Native 24P/25p settings/problems discussion

rambooc1 said:
From here Pappas, 1.2 times bigger.
Click to expand...

Thanks rambooc1!


cbrandin said:
I have posted testing results in the AVCHD research thread in the GH13 Development/Testing section that prove that there are no "rogue" frames - it's just a bug in StreamEye.

Therefore, whatever is causing write errors in 24pN mode has nothing to do with any rogue frame bug - the streams look perfectly fine. Time to go back to experimenting with the other settings. There appears to be nothing wrong with the Native 24p/25p setting itself. Native 24p/25p does produce more color data than wrapped FHD mode (per Barry); that probably explains why you can't use as high data rates with it checked. Same thing with SH mode versus FHD mode.

Chris
Click to expand...

Very interesting Chris- Good work!

Pappas
 
svecher said:
I've been testing below settings on GH13 for about a week now and have not had any issues whatsoever, so it's time to share.

Code: 
[Information]
Comment=Working Settings:<br>1) Native, Surfer AVCHD, graduated H/L and Preset @ <br>50Mbit<br>2) modulr's MJPEG settings with 4:2:2 color
SD_Card=Sandisk Extreme Class 10
Camera=GH1 v1.32
[Settings]
Prevent version compare=Checked
720p30 420->422 color sampling=Checked
E1 Quality=188
E1 Table=4
E2 Quality=188
E2 Table=4
E3 Quality=188
E3 Table=4
E4 Quality=185
E4 Table=4
Native 24p/25p=Checked
Video Bitrate FHD/SH=50000000
Video Bitrate H=40000000
Video Bitrate L=32000000
Overall Bitrate=56000000
Limiting Bitrate=60000000
[B]Preset bitrate=500
Preset bitrate 2=500[/B]
Click to expand...

I just downloaded today's version of PTool v3.40d, but I can't find a "Preset bitrate" setting anywhere. Am I missing something? Thanks.
 
emory said:
I just downloaded today's version of PTool v3.40d, but I can't find a "Preset bitrate" setting anywhere. Am I missing something? Thanks.
Click to expand...

It's in ”Patches for testers" folder.
Patches for testers->AVCHD Movie Mode -> AVCHD Compression -> preset bitrate.
 
Thanks, Churasan. It never occurred to me to check the Tester's folder; I always considered it kind of "off limits..."
 
rambooc1 said:
I used to get both card speed AND card write errors in NATIVE mode, i have only ever had card speed errors in 50i, and that's when i shoot the static chart 100%. I still think a class 10 will fix that too.
Click to expand...

PappasArts said:
Now that's interesting, I wonder if that has to do with your settings not being to high...
Click to expand...

Yeah, I don't feel the need to push bitrates any higher - a good class 10 card already enables 3x extra bitrate bandwidth available for complex scenes in FHD mode as compared to stock setup (57 vs 18). That's plenty enough for me :2vrolijk_08:
 
svecher said:
Yeah, I don't feel the need to push bitrates any higher - a good class 10 card already enables 3x extra bitrate bandwidth available for complex scenes in FHD mode as compared to stock setup (57 vs 18). That's plenty enough for me :2vrolijk_08:
Click to expand...

Yeah I tried the settings in 24PN today on some death shots side by side with 24PSF and the 50-56-60 24PN would get write error- Then I loaded up my 4X settings with 24PSF and the same shots would record forever. I can force a write error with 4x 24psf if I try. However its way more stable then 24PN 50-56-60. I hate loosing footage and pulling the battery with 24P native- 24psf 4x is so high of a bit rate that it's hard to complain- I don't expect any camera even the AF100 can be anywhere as clean- that is solely the fact we are shooting in mbits 3X higher than 24mbits avchd. I'm shocked at how clean the images are. It's all gravy no matter how its sliced....

Pappas
 
cbrandin said:
I have posted testing results in the AVCHD research thread in the GH13 Development/Testing section that prove that there are no "rogue" frames - it's just a bug in StreamEye.

Therefore, whatever is causing write errors in 24pN mode has nothing to do with any rogue frame bug - the streams look perfectly fine. Time to go back to experimenting with the other settings. There appears to be nothing wrong with the Native 24p/25p setting itself. Native 24p/25p does produce more color data than wrapped FHD mode (per Barry); that probably explains why you can't use as high data rates with it checked. Same thing with SH mode versus FHD mode.

Chris
Click to expand...

