5
$\begingroup$

I am trying to use Blender to edit game play videos that I capture with Nvidia's Shadowplay. I have learned what I know so far by watching these Blender tutorials which will be linked below.

My problem is that the original video source is 1080p at 60fps. When I finish editing and attempt to save the video as 1080p at 60fps, the quality to bad. The video is grainy and blurry. Nothing like the source video which is clear.

So my question is what can I do to improve the quality of the video the way a 1080p video should be?

This is to the tutorials I mentioned before. https://www.youtube.com/playlist?annotation_id=annotation_4173738165&feature=iv&list=PLjyuVPBuorqIhlqZtoIvnAVQ3x18sNev4&src_vid=xSGIPmQdV6M

My settings

$\endgroup$
  • $\begingroup$ Change the export settings to a better format then xvid $\endgroup$ – GiantCowFilms Feb 17 '15 at 16:49
  • $\begingroup$ What do you consider to be a better format? $\endgroup$ – Brett B Feb 17 '15 at 17:21
  • $\begingroup$ I watched and followed the instructions in the video linked below. It's not 100% perfect, but I am happy with it. The quality is a lot better than the previous attempts. youtube.com/… $\endgroup$ – Brett B Feb 17 '15 at 18:42
3
$\begingroup$

Try changing from Xvid to H.264:

enter image description here

$\endgroup$
  • $\begingroup$ I did what you said and what was shown in your picture. And the end result is the same. $\endgroup$ – Brett B Feb 17 '15 at 17:23
  • $\begingroup$ Could you take another screenshot with your settings changed? $\endgroup$ – Einar Feb 17 '15 at 17:39
  • $\begingroup$ Can avi even take h264 codec? If not, it probably defaults back to xvid or so. Rather put it in an mp4. $\endgroup$ – Greg Zaal Feb 19 '15 at 12:16
  • $\begingroup$ These are my actual settings, that I use for animations. They should work fine, as far as I know =) $\endgroup$ – Einar Feb 19 '15 at 16:14
1
$\begingroup$

You need to increase the bitrate settings in the encoding panel. The bitrate defines the amount of space that can be used for each frame, the lower the bitrate the more the data needs to be compressed to fit within that rate.

For comparison consider a jpeg image, at 90% quality it compresses to 189k and at 10% quality it compresses to 44k. The 90% image appears much cleaner than the 10% image.

In a video the bitrate has the same effect, if the bitrate says it can only use 50K for this part of the video it needs to compress at 10% quality leaving you with a poor quality video. By increasing the bitrate you are allowing the encoder to use a higher quality compression to get a better result. Also by running at 60fps you would expect to want a bitrate almost double a video running at 30fps. You will also want the maximum rate to be higher than the bitrate.

An image at 1920x1080 using 3 bytes per pixel (RGB) takes 6MB of ram, times that by 60fps and you can easily see that the default bitrate of 6000kb/s will need to use some heavy compression.

While I am skipping the more detailed explanation, the basic concept is there, in reality many things like amount of the frame that changes and GOP size also plays a part in the final quality.

You can find many opinions on what bitrate to use, at the end of the day it comes down to your video and what quality level looks good to you and how big you want your final video file.

$\endgroup$
0
$\begingroup$

Personally (if you have the harddrive space) I would render the final composition to a lossless format (AVI for example) and then use multipass h.264 compression with this open source video compression software - https://handbrake.fr/

Rendering h.264 in multiple passes gives you better quality compression though it takes longer. Having the ability to compress the video and compare that to your uncompressed AVI will allow you to tune your settings to keep the best quality for the lowest file size. They even have a youtube preset!

$\endgroup$
  • 2
    $\begingroup$ Rendering in multiple passes doesn't give you better quality. It simply improves the bitrate management. But yes, using an external encoder would be best. $\endgroup$ – Mike Pan Feb 18 '15 at 4:31
  • $\begingroup$ sorry, i should have stated it that way, I using quality to refer to the compression more than the image its self $\endgroup$ – Haakon Feb 18 '15 at 19:50
  • $\begingroup$ Unless you need to target a specific file size, multipass isn't the best option with H.264. CRF (acronym for Constant Rate Factor) is better. It produces very predictible quality, but depending on the properties of the input, the bitrate, and thus the final file size, may vary greatly. $\endgroup$ – user7952 Feb 18 '15 at 23:58

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

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