Personal View site logo
Official Low GOP topic, series 4
  • 586 Replies sorted by
  • @artiswar And the lenses i used were the Lumix 14mm 2.5 and the Sigma 24-70 Canon mount with the kipon adapter.

  • @artiswar Thank you! :) -the color profile was smooth, all -2. -I edited the clips in Premiere Pro CS5 avchd timeline. -Exported to prores422HQ. Put it in FCP, recut the video, and sent it to color then back to FCP it went. -Then compressed to h.264 with compressor. then up to vimeo and youtube.

  • @driftwood : You're probably right, the video bitrate was stable and well above 80Mbps.

  • @stanlymanly2

    That was exceptionally beautiful. You've got some really great grading and compostion there. What were your settings (Color profile, shutter speed, etc.)? What lens(es) did you shoot on? What was your post production process?

    Some really great softening in there, it seems.

  • Hey everybody! Id like to show you what I was able to do with the Aquamotion patch. The driving scenes were shot with 24L toned down to 66mbps. The rest are shot with the 24H 100mbps mode. I know its a little behind but I just wanted you guys to check it out and if possible pass me some tips on how to utilize the hacks better. Now I am trying the Terraquake patch with 24H toned down to 132mbps and 24L down to 100mbps just because my card can handle those bitrates better. Thanks again and I hope that you guys can give me any type of criticism to help me out.

  • @duartix "Average Video Bitrate = -19.813.054"

    • Looks like youve spotted an integer error in streamparser. @cbrandin Chris?
  • @cinepaint, ISO 1250 may be problematic if you don't know about the discussion here:

    http://www.personal-view.com/talks/discussion/1280/gh2-iso-noise-bug-at-320-640-and-1250

    Otherwise, please post a frame and someone may have some feedback that may assist you. Regards, WR

  • I thought some of you would like to know that I tried Quantum v7 at night, with ISO 1250 opened to f2. Everything in the image, including the well lit areas was very noisy. Wondering why that was?

    Thanks.

  • That first test was around 1min ~1GB.

    Now I've tried the same torture on 720p60 and gave up at around 5:30s (because I believed it should have already spanned) but the file ended up short of 3.5GB. Not only that but my shaking abilities have been seriously challenged because the report shows much less bitrate:

    File: A:\PRIVATE\AVCHD\BDMV\STREAM\00002.MTS
    
    Size = 3.553.880.064 Bytes (18.509.792 Packets)
    
    Stream: Max Speed = 85.509.280 bps
    
    Mode = 720/50p
    
    PCR Timing Interval = 0,095 seconds
    
    Clip Time = 0:05:45,72
    
    Average Total Bitrate = 82.237.187
    
    Average Video Bitrate = -19.813.054
    
    Average Audio Bitrate = 235.005
    
    Average Other Bitrate = 2.429.206
    
    Max GOP Size = 721.152 Bytes (5.769.216 bits)
    

    I don't know if this is due to the nature of 720p or if the use of P-frames is helping (still I'm not buying it because they are very marginally smaller than the I-frames)

    Thanks again.

  • Loaded Quantum 100-50 (Documentary).

    Loaded http://www.amazon.co.uk/Samsung-16GB-Class-Extreme-Speed/dp/B00569J5JC into my GH2.

    Loaded this texture in the browser (full screen).

    Went straight to 80%VFR mode.

    Set the 45mmF1.8 lens to F5.6, set ISO to 3200.

    Shaken the hell out of the camera for a little over 1min.

    I don't know if it spans or blends but this is what I got in StreamParser:

    File: A:\PRIVATE\AVCHD\BDMV\STREAM\00001.MTS
    
    Size = 1.081.583.616 Bytes (5.633.248 Packets)
    
    Stream: Max Speed = 106.611.800 bps
    
    Mode = 1080/24pN
    
    PCR Timing Interval = 0,095 seconds
    
    Clip Time = 0:01:27,09
    
    Average Total Bitrate = 99.358.262
    
    Average Video Bitrate = 96.028.338
    
    Average Audio Bitrate = 235.136
    
    Average Other Bitrate = 3.094.787
    
    Max GOP Size = 1.300.032 Bytes (10.400.256 bits)
    
    Video Frame Size (Min/Ave/Max)= 499.008 / 511.046 / 788.160
    

    I don't really know which was the better deal, the camera with kit lens and two batteries for 600€, the 16GB Samsung SDHC card for 18€ or Vitaly+Driftwood for free.

