A few suggestions, for improving video quality

Issue #1030 on hold
Former user created an issue
  1. If we know that, (in video mode), the sensor reads out a native 1620x1080 from the CMOS sensor, ideally we should be writing this stream to SDHC & not the artificially interpolated 1920x1080 file. All up-scaling would be done in software using something like a Lanczos resizing algorithm. You would then crop to a 16:9 aspect ratio.

  2. i-frame only h.264 compression, similar to the low GOP method our GH2 cousins are currently using.

Comments (4)

  1. Former user Account Deleted

    To improve image quality, I tried to record the ML clean output from my 60D HDMI to a nanoflash 220 mbps i-frame codec. This produces a 1920x1080 image with the 1620x1080 inside and lateral black wipes. In the editing soft (FCP) I had to zoom-in 118,5% to get rid of the bars and by the way I lost the upper and lower part of the picture. At the end I recovered the same picture than the one recorded directly by the 60D but I did not noticed any improvement in quality. It vas even worse as the 18,5% zoom make the noise bigger in a range that the benefit of the 220 i-frame codec do not compensate. Any ideas to test a better workflow? Pascal Carcanade

  2. Alex

    It is possible to set the size for I-frames and P-frames (for example, make all frames equal in size). This won't make all of them I-frames, but it should be close.

    Do you have an example where compression artifacts are visible?

  3. Former user Account Deleted

    Wouldn't intra-only lower quality if the bitrate isn't boosted to compensate? If I understand correctly, P-frames (and B-frames, though they aren't available in baseline profile) are used to reconstruct frames from previous and/or future frames to save space. Having intra-only would probably reduce quality at a given bitrate since each frame is a "full" frame. The only practical reason to have it is to allow faster/accurate seeking (no frames have to be reconstructed from the nearest keyframe).

    As for the "native" recording resolution, I'm fascinated to know how that works (though my technical knowledge is severely limited). If the video is truly upscaled 1620x912, bypassing the upscaling process should help reduce file sizes and perhaps it'll give some headroom to boost the bitrate further? Or even more fantastical, one could alter the recording area to full 1920x1080 and overscan the 3:2 frame. Doubtful though, since the recording surface is probably set in stone.

  4. Log in to comment