is still possible, thanks for any new feedback..

Svart
06-28-2010, 09:24 AM
was the GOP=6 shot with 24p/60i or native 24p? I'd like to see what a low GOP in 24p/60i looks like compared to these tests. Also I would like to see GOP=12 for 24p/60i since that is what is suspected to happen in Native 24p anyway.

PappasArts
06-28-2010, 01:35 PM
Wow...look how much noisier the 60i is to the native 24p shot.

Hey Ian-T,

Interlace yuck- I gather...




Hi Pappas. I posted it before to spark a discussion but nothing happened. Here it is again. You can notice AVCHD has more processing than MJPEG. So far your tests indicates more NR in AVCHD.
Your MJPEG vs AVCHD snaps showed a softer MJPEG image.
Why are we not considering that the MJPEG is closer to the unprocessed raw sensor signal and the sharpness in AVCHD is simply extra processing ?
I didnt see anything in your tests that could not be explained by a sharpen filter applied to MJPEG.

So, I am suggesting that AVCHD = Sensor + NR + Sharpen
while
MJPEG = Sensor signal.

Another reason for this is that upscaling does take a fair bit of time. And there is no reason for the sensor to upscale in 1080p MJPEG mode. Since its firmware, its a software RTOS implementation.
If there would be an upscale it would be for both AVCHD and MJPEG. Just an upscale for MJPEG makes absolutely no sense.
The other alternative where MJPEG is less processed does make more sense as Panny never intended MJPEG to be used at 1080p.


Could be. Often straight from sensor looks soft. In photography forums- , it's filled with complaints I have read many times- like " why are my raw images so soft"... So what your saying could be dead on.

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

PappasArts
06-28-2010, 01:38 PM
was the GOP=6 shot with 24p/60i or native 24p? I'd like to see what a low GOP in 24p/60i looks like compared to these tests. Also I would like to see GOP=12 for 24p/60i since that is what is suspected to happen in Native 24p anyway.


Hey Svart,

Low gop 24/60i tears up and it was horrible. I tried it with a motion wheel test on 6 and 8- Not pretty!

I went with the 6-gop 24PN same as JVC did with their HD at 24FPS. Good so far!




By the way guys..any known progress with regarding 24p in mjpeg? i know Vitaliy is working hard, but im not sure if 24p mjpeg is still possible, thanks for any new feedback..

Hey Sammy,

Don't I wish! Nothing yet as I know.

What would be nice if we can get the 24PN worked out in AVCHD; this would allow us to use higher bit rate formulas with less chance of write failures-- This would be awesome..

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

kmoreau893
06-28-2010, 06:01 PM
I'm sorry, I'm trying to like JPEG in 1440x1080 30P but I find it to be occasionally noisy and blocky. And in general I just find it kind of noisy, even in low ISO, like 100. Im using the 420 color space mode because i read in one of the threads 422 wasn't doing anything, maybe this is my mistake? I'm sure it's my fault but with all the hundreds of posts I'm just trying to keep it simple and using the 'stable' settings thread (or what I think they are -- not sure). I'm sure it's my problem but if I'm not 'adding to the noise' of all these posts it would be appreciated tell me what settings I have wrong. Thanks much.

In the attachment, notice the blocking in the man's shirt, this seems to happen occasionally in darker areas, very unacceptable. Here are my settings:

MJPEG size 1280m->1920m checked
MJPEG 1280m width 1440
MJPEG 1280m height 1080
MPEG 1280m 420->422 color sampling not checked
MJPEG E1 Quality 384
MJPEG E1 Table 24
MJPEG E2 Quality 300
MJPEG E2 Table 24
MJPEG E3 Quality 300
MJPEG E3 Table 24
MJPEG E4 Quality 276
MJPEG E4 Table 24
Video Bitrate Adjustment Simplified 50000000
Overall Bitrate Adjustment 52000000
Limiting Bitrate Adjustment 60000000

PappasArts
06-28-2010, 06:20 PM
I'm sorry, I'm trying to like JPEG in 1440x1080 30P

Remember that AVCHD is applying it's own algorithms, and there is some form of smoothing possibly. MJPEG doesn't, it appears.

Also the better the compression ( or the least ) you start to see the sensor noise because it to now is getting compressed and not thrown out from the codec.

Here is an example ( link below ) of how a 35Mbit codec will render a sky and the sensor noise entirely different- however when a 100mb long gop is done at the same sky- look at the sesnor noise and image. That just means your getting more of how the sensor looks and not blurred out.

