View Full Version : FFmpeg commands for converting 24-bit multichannel DTS to 16-bit stereo W64?
eddman
27th February 2015, 14:32
I'm looking for ffmpeg commands for the aforementioned conversion.
Also, is there a way to use the dolby pro logic method, like eac3to? I assume that gives a better result than simple, right?
Reino
1st March 2015, 12:53
Something like this?
ffmpeg.exe -hide_banner -i input.dts -af "aresample=matrix_encoding=dplii" -ac 2 output.w64
eddman
1st March 2015, 17:35
Something like this?
ffmpeg.exe -hide_banner -i input.dts -af "aresample=matrix_encoding=dplii" -ac 2 output.w64
Thanks. I don't see a bit depth command. Does ffmpeg convert to 16-bit by default? If so, what'd be the command to convert to/keep the same 24-bit depth?
poisondeathray
1st March 2015, 17:58
Yes, 16bit by default. Your title asked for 16bit
-c:a pcm_s24le for 24bit little endian
eddman
1st March 2015, 18:05
Yes, 16bit by default. Your title asked for 16bit
-c:a pcm_s24le for 24bit little endian
I know. I was just surprised that it didn't keep the original bit depth by default. It's kind of an odd thing.
Actually, I wanted to know the proper bit converting command; 16-bit was an example.
Thanks again.
little endian
I don't really know what that is and why it should be like that.
foxyshadis
2nd March 2015, 00:59
I don't really know what that is and why it should be like that.
Just the byte order to correctly read the samples, and since it's always tagged, it doesn't matter. Technically: x86/x64 are little endian, Motorola 6800 and its descendants were big endian, and all that means is that if the file doesn't match the CPU then the bytes have to be swapped internally to work with them. (ARM can be both endians.)
tl;dr: Always use le.
pandy
2nd March 2015, 10:20
Sometimes surround matrices may give suboptimal results (some TV's equipped with voice processing fail) then this perhaps can be useful - works fine for me.
@ffmpeg -threads %NUMBER_OF_PROCESSORS%*1.5 -i %1 -vn -c:a pcm_s16le -af "aformat=sample_fmts=fltp,pan=stereo|FL < FL + 1.414FC + .5BL + .5SL + 0.25LFE|FR < FR + 1.414FC + .5BR + .5SR + 0.25LFE,aresample=resampler=soxr:osr=48000:dither_method=shibata,aformat=sample_fmts=s16:channel_layouts=stereo[out]" -y %1.w64
vBulletin® v3.8.11, Copyright ©2000-2026, vBulletin Solutions Inc.