Video Compression

[PkmoNk]

In VirtualDub:
Video>
Direct Stream Copy
Fast Recompression
Normal Recompression
Full Compression Modes

Does it make a difference if I don't use Filters nor Color Depth and I use Fast Recompression?

[neuron2]

Originally posted by PkmoNk
In VirtualDub:
Video>
Direct Stream Copy
Fast Recompression
Normal Recompression
Full Compression Modes

Does it make a difference if I don't use Filters nor Color Depth and I use Fast Recompression? Well, it's faster!

Can you state your question more clearly?

[PkmoNk]

I'm more concerned with the quality wise. If I enable Fast Recompression and turn off Full Processing, will it result with the same output quality?

[fccHandler]

"Fast recompress" first attempts to pass the video data in a YUV format; if that fails, then it attempts to pass the data in an RGB format (converting it if necessary).

Obviously, if the video is originally stored in a common YUV format (like YUY2), and the compressor supports the same format for input, then "fast recompress" is much faster than the other processing modes, because it avoids YUV->RGB / RGB->YUV conversions.

However, in my experience YUY2 data converted to RGB and then passed to the DivX codec seems to compress better than YUY2 passed directly via "fast recompress." I wonder why that is...

[PkmoNk]

sorry fccHandler, but I don't really understand you :D

If you can simplify that for noobs like me, it'll be great hehe

[manono]

Hi-

If you're feeding VDub an .avs, then you use Fast Recompress. Not only is it faster, but it's faster because it avoids the RGB color conversion that slows it down and degrades the quality by a little bit.

Direct Stream Copy isn't for encoding. It's used for making an exact copy of the video stream like, for example, when you add the audio to the video, or when you split a 2 CD AVI.

[PkmoNk]

What if my input is MPEG1?

[fccHandler]

Originally posted by PkmoNk
What if my input is MPEG1?
All MPEG video is YUV, so yes, "fast recompress" will be faster and more accurate, but only if the codec you're compressing to supports YUV input (I think all modern codecs do).

Originally posted by PkmoNk
If you can simplify that for noobs like me, it'll be great hehe
YUV format (and YCbCr which is very similar) stores video information by separating the colors into a luminance (brightness) channel and two chrominance (color) channels.

RGB format stores video information by specifying the relative amounts of Red, Green, and Blue, which (when combined) produce the colors you see.

You can convert from YUV to RGB and vice versa, but the conversion tends to be slow, and because of rounding errors the conversion may degrade the video slightly.