Take a look- half way down the page: http://www.xdcam-user.com/?page_id=358


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

kmoreau893
06-28-2010, 06:50 PM
It's hard for me to believe the blockiness is lack of smoothing filters. Look at the screenshot, it's horrible. It seems like a bug in the codec or something wrong with my settings.

Pappas, are, are you now trusting / shooting in mjpeg or AVCHD? I'm not going to again until I can resolve these issues. In a lot of posts it's sometimes hard for me to tell if people are talking about MJPEG or AVCHD. The 24pN AVCHD I've shot is pretty stunning, if not overly sharp, but I hear there are ways to help this. Thanks for any advice.

PappasArts
06-28-2010, 07:28 PM
It's hard for me to believe the blockiness is lack of smoothing filters. Look at the screenshot, it's horrible. It seems like a bug in the codec or something wrong with my settings.

Pappas, are, are you now trusting / shooting in mjpeg or AVCHD? I'm not going to again until I can resolve these issues. In a lot of posts it's sometimes hard for me to tell if people are talking about MJPEG or AVCHD. The 24pN AVCHD I've shot is pretty stunning, if not overly sharp, but I hear there are ways to help this. Thanks for any advice.

The Mjpeg has been for me- very good. Your settings are normal. I'm shooting in both- like many others. Yes it's sharp, however it's a very nice image.

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

Svart
06-28-2010, 08:25 PM
Hey Svart,

Low gop 24/60i tears up and it was horrible. I tried it with a motion wheel test on 6 and 8- Not pretty!

I went with the 6-gop 24PN same as JVC did with their HD at 24FPS. Good so far!




Yeah mine tore up with anything less than 10 with 24p/60i.

I'd still like to see the comparison between GOP=12 24p/60i and your native 24p stuff.

PappasArts
06-29-2010, 12:19 AM
Yeah mine tore up with anything less than 10 with 24p/60i.

I'd still like to see the comparison between GOP=12 24p/60i and your native 24p stuff.

Right now I'm playing with 64-72-80/ 15GOP- 24 over 60i.. Quality is insane if I don't get a write error. I'm totally convinced it's the cards not being able to keep up. Where are those class 20's :-/

If someone could rig up a SD dummy card with a cable to a hi end CF card... Mmmm- Kinda like what people were trying to do with P2 way back..


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

Svart
06-29-2010, 10:23 AM
I shot some stuff before work this morning at 56-63-70 with GOP=12 in 24p/60i. I didn't have time to view it though. I'm interested to see what it looks like compared to "C" settings in GOP=12.

I did get the *real* card speed error with highly detailed scenes while panning. Camera did NOT lock up so I know it's the real deal.

Anyway, I pretty much have to be able to view video in camera/hdmi out so I can't run any GOP less than 12 anyway.

As for that, I compared GOP=15 and GOP=12 on my TV last night and I'd swear that some of those artifacts around the leaves and flowers were not present on the GOP=12 setting under 24p/60i.

aczelkri
06-29-2010, 03:34 PM
Nobody here is talking about the inferior audio quality that comes with MJPEG. Its sampling rate (not bitrate!) is only 16000Hz in contrast to the 48000Hz with AVCHD. MJPEG audio is audibly worse. Or has it been solved yet?

If you are planning on flushing the audio down the toilet anyway and use some background music instead, then ok, it makes no difference. But in any other case I don't think that whoever cares so much about pixel-peeping the video image to get the highest possible quality would settle with such a low quality audio. Especially when the device is otherwise capable of recording really nice audio.

PappasArts
06-29-2010, 04:41 PM
I shot some stuff before work this morning at 56-63-70 with GOP=12 in 24p/60i. I didn't have time to view it though. I'm interested to see what it looks like compared to "C" settings in GOP=12.

I did get the *real* card speed error with highly detailed scenes while panning. Camera did NOT lock up so I know it's the real deal.

Anyway, I pretty much have to be able to view video in camera/hdmi out so I can't run any GOP less than 12 anyway.

As for that, I compared GOP=15 and GOP=12 on my TV last night and I'd swear that some of those artifacts around the leaves and flowers were not present on the GOP=12 setting under 24p/60i.


Hey Svart,

64-72-80- gop15 testing at the moment. Nice image.

I tried to others today 74-82-90 - 84-92-100- at GOP 15. Both were muddy as hell. Worse than
the stock codec. Weird as the numbers went up.

If a lower number set is good- then a middle set is better- and a higher set is worse. This must mean there is an optimum set, possibly somewhere in between.

