Question about ffmpeg

[koliva]

Dear all,

I have asked a question about encoding step by using ffmpeg. Now the problem is decoding. Whatever I do, I got always "could not find codec parameters",

For encoding,

ffmpeg -i out_%d.bmp -f rawvideo output.yuv

This is okay but how I can get my images back from output.yuv file?
I have tried

ffmpeg -i output.yuv -s 1400x800 rawvideo -f image2 out1_%d.bmp

but it didn't work.

Another question is, rawvideo means 4:0:0 or 4:2:2 ?

[Boolsheet]

You have to put the resolution, framerate and colorspace for the raw file before the -i parameter. This way FFmpeg knows how to apply them correctly:

ffmpeg -s 1400x800 -vcodec rawvideo -f rawvideo -pix_fmt yuv422p -i output.yuv -vcodec bmp -f image2 out1_%d.bmp

(I don't think you're supposed to save YUV images in BMP. It gets converted to RGB anyway, so this is not a lossless conversion)

For a list of the colorspaces execute:
ffmpeg -pix_fmt list

Another question is, rawvideo means 4:0:0 or 4:2:2 ?

Rawvideo means uncompressed and unencoded frames in the chosen colorspace.
4:2:2 and 4:2:0 is the notation for chroma subsampling used in YUV, it says how many pixels share the same color information.

[koliva]

Originally Posted by Boolsheet

You have to put the resolution, framerate and colorspace for the raw file before the -i parameter. This way FFmpeg knows how to apply them correctly:
(I don't think you're supposed to save YUV images in BMP. It gets converted to RGB anyway, so this is not a lossless conversion)

For a list of the colorspaces execute:
ffmpeg -pix_fmt list



Rawvideo means uncompressed and unencoded frames in the chosen colorspace.
4:2:2 and 4:2:0 is the notation for chroma subsampling used in YUV, it says how many pixels share the same color information.

Thank you for your reply. I stil have l a problem.

I used for encoding:

ffmpeg -i out_%d.bmp -vcodec rawvideo -f output.yuv

and decoding,

ffmpeg -s 1400x800 -vcodec rawvideo -f rawvideo -pix_fmt yuv400p -i output.yuv -vcodec bmp -f image2 out1_%d.bmp

First, I have 2 grayscale(8 bit) images(named out_1.bmp, out_2.bmp) to be encoded. Therefore, I dont have Cb Cr components in yuv format. When I decode that output.yuv file, I got only one greenish bmp file. I tried many different configurations but it seems that there is a problem when I am combining my images.

My file sizes are:

out_1.bmp: 1095 KB
out_2.bmp: 1095 KB

output.yuv: 2190 KB

out1_1.bmp: 3282 KB ???

[Boolsheet]

You can convert your bmps to yuv400p with the same pix_fmt parameter:

ffmpeg -i out_%d.bmp -vcodec rawvideo -f rawvideo -pix_fmt yuv400p output.yuv

[koliva]

Originally Posted by Boolsheet

You can convert your bmps to yuv400p with the same pix_fmt parameter:

It worked. Thank your so much