Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion. Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules. |
30th March 2021, 21:51 | #1 | Link |
Registered User
Join Date: Jul 2007
Location: Durban, South Africa
Posts: 11
|
DTS-HD to DTS Core Extraction
I was looking around the forum for a way to extract DTS from DTS-HD and I discovered this can be done with tsMuxeR using the Downconvert option. Another method involves demuxing the full DTS-HD track and using eac3to -core to extract the DTS track. Both methods produce a file of identical size, except eac3to patches the file to 24 bits whereas tsMuxeR creates a 16 bit file.
I found this post which gives an explanation for this behaviour but I don't fully understand it and I am still left unsure about which method is best to use? Is it better for the file to be patched to 24 bits or is the resultant 16 bit file created by tsMuxeR's downconversion okay? |
30th March 2021, 22:01 | #2 | Link |
Registered Developer
Join Date: Mar 2010
Location: Hamburg/Germany
Posts: 10,348
|
DTS core, like most lossy codecs, do not have an actual bitdepth. The difference here is that the header contains a flag called "original bitdepth", which some decoders use to decide on how to output the audio. If its 24-bit, there is a chance you might get a bit more precision.
Other decoders ignore this flag entirely and always output 16-bit, or always output floating point. In summary, its really not a huge difference, it doesn't change the audio, just a bit in the header. Overall the "24-bit version" is probably slightly better.
__________________
LAV Filters - open source ffmpeg based media splitter and decoders |
31st March 2021, 08:14 | #4 | Link |
Moderator
Join Date: Feb 2005
Location: Spain
Posts: 6,915
|
The "original bitdepth" of lossy encoders must be ignored always.
The internal bitdepth of lossy codecs are always imprecise (in function of bitrate) float samples in frequency domain. Like lossy, the exact original bitdepth never can be recovered no matter if is 16 or 24 bits. The first pass of decoders is convert the float samples in frequency domain to float samples in time domain, downconvert float samples to 16 bits int always can introduce more rounding errors than convert to 24 bits int. In prevention of decoders than not ignore the "original bitdepth" is always better patch that useless flag in dts header.
__________________
BeHappy, AviSynth audio transcoder. |
31st March 2021, 14:49 | #5 | Link |
Guest
Posts: n/a
|
The file created by tsMuxeR is okay. If you were focused on the nuances of how various decoders might handle the file, I doubt that you would be ditching the DTS-HD in favor of the core in the first place, so use whichever method is easiest. In other words, if you want the file for day-to-day listening and just want to watch a movie, you don't need to go out of your way to use eac3to just because it patches the track.
|
Tags |
dts, dts-hd |
|
|