100% Quality Factor MJpeg Settings (very low compression - class 10 required!)

100% agree, when you fail the chart you start thinking these values are not for me, at least i thought like that..but in real world situation like last night ( i started shooting a music video) , with avchd 57-66-74 24pn gop 12 for sure it would have failed the test..but did not fail me once doing green screen, fast pans,low light situation for clips running 5min+ ,not once.. by the way with these settings i was able to get a couple of shoots at iso 1250 that with out the hack would look like crap..

mpgxsvcd said:
Both of your guys work is very much appreciated. There is no question on that. However, I just think the new chart is overkill for the average user. Most people can use the settings you guys have come up with for 100% of their shooting needs.

I don't see why those average users should step their quality down in order to have the ability to shoot a test pattern that they will never shoot again.

This is just my opinion but I think our time would be better spent trying to optimize 1080p MJPEG. So far I haven't seen any settings that are stable for even the star chart. I would also like to see some proof as to whether it is really scaled 720p or not. Just my opinion though.

I have tested the 1080p settings extensively but I still can't find anything stable. Got any ideas?
Click to expand...
 
mpgxsvcd said:
would also like to see some proof as to whether it is really scaled 720p or not. Just my opinion though.
Click to expand...

Y'know I myself am conviced its not just a upscale...if it is...then its a damn better upscaler in cam than what I use in my editing software.


Anyway kinda back on topic...I set off today using the 400 settings on a gf1 using the 1440x1080 (no 4.2.2) res...had 6 write errors out of 20 shots...I was using my verbatim (which hadn't failed me yet) will try my sandisk next time.

Will be upping a vid on vimeo in a bit
 
Last edited:
mimirsan said:
Y'know I myself am conviced its not just a upscale...if it is...then its a damn better upscaler in cam than what I use in my editing software.


Anyway kinda back on topic...I set off today using the 400 settings on a gf1 using the 1440x1080 (no 4.2.2) res...had 6 write errors out of 20 shots...I was using my verbatim (which hadn't failed me yet) will try my sandisk next time.

Will be upping a vid on vimeo in a bit
Click to expand...

I am doing a test now to try to confirm whether upscaling in post processing is better than shooting with 1080p.

Since MJPEG actually stores the first frame as a jpg image I can compare those to see if there is any scaling going on.
 
Here is the chart with my 1440X1080 settings. I stopped the camera- If I can get decent recording time with the new chart and the star chart- I know that it's closer to being stable. I'd prefer it never fail, however the other point is to see how long we can go until the codec muscle tears- Then we know the settings that held the longest before failure, are the most stable. That's the point for the chart too!

Picture58.jpg


Link to 1440x1080 MJPEG 70mbits settings: http://www.dvxuser.com/V6/showthread.php?t=213364

MIKOS • MLPappas
http://TWITTER.COM/PAPPASARTS
http://PAPPASARTS.WORDPRESS.COM
http://MIKOSarts.wordpress.com
http://MIKOSarts.wordpress.com/page-2-of-blog
http://MIKOSarts.COM
 
Last edited:
I find the charts useful for testing purposes. I used your star chart link + real world shooting in order to settle on Q4 for this thread's settings. The charts provide consistency when trying to dial in the numbers. The new chart would definitely result in nearly bulletproof settings. If you really want hardcore failure protection, you'd simply tune Q3 to the star chart, and tune Q4 to the new chart.

So something like... (based on a quick test where I recorded the new chart at 165 for roughly 30s).. granted it could test longer (so the number might be less), but I'm at work.

Q: [300, 250, 188, 165] / T: [4, 4, 4, 4]

I'd venture to say that the above would qualify as the uber fail safe version in a real world shooting. You'd potentially sacrifice some quality on the simpler scenes in return for it being absolutely bulletproof on the complex stuff. I'd would guess in a complex real world setting it would stay at Q3, unless you run across something hardcore... like a massive life size version of pappas new test grid while strolling the streets... hehe.

sticking with the thread's 400 settings tho... So far, it's been very practical for real world shooting even if it's a bit more on the edge.

regarding the 1440x1080 stuff, I'm pretty sure the Q's will have to be shifted down significantly if your using the 400 settings. Definitely have to explore that... :)
 
Last edited:
GF1 test video

http://vimeo.com/13134055
No raw vid sorry...had to compress down for vimeo :-(

after comparison
Im currently now using Pappasarts setting with the gh1 (as it has a slightly softer image giving less moire)
and the this threads 400's for gf1 (because the sensor is smaller it benefits from the sharper image as it was a lot softer in distant objects than the gh1 )
Thanks again guys!!
 
I was testing Pappas chart yesterday with my GH1 and noticed one strange thing. When at AVCHD ('c' setting FHD mode) I positioned the camera facing the LCD monitor with Pappas' chart and press the record button (creative movie mode 'M' ) the camera stops recording after 2 seconds. If you press record button again it will freeze the camera. It does not happen with AVCHD SD thought, only AVCHD FHD .

I just wonder if anybody else experienced the same issue.
 
abolit said:
I was testing Pappas chart yesterday with my GH1 and noticed one strange thing. When at AVCHD ('c' setting FHD mode) I positioned the camera facing the LCD monitor with Pappas' chart and press the record button (creative movie mode 'M' ) the camera stops recording after 2 seconds. If you press record button again it will freeze the camera. It does not happen with AVCHD SD thought, only AVCHD FHD .