I'm still getting these empty P frames on high static detail shots, the massively large rogue P frames (per streameye) were never a problem wrt IQ, but the blank/repeated P frames that precede a muddy I frame are definitely still there in 25p. I had a few minutes this morning to try out the latest ptool and checked off the latest add-ons and was still able to easily repeat the blank P/muddy I bug. It can still crash the camera though it did seem to go longer this time (about 25 secs). If time allows I'll try adjusting some of these new settings with the latest ptool
 
richtrav said:
I'm still getting these empty P frames on high static detail shots, the massively large rogue P frames (per streameye) were never a problem wrt IQ, but the blank/repeated P frames that precede a muddy I frame are definitely still there in 25p. I had a few minutes this morning to try out the latest ptool and checked off the latest add-ons and was still able to easily repeat the blank P/muddy I bug. It can still crash the camera though it did seem to go longer this time (about 25 secs). If time allows I'll try adjusting some of these new settings with the latest ptool
Click to expand...
Please also post post a screenshot of footage with blank P frames analysed by GH13StreamParser.
 
richtrav said:
I'm still getting these empty P frames on high static detail shots, the massively large rogue P frames (per streameye) were never a problem wrt IQ, but the blank/repeated P frames that precede a muddy I frame are definitely still there in 25p. I had a few minutes this morning to try out the latest ptool and checked off the latest add-ons and was still able to easily repeat the blank P/muddy I bug. It can still crash the camera though it did seem to go longer this time (about 25 secs). If time allows I'll try adjusting some of these new settings with the latest ptool
Click to expand...
Steameye is buggy, there are no rogue P frames.
 
PappasArts said:
I don't expect any camera even the AF100 can be anywhere as clean- that is solely the fact we are shooting in mbits 3X higher than 24mbits avchd. I'm shocked at how clean the images are.
Click to expand...

Well, by "clean" I assume you are referring to the absence or reduction of compression artifacts, correct (?) ...because the pure resolving power (nowhere near an EX-1) has not changed. And I would imagine, if Panny is serious about the AF100, they need to at least match, but preferably beat, the resolving power of the EX-1 - something that seemingly Canon's XF line does, in fact, do in addition to 4:2:2.
 
ajcourtney said:
Well, by "clean" I assume you are referring to the absence or reduction of compression artifacts, correct (?) ...because the pure resolving power (nowhere near an EX-1) has not changed. And I would imagine, if Panny is serious about the AF100, they need to at least match, but preferably beat, the resolving power of the EX-1 - something that seemingly Canon's XF line does, in fact, do in addition to 4:2:2.
Click to expand...


This is something that perplexes me a bit. Why doesn't the GH1 resolve near the limits of the 720p and 1080p format? This is an especially interesting question if, in fact, it does not do any line skipping.

My current best guess is that the GH1 forms its image using these three basic steps.

1) Bilinear or nearest neighbor demosaicing. Super simplistic and super fast. The evidence for this is color aliasing that shows up in video files but is absent in still photos.

2) & 3) A digital low pass filter (blur routine) applied before bilinear or bicubic downsampling. The evidence for this is that the aliasing is less pronounced than if you apply downsampling to full resolution GH1 stills (12MB) without first blurring the image.

Even if a lot of resolution is lost due to the simplistic demosaicing, There should still be way more picture detail than can be held in a 1080p image. So it seems like the problem is the aliasing with the aliasing artifact obliterating/swamping fine detail. If I'm right, then feeding the sensor a softer image to begin with (optical filter or very small aperture) to reduce the aliasing and then increasing sharpening might actually create an image that has more real detail - a "cleaner" image.
 
I just tried your settings for an interview outside in the shade against some blowing cedar branches... it seized up 5 times in an hour, but the footage was unaffected. If I change to lower settings, do I need to adjust the preset bitrates? Footage looks amazing anyhow...
 
Quick question:

Is 24PSf the same as 60i?

In other words, 24p NATIVE (25p in PAL-Land) is simply left UNCHECKED in ptool3?

Or is there more to it than that?
 
These settings proving very robust

These settings proving very robust

Adventsam said:
PAL, 50i, 50 FHD, no H or L ticked, 52, 60 overall and limiting and only preset bitrate ticked, 500 not preset 2. very stable, very good low level.
Click to expand...

Not one lock-up! mind havent tried 50p since I got these so robust. Low detail scenes 20Mb+, High detailed 43+ but generally in upper 30's and looking awesome. Also, the 50i has no RS as far as I can tell it solid-solid-solid and the panning, will upload some panning I did looked like amazing.
 
Adventsam said:
Not one lock-up! mind havent tried 50p since I got these so robust. Low detail scenes 20Mb+, High detailed 43+ but generally in upper 30's and looking awesome. Also, the 50i has no RS as far as I can tell it solid-solid-solid and the panning, will upload some panning I did looked like amazing.
Click to expand...


Yeah i had success with those settings a week ago and thought the low level was good. Might have to revisit.
http://www.dvxuser.com/V6/showpost.php?p=2045075&postcount=24

From that post.........>


The 100% error free settings for me is ...
Version compare checked
Battery mod Checked
25pN unchecked
C settings
FHD/SH 50,000,000
H unchecked
L unchecked
Overall 52,000,000
Limiting Bitrate 60,000,000
Preset Bitrate1 - 234 (default)
Preset Bitrate2 - not ticked
 
rambooc1 said:
The 100% error free settings for me is ...
Version compare checked
Battery mod Checked
25pN unchecked
C settings
FHD/SH 50,000,000
H unchecked
L unchecked
Overall 52,000,000
Limiting Bitrate 60,000,000
Preset Bitrate1 - 234 (default)
Preset Bitrate2 - not ticked
Click to expand...

What card were you using?

Stable in static high detail too?
 
You must log in or register to reply here.
Back
Top