2

I am a software developer. I understand the video concepts, but I am no expert.

I have a Sony Action Cam AS-15. It records ate 1080/60fps/25Kbitrate. The 60FPS videos just dont; play smooth maxemised on my pc, xbox, ps etc.

I am looking for the best way to drop some bitrate, but keep decent video quality.

I have tried numerous converters/settings, but I just cant get a decent quality video going... I can, for instance, in no way, get videos to look as good as the uploaded videos for the Action Cam, on the Sony action cam youtube channel, for instance. Not even to mention nearing in any way the gopro uploaded video quality.

Sony Youtube Video:

My Youtube Videos:

This Last upload was: 13 minutes, 1GB, 720p, 30fps, 9259kbps, MP4 (Windows Movie Maker), Gopro Black, Sony Action Cam, Fuji xp-100 cameras

The little AS-15 is not as good as the gopro, but the "professional" post videos uploaded to Sony's channel, look damn good. I am not talking about saturation, effects etc. I am talking purely about (pixelation/artefacts)

The original 1080/60 videos look pretty good, but I cant keep them looking good in a more usable format... I would like to keep 60fps, but just lower the bitrate.

I have tried:

  • Any Video Converter Software, convert to 30fps/24kbitrate
  • GoPro Studio (Best quality)
  • Keep high original bitrate, converted to 720p
  • .h264
  • .mp4
  • I don't really want to buy Vegas Pro, or Adobe premier, as this is a small hobby.

I see decent quality movie (feature films) encodes, running at 1080p, with only 2K bit-rate...? Which makes for a decent file size of 1.5gb/90 minutes. My encodes are 1GB for 15 minutes, and look just terrible.

What am I missing?

P.S. I know from research, and observation, that the Sony AS-15 totally messes up with video compression. There are quite a lot of artefacts on very detailed scenes... Can this be the problem, that the source just isn't good enough to convert / work with?

If this is the case, are there any software to keep the original file bit-rate and FPS, but just scale down to 720p, in order to keep the initial quality as it was recorded, but make it a bit more usable..

2
  • let us continue this discussion in chat
    – AJ Henderson
    May 14, 2014 at 14:01
  • Thanks for details but still... After a good grading, I still see a loss of global quality after uploading to Youtube. I think it's their compression method that needs some precise curves, colors and constrastm maybe some surface denoising, suiting their need. Inside Youtube Studio, they have tools to enhance ("fill light", contrast, stabilization), but I never tried. > wikihow.com/Enhance-Video-Quality-on-YouTube Google tells us this: - Video codec: H.264 - Progressive scan (no interlacing) - High Profile - 2 consecutive B frames - Closed GOP. GOP of half the frame rate. Apr 11 at 7:15

1 Answer 1

2

Taking a closer look at the footage now that I'm at home, it's mostly the level of activity and motion in the shots. Even in the Sony video, the quality absolutely falls apart any time that there is high motion and unstable video.

The way video compression works, it depends on predictable and smooth motion to achieve good compression results, especially for on the fly recording like a camera has to do since it can't look forward in time and see how things are going to move any further than it can buffer, which is generally not going to be very far (probably sub-second).

There are two places you could get quality loss. The first is in the recording itself. If you aren't recording at sufficient data rates, then if there is too much motion, fine detail is going to get lost. Uncompressed video is absolutely massive and even most video cameras record with moderate to heavy compression already applied. If your source footage lacks detail (and isn't out of focus) then you need to work on stabilizing the camera to ensure it can better compress the video. There are a number of different ways you can shock mount a camera to help provide some stabilization, though they also generally add weight.

If the source footage is ok, then the problem lies in the final compression for web/end-user consumption. The fundamental problem is still the same, but we now have the option of using digital video stabilization to help make movement less random. We also can use VBR 2-pass compression since we can now look ahead and see how thing move in the future as well as the past, which allows for far more accurate compression and more efficient use of bandwidth.

Additionally, some of the footage could use color grading to enhance the contrast. Ideally footage should make full range of luminosity (brightness). Your blacks should be dark black, but without obscuring detail in the shadows. Your whites should be bright, but without fine detail bleeding in to pure white. You can start with a white point and black point, but really, using something like curves is the best way to tweak the video to make the most use of the available range of luminosity. This can also be quickly, but sloppily accomplished by increasing the contrast, but results often look far more artificial that way. Color grading is best learned by doing it and it can vary throughout a clip. It takes time and practice to get good at it.

2
  • Thanks for the detailed reply. I am happy with the source video, and am looking at the best way to get that into a more usable format, looking good, at 60fps. I did a encode using handbrake last night, and it came out pretty good...at 1080/59fps/9889kbps,h.264. From 263MB -> 105MB (original was 25k kbps) May 14, 2014 at 6:37
  • @LouisvanTonder - cool, glad it came out better. I would recommend using software stabilization as well as it will give the encoder a far easier time with your video if it is at all shaky.
    – AJ Henderson
    May 14, 2014 at 13:25

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge that you have read and understand our privacy policy and code of conduct.

Not the answer you're looking for? Browse other questions tagged or ask your own question.