    Enough said. It kicks ass.

    THANK YOU.

    Colour_Texture_06_by_Tackon.jpg
    1920 x 1080 - 4M
  • "I think you can probably shut down lens correction by going to "SHOOT W/O LENS" and setting this to "ON". Panasonic lenses will still work."

    Does nothing to stop these errors. Shoot wo lens doesn't stop any of the panasonic lens's functionality as far as I've experienced. I've always had it set to on. I don't know if it turns on and off lens correction. Or if that has anything whatsoever to do with the errors.

  • Incidentally, the lack of noise with Quantum 9b in 24H mode and high ISOs (1600 to 6400) seems unreal.

  • I think you can probably shut down lens correction by going to "SHOOT W/O LENS" and setting this to "ON". Panasonic lenses will still work.

  • I was about to point that digital lenses might just be sharper and so put more pressure on the codec, but RRRR has made a giant point. The interesting part is that it's putting legacy glass to advantage. :O

    And now I wonder how long before people start asking for an hack to shutdown image correction. :)

  • re: lumix lenses and failure - it may well have to do with lens correction being done by the camera rather than AF, OIS and so on.. Lens correction would by far put the bigger load on the processor imo (than AF, OIS) since it's real-time image processing.

  • @ehr, I agree too, but let's let our poor man driftwood perfect the 24h (whilst he's tinkering with 720 / 1080i as he can) and then I'm sure we'll get our branched -- I think @sage has already created a branch with quantum 7 and I'm sure he'll do so with 9 as well -- actually doing some twixtor tests with his patch tuesday so I'll let you know how it goes.

  • "I also think the final Quantum should have a 720p and a 1080i branch."

    +1. I need me some fine 60p.

  • About quantum 9b, 720p 60 with both 95mb 64gb sandisk and 30mb sec sandisk.

    from my tests with 14-140 and 20mm lumix, Af and ois are not factors in the write fails. With them off, they both still fail it just takes a couple of seconds longer. Even with the camera motionless and no focusing or zooming they fail. Always within about 4- 20 seconds.

    I have an ancient 50mm minolta and I can focus and refocus all day with that and it never fails. Same with voight.

    If I was to guess it has to do with the increased communication between a lumix lens and the camera. Even if af and ois are off, there's probably still more info flying back and forth during the card write process.

  • I was using the Q9 in 720/60P and was good at 50 fps but at any higher FPS it said my card was not fast enough? Though I would get about 7 seconds of pure butter before it stopped. That was using a Canon 50mm 1.8. Don't know if this helps or not.

  • @driftwood et al Panasonic lenses have a very high resolution and wide angel lenses also cause more problems for the codec. Furthermore, those lenses have IS and AF that burden the processors that also process the image data into AVCHD files. I think the most important things is that these things are documented so we can all make our own decisions. People with manual lenses generally have less problems. I also think the final Quantum should have a 720p and a 1080i branch. Most people never use 1080i while @proaudio4 is right, 720p60 is great for slomo!

  • A lot of people are using 1080i and prefer it once deinterlaced properly.

    With Q9 I get incredible results/stability with 720p on the voigtlander then shit on the same shutter / iso / fstop etc... settings with a panny 14-140mm. Frustrating. Equally sdcard to card the results vary. Hence my reliance on other people testing and reporting back. We're defo after a good balance.

  • Nick, If we can't have both at best quality, which I believe this is the case, my vote goes for the best 720 60P!

    Having said that, you could always throw out a poll, but man we need 60P for slomo!

  • @Kolbasa Im d/l your file and will report back.

  • BTW: The big thing left to fully work out is the 'Frame Buffer Size' patches in ptools. If anyone is prepared to alter all up and down by small percentages at a time to see how it effects the settings - just testing those on Quantum - please pm me.

  • Thanks for the reports. Looking into it. Its all about balancing 60i with 720 on the frame limits which is still killing 720.

This topic is closed.
← All Discussions