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

cbrandin
06-29-2010, 05:17 PM
Pappas,

You might consider posting this info on the AVCHD research thread. Seems pretty significant to me.

Chris

PappasArts
06-29-2010, 07:42 PM
Ok- tried some new settings at 58-60-68 15GOP - that seem to give me what 24/60i at 64-72-80 15GOP did days ago in testing.

Make sense since there's 36 less frames to compress with 24 over 60fields . So at 24 discrete frames there's an increase of Mbits thrown at it, probably like 8 to 10Mbits more.
Still got to fully test to be sure....

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

Svart
06-29-2010, 09:02 PM
So my 56-63-70 at 12GOP looks fantastic. It just gives me *real* card speed errors(the kind that don't lock the camera up) and I'm seeing around 30-45 mbit on average and I caught a few 103mbit in peak bitrates before the camera pooped out. I believe this is the max of the Patriot LX class 10, somewhere in the low 100mbit range.

It would be interesting to see a matrix of bitrates and images and then GOP and images. I bet we can find an overlapping point of 24p/60i at higher bitrates/lower GOP and lower bitrate native 24p.

My goal is to find a point where people can get settings equivalent to Native 24p without having to use Native 24p. This keeps cheapskates like me from needing to buy high dollar cards until Vitaliy can eventually fix the Native 24p patch(if possible).

Filmkid2003
06-29-2010, 10:19 PM
I think the 24p/60i with high bit rate settings looks better than any of the native 24p footage I've shot with the GH-1 anyway.

Brian@202020
06-29-2010, 10:24 PM
...I believe this is the max of the Patriot LX class 10, somewhere in the low 100mbit range...

My tests indicate that the Patriot 32gig class 10 card can only be stable at around 53-54mbp/s. any spike above that and it fails. Believe me it's very disappointing for a class 10 card.

0100
06-29-2010, 10:26 PM
So my 56-63-70 at 12GOP looks fantastic. It just gives me *real* card speed errors(the kind that don't lock the camera up) and I'm seeing around 30-45 mbit on average and I caught a few 103mbit in peak bitrates before the camera pooped out. I believe this is the max of the Patriot LX class 10, somewhere in the low 100mbit range.

It would be interesting to see a matrix of bitrates and images and then GOP and images. I bet we can find an overlapping point of 24p/60i at higher bitrates/lower GOP and lower bitrate native 24p.

My goal is to find a point where people can get settings equivalent to Native 24p without having to use Native 24p. This keeps cheapskates like me from needing to buy high dollar cards until Vitaliy can eventually fix the Native 24p patch(if possible).

Nice I am going to try 56-63-70 at 12GOP, but this is my first time messing with GOP...

I only see 1080p24 GOP, I am assuming that is where you are entering the 12 GOP. Is this 1080p24 GOP for both native 24p and 24p/60i? My guess is yes but just want to be sure.

Thanks!

Oh and I will be testing with a transcend 8gb class 10 and 6.

Filmkid2003
06-29-2010, 10:33 PM
My tests indicate that the Patriot 32gig class 10 card can only be stable at around 53-54mbp/s. any spike above that and it fails. Believe me it's very disappointing for a class 10 card.

I would have to agree. I have a Patriot LX 16GB Class 10 card and the highest bit rate settings I can get up to without any errors are 50,52,70... and that's without native 24p checked. Any higher than that and I start getting write speed errors.

PappasArts
06-30-2010, 12:47 AM
Gop 15 is just fine- Good enough for convergent designs NanoFlash, good enough for this. Plus if your using 24PN- it automatically is 12GOP. Thing is, when you lower gop- you need to increase the bit rate as well. 15gop is good for 24 over 60i... and 12 is good for 24PN.

This is how I understand it- I could be entirety wrong..

Barry Green could elaborate on this more though...


Tonight I shot anamorphic with 58-60-68 24PN with GOP unselected since it automatically is 12GOP by default. Looks good.

Pappas

adys
06-30-2010, 01:59 AM
In Mpeg 70 setting, panning is very jumpy, while in the same conditions AVCHD C setting with not native is smooth like oil..

Did anyone else notice it?

Sorry if I am asking an old question, very hard to follow all this hack details...

Svart
06-30-2010, 07:30 AM
My tests indicate that the Patriot 32gig class 10 card can only be stable at around 53-54mbp/s. any spike above that and it fails. Believe me it's very disappointing for a class 10 card.