I just wonder if anybody else experienced the same issue.
Click to expand...

Have you got 24/24p native ticked? Causes a lot of probs.
 
No, I did not select 24/25p.
What is weird is when I start recording pointing the camera away from the screen and then go back to pappas chart it works fine .

Could anybody of you guys replicate this scenario? (AVCHD FHD, setting 'C', creative movie mode with 'M'. shutter 60)
Open Pappas' chart (the last one, not the star) , focus on it and push record button to see if it starts recording.
In my case it stops right away.

thanks!!!

mimirsan said:
Have you got 24/24p native ticked? Causes a lot of probs.
Click to expand...
 
Last edited:
@abolit - Even tho FHD and SH use the same setting, SH is usually recording at a lower bitrate (for complex scenes)... It's likely that chart it too complex for FHD.

@mimirsan - I really dig those super wide shots.. the details off in the distance. Makes me want a GF1 to tote around 24/7.
 
Last edited:
abolit said:
What is weird is when I start recording pointing the camera away from the screen and then go back to pappas chart it works fine .
Click to expand...

I can confirm this weirdness- I've seen this happen too. Also- cover the lens with your hand etc, and slowly reveal the chart, and it will record for a short while before error.

If recording is started with chart exposed- no covering- error right out the gate. This was in 24PN research though. Still the same thing.

There must be some kind of codec bit rate analyzer that checks the image before recording, just a guess

MIKOS • MLPappas
http://TWITTER.COM/PAPPASARTS
http://PAPPASARTS.WORDPRESS.COM
http://MIKOSarts.wordpress.com
http://MIKOSarts.wordpress.com/page-2-of-blog
http://MIKOSarts.COM
 
Last edited:
Thanks Pappas!
Thought somethings was wrong with my camera.


PappasArts said:
I can confirm this weirdness- I've seen this happen too. Also- cover the lens with your hand etc, and slowly reveal the chart, and it will record for a short while before error.

If recording is started with chart exposed- no covering- error right out the gate. This was in 24PN research though. Still the same thing.

There must be some kind of codec bit rate analyzer that checks the image before recording, just a guess

Pappas
Click to expand...
 
modulr said:
@abolit - Even tho FHD and SH use the same setting, SH is usually recording at a lower bitrate (for complex scenes)... It's likely that chart it too complex for FHD.

@mimirsan - I really dig those super wide shots.. the details off in the distance. Makes me want a GF1 to tote around 24/7.
Click to expand...

Yeah distant detail has vastly improved...no a patch on the gh1s but superb for a camera that you can pick up so cheap now
 
Here is an indoor video I shot of my son climbing last night. I used the settings from the first post and I enabled 4:2:2 color. I think the quality of the footage was outstanding considering how poor the lighting is in there. I had trouble with holding the camera steady on this shot. Sorry about that. There is also a youtube clip below that shows Hacked MJPEG clips and AVC-HD clips together. The 720p AVC-HD is good but it really just can't hold upto the 100 mb/sec MJPEG right now.

http://vimeo.com/13170628

http://www.youtube.com/watch?v=lT1UweeWVSQ
 
So what was your setting?
From what you wrote it's clear that you were using 4:2:2.
How about Q/T settings?

Off topic: Was anybody securing your son climbing that high? I can't see any guys down there.


mpgxsvcd said:
Here is an indoor video I shot of my son climbing last night. I used the settings from the first post and I enabled 4:2:2 color. I think the quality of the footage was outstanding considering how poor the lighting is in there. I had trouble with holding the camera steady on this shot. Sorry about that. There is also a youtube clip below that shows Hacked MJPEG clips and AVC-HD clips together. The 720p AVC-HD is good but it really just can't hold upto the 100 mb/sec MJPEG right now.

http://vimeo.com/13170628

http://www.youtube.com/watch?v=lT1UweeWVSQ
Click to expand...
 
Another thing I've noticed running MJPEG clip on VLC player (mac) with 400,300,250,188 setting .

VLC > Window>Media Info> Statistics> this screen shows dropping frames which was greater than zero. I'm not sure if it means something problematic but I didn't see any dropping frames playing back AVCHD 1080.

Any ideas?
 
abolit said:
Another thing I've noticed running MJPEG clip on VLC player (mac) with 400,300,250,188 setting . Any ideas?
Click to expand...

I can't confirm that. Just tried it with a longer clip and zero lost frames. Current computers lough when asked to playback MJPEG and so should yours. :cheesy:

Thanks for mentioning the VLC player. I just saw that it gives a live display of the bitrate not just the average like the QuickTime player.
 
abolit said:
So what was your setting?
From what you wrote it's clear that you were using 4:2:2.
How about Q/T settings?

Off topic: Was anybody securing your son climbing that high? I can't see any guys down there.
Click to expand...

I used the following MJPEG settings
Q1=400, T1 = 4
Q2=300, T2 = 4
Q3=250, T3 = 4
Q4=188, T3 = 4

I know it looks high but it is actually about 12 feet to the top out. He has an 8 inch thick gymnastics pad under him and a spring loaded floor under that. As long as the climber knows how to fall properly, which my son does, the climber can jump from those heights without injury.

Bouldering like this is actually a lot of fun and can be done safely if properly trained.
 
You must log in or register to reply here.
Back
Top