Re: [FFmpeg-user] SAR changes when stream copying y4m video to Matroska when SAR is present in source

Kieran O Leary gmail.com> writes: > > I originally wanted to suggest replicating with FFmpeg alone > > (without an external source file) but feel free to use the > > mentioned file. > > > > Do you mean that I should use -f lavfi to generate a sample? If > so , I can try to do this. I first

Re: [FFmpeg-user] SAR changes when stream copying y4m video to Matroska when SAR is present in source

On Mon, Jul 25, 2016 at 10:34 AM, Carl Eugen Hoyos wrote: > > I originally wanted to suggest replicating with FFmpeg alone > (without an external source file) but feel free to use the > mentioned file. > Do you mean that I should use -f lavfi to generate a sample? If so , I

Re: [FFmpeg-user] SAR changes when stream copying y4m video to Matroska when SAR is present in source

Kieran O Leary gmail.com> writes: > > Feel free to open a ticket. > > Will do. Let me know if any further info or testing should > go into the trac report. I'd love to see a saner FFmpeg version string there (but that's only me) ;-) I originally wanted to suggest replicating with FFmpeg

Re: [FFmpeg-user] SAR changes when stream copying y4m video to Matroska when SAR is present in source

Hi Carl, Thanks for replying. On Mon, Jul 25, 2016 at 9:47 AM, Carl Eugen Hoyos wrote: > Which - without really verifying myself! - are probably > the correct display dimensions for the given input. > > I don't know why FFmpeg writes display dimensions instead > of DAR: My

Re: [FFmpeg-user] SAR changes when stream copying y4m video to Matroska when SAR is present in source

Kieran O Leary gmail.com> writes: > I posted a similar issue recently where an unknown SAR was > declared as 1:1 when stream copied to matroska. Allow me to repeat that this is not correct: If SAR is unknown, no sar is written to mkv output (when stream copied or reencoded). The demuxer does

[FFmpeg-user] SAR changes when stream copying y4m video to Matroska when SAR is present in source

Hello, I posted a similar issue recently where an unknown SAR was declared as 1:1 when stream copied to matroska. I think this is a different issue because the SAR is declared in the source in this instance. I am running some tests on the xiph/derf collection and found that a lot of the videos