How are you testing? You have to be careful since some of the card readers on the market will actually bottleneck benchmarking programs you run on SDHC cards.

I'm not saying these cards are good at all, but maybe I've been lucky, they seem to work OK for me.

Barry_Green
06-30-2010, 08:06 AM
Thing is, when you lower gop- you need to increase the bit rate as well.
That's exactly correct. The shorter you make the GoP, the higher the bitrate needs to be. The longer you make the GoP, the lower the bitrate can go.

We're talking about a long-GoP codec here, not an intraframe codec. The whole point of a long-GoP codec is to USE the GoP.

A GoP codec gets its efficiency from only encoding the changes between frames. The shorter you make the GoP, the less opportunity it has to do that.

In a long-GoP codec, especially one like this one without B-frames, there are two types of frames: the Intraframe, and the Predicted frames. In each Group of Pictures, there is one Intraframe followed by a series of Predicted frames (I = Intraframe, P = Predicted, and the Group of Pictures looks like this: IPPPPPPPPPPPP). In that group, the I frame might take up half the available bandwidth, and the P frames all combine to make up the second half of the bandwidth. So one frame takes up 50% of the bandwidth, and then 11 frames combine to take up 50%. So what happens if you make a GoP of six instead of 12? Well, you've just doubled the number of I-frames per second (instead of IPPPPPPPPPPPP it's now going to be IPPPPP IPPPPP). And each I-frame takes up the same amount of space, so -- the I-frames alone are going to use up the same amount of bandwidth as the ENTIRE group of 15 frames used to! So where does the bandwidth to encode the 10 "P" frames come from? Yep, you have to *increase* the bitrate, to make room for those frames. Note, you're not increasing quality, you're just trying to accomodate the additional I-frame you've added into the mix.

Experiment with the GoP all you want, we're all interested in hearing your experiences, but don't go into it thinking that there's some absolute correlation that says "smaller GoP = higher quality" because it may not work out that way. You may find that lowering the GoP just unnecessarily inflates the bitrate without delivering an actual increase in visual quality. Then again, because it's still in the testing phase, you may indeed find that lowering (or raising) the GoP results in a small or big jump in quality. Try it, try everything, experiment, let's find out.

sammysammy
06-30-2010, 08:19 AM
im using 58-67-74 gop12 24pn ..with sandisk extreme iii 16gb class 10 ..ok what i have found out with my card is that if i go beyond 58 i will start getting the blocking and artifacts..this has been stable in 1080p mode, in sh mode i got a few card error's but only on very high detail shoots with high jerky motion, the same detailed shot with normal camera movement the card can work with ..so far im happy with the result ..the only thing i can say is that, the panning when shooting 24pn is very weird, kind of choppy and strobe like, i have to be very very careful to get it to look right, compared to mjpeg 30p shooting or even sh mode..anyone notice that?

i have tried many mjpeg settings with my gh1, i got close to 110mb on a few shoots but only for a few seconds.. the most stable setting for me has been 396,338,306,292 all 24 with 422 ,this gave me clips from 60-80's with out any major issue .. im not sure if the gh1 like the gf1 can get into 100's being without many card errors, i tried the format trick and recording after the first error and still got errors with a setting like 400,400,400,400 with all 24 and 422

Brian@202020
06-30-2010, 08:57 AM
How are you testing? You have to be careful since some of the card readers on the market will actually bottleneck benchmarking programs you run on SDHC cards.

I'm not actually benchmarking the card. I'm testing it in a real world situation and seeing which is the largest bitrate I can write on the card before it fails with the hack. I've done quite a few tests and tried nearly every number posted on this forum since day one and well as many I came up with and I get the same results. Any footage that is around 53-54mbp/s or less works fine and any footage that creeps above that fails.

sammysammy
06-30-2010, 08:59 AM
Brian,I also found that 53-54 mbp/s number to be the limit also for my card..

adys
06-30-2010, 09:10 AM
Guys, no one else encounter jumps when panning with this MPEG settings (the in this thread) ?

sammysammy
06-30-2010, 10:07 AM
for me panning in 30p mjpeg is much more smoother than 24pn avchd ..

mimirsan
06-30-2010, 10:35 AM
Guys, no one else encounter jumps when panning with this MPEG settings (the in this thread) ?

Are you playing directly from the card on the computer? I had the same issue. Once I copied files over to HD it was fine.
You could get jerky playback issue on computer due to reading speed...once videos are reedited..compressed etc all should be fine.
Mjpeg is nice and smooth panning for me also.

adys
06-30-2010, 11:01 AM
Well I am using this settings http://www.dvxuser.com/V6/showthread.php?t=213486

With gop unselected and 422 unselected.

Photo with Iso 800, lots of jumps, SanDisk 16 gig class 10.
In camera, can't play at all.
In computer, jump.
Inside avid after imported, jump.

Again, only in panning test, like the one you do for showing rolling shutter effect, a bush full details.

mimirsan
06-30-2010, 12:37 PM
Well I am using this settings http://www.dvxuser.com/V6/showthread.php?t=213486

With gop unselected and 422 unselected.

Photo with Iso 800, lots of jumps, SanDisk 16 gig class 10.
In camera, can't play at all.
In computer, jump.
Inside avid after imported, jump.

Again, only in panning test, like the one you do for showing rolling shutter effect, a bush full details.

wierd...
I have the same "pappa" setting for mjpeg...ptool recommended "c settings"...422 unselected I have it res 1440 x1080
Record on a sandisk 16 gb class 10.
No playback on cam (which is the norm currently for high bitrate mjpeg anyway)
No problem with playback on computer (have a cheap AMD phenom 64)

Not sure what could be causing the jumpy panning.

sammysammy
06-30-2010, 01:59 PM
2 questions for those who may know..

1)do we have any progress regarding the possibility of 24p mjpeg even at 720p mode?

2)has any one found a combination for increasing the overall avchd bit avg on wide open shots (at f 4 and faster) ?even with 58.67,74 combo i still get an avg of 15-23 mb on not so detailed shots or wide open shots..high detailed f8 and above i get anywhere from 35-54 mb
thanks

adys
06-30-2010, 04:06 PM
wierd...
I have the same "pappa" setting for mjpeg...ptool recommended "c settings"...422 unselected I have it res 1440 x1080
Record on a sandisk 16 gb class 10.
No playback on cam (which is the norm currently for high bitrate mjpeg anyway)
No problem with playback on computer (have a cheap AMD phenom 64)

Not sure what could be causing the jumpy panning.

Ohhh I got the problem ... I am with MPEG 24X and not 30...

0100
06-30-2010, 08:01 PM
Nice I am going to try 56-63-70 at 12GOP, but this is my first time messing with GOP...

I only see 1080p24 GOP, I am assuming that is where you are entering the 12 GOP. Is this 1080p24 GOP for both native 24p and 24p/60i? My guess is yes but just want to be sure.

Thanks!

Oh and I will be testing with a transcend 8gb class 10 and 6.

anyone?

notsociable
10-11-2010, 09:41 AM
:thumbup::thumbup::thumbup::thumbup::thumbup:
That's exactly correct. The shorter you make the GoP, the higher the bitrate needs to be. The longer you make the GoP, the lower the bitrate can go.

We're talking about a long-GoP codec here, not an intraframe codec. The whole point of a long-GoP codec is to USE the GoP.

A GoP codec gets its efficiency from only encoding the changes between frames. The shorter you make the GoP, the less opportunity it has to do that.

In a long-GoP codec, especially one like this one without B-frames, there are two types of frames: the Intraframe, and the Predicted frames. In each Group of Pictures, there is one Intraframe followed by a series of Predicted frames (I = Intraframe, P = Predicted, and the Group of Pictures looks like this: IPPPPPPPPPPPP). In that group, the I frame might take up half the available bandwidth, and the P frames all combine to make up the second half of the bandwidth. So one frame takes up 50% of the bandwidth, and then 11 frames combine to take up 50%. So what happens if you make a GoP of six instead of 12? Well, you've just doubled the number of I-frames per second (instead of IPPPPPPPPPPPP it's now going to be IPPPPP IPPPPP). And each I-frame takes up the same amount of space, so -- the I-frames alone are going to use up the same amount of bandwidth as the ENTIRE group of 15 frames used to! So where does the bandwidth to encode the 10 "P" frames come from? Yep, you have to *increase* the bitrate, to make room for those frames. Note, you're not increasing quality, you're just trying to accomodate the additional I-frame you've added into the mix.

Experiment with the GoP all you want, we're all interested in hearing your experiences, but don't go into it thinking that there's some absolute correlation that says "smaller GoP = higher quality" because it may not work out that way. You may find that lowering the GoP just unnecessarily inflates the bitrate without delivering an actual increase in visual quality. Then again, because it's still in the testing phase, you may indeed find that lowering (or raising) the GoP results in a small or big jump in quality. Try it, try everything, experiment, let's find out.