My lastest attempt was – Audio-only as ‘Flac’ But it’s creating a file that is larger than the reported m2ts file size, and it’s not done yet. I have no idea what monster will be created. But I doubt it will be usable.
You said you had “FFMpeg. N-75841-g5911eeb”. You can use that to try to convert the file. Hold SHIFT and right-click over the folder that contains ffmpeg.exe, and choose “Open Command Window Here”.
In the command window, type:
ffmpeg -i filename.m2ts 8_channel.wav
Latest Git FFmpeg should support raw DTS-HD, according to General Documentation.
Gale
It is doing something. But I doubt that is the only parameters I needed, take a look at the command window I got:
Let us know if the file completes or not, but for detailed support with FFmpeg I suggest you contact the ffmpeg-user list: https://ffmpeg.org/mailman/listinfo/ffmpeg-user.
Gale
My lastest attempt was – Audio-only as ‘Flac’ But it’s creating a file that is larger than the reported m2ts file size, and it’s not done yet.
Yeah… That doesn’t seem right… I’d expect the FLAC to be about the same size as the lossless DTS-HD stream, although a FLAC would be larger than a “regular” lossy DTS file/stream.
As a very rough estimate, I’d expect an 8-channel, 24-bit, 48kHz FLAC to be about 40MB per minute (about 60% of the size of the uncompressed PCM data).
It completed, and I got a 4gb file that I can play with Foobar by itself. Of course it is 6CH. If you looked at the image it seemed to grab the DTS-HD track. I also loaded it into Audacity, it reads very quickly, I can edit and mess with it. Let’s see if this version will produce better results?
The unfortunate truth here seems to be that the 2 rear-tracks appear to be matrix’d…I don’t need 7.1 with my current system, however it will sound weird, as it already has if they are played together as 2-tracks and not as 4 and not decoded, or the extra data not stripped off. I like to create 6CH ‘Apple-Lossless’ .m4a (s) Well see how it goes this time.
Oh duh… I could play it right now, in my theater room. But I can’t my wife is hanging out there. Later. LOL
Note: I got 16bits again, I was hoping for 24. This might be a setting issue with ffmpeg.
I haven’t had a chance to play the 4gb-wav data I recorded yet. But I was studying this article:
https://en.wikipedia.org/wiki/DTS-HD_Master_Audio The sub-heading ‘Combined lossless/lossy compression’ gets a bit over-my-head, and I don’t understand it much. I was wondering if this is actually explaining how the 7.1 is laid out on the disc. Somehow it uses 2-different audio tracks to conjure up the decoding of 1 (maybe). This makes it sound like, it’s not just as simple as the rear-channels being matrix’d.
I was wondering if this is actually explaining how the 7.1 is laid out on the disc. Somehow it uses 2-different audio tracks to conjure up the decoding of 1 (maybe). This makes it sound like, it’s not just as simple as the rear-channels being matrix’d.
The lossy/lossless concept is not related to the number of channels. And, the lossy & lossless information and all of the channels are in the same DTS-HD stream (or file).
DTS-HD Master Audio contains 2 data streams: the original DTS core stream and the additional “residual” stream which contains the “difference” between the original signal and the lossy compression DTS core stream.
If you save the difference information you can use it to re-create the original lossless data.
“Philosophically”, there’s no advantage in that compared to other methods of lossless compression except that it’s compatible with the “old” lossy decoder, as long as the decoder can ignore the additional difference information (which it doesn’t know anything about or what to do with).
Ok I had a chance to listen to the 4GB-wav created directly with ffmpeg. I can already tell the difference. It is great.
I’m also creating a new one with this setting: It will likely be 8GB…
-acodec pcm_s32le Because the default setting is -s16le…
So I’m good. Still no 7.1, but I’m ok with that for now. So anyone looking at this thread to figure out 7.1, will find not too much of value.
Thankx for all your help folks!!
Ok I figured out what was causing my original problem…
It was reencoding selected audio to Apple-lossless m4a. For some reason this is not working so well, or I’m losing something in the exchange.
After creating a new 32bit file as described above. Audacity could not read the resulting wav file… or rather it only read 1hr or so out of a 2hr file. Then I tried dumping it as ‘raw’. Then audacity would crash in the reading, and this time it was using ffmpeg in the reading anyway.
I switched back to reading the entire file the original way in the first place. The picking the file through audacity with ffmpeg installed. Back to square 1, and assuming my problem was my ‘reencoding’, when I sliced pieces.
I have another hypothesis… Is it also possible, that although the disc is marked 7.1 DTS-HD Master, that they are lying? Clearly the 7.1 is not discrete.
4 GB is the maximum size of WAV.
Import RAW uses libsndfile, not FFmpeg.
Why do you need a 32-bit WAV? Why not try encoding the M2TS file to FLAC with FFmpeg, which does not have a size limit?
Gale
OK, but how do you get at the difference information if it is not actually one of the streams listed when you import the file?
Gale
I did… I believe that is what audacity is doing anyway.
Here is another problem, some of the music just has the rear channels too loud, I’m guessing because of the 7.1… so I’m back to the 7.1 being matrix’d. So I’m having to Amplify other parts, and remix it to make it sound right.
Agreed… I only see one DCA — Only one DTS-HD stream all the rest are mared AC3. And some are different languages.
I thought that was in VLC? I meant, use your FFmpeg git using a similar command to the one I showed you but name the output file with .flac extension.
When Audacity imports the M2TS file using FFmpeg, it expands the file to PCM, not to FLAC.
Gale
I’m not sure what I’m going to gain from this, unless you mean that ‘expands the file to PCM’ is a bad thing. I’m going to cut pieces and make edits anyway. My final listenable version(s) that I will store will not be one whole large ‘.Flac’/‘.Wav’ with the entire movies audio.
Ok… When I do this command: ffmpeg -i “d:\bdmv\stream\00005.m2ts” i:\flac.flac
I get Stream #0:0: Audio: flac, 48000 Hz, 5.1(side), s32 (24 bit), 128 kb/s
I’m guessing the last part isn’t what I want…
The flac it created has a bitrate of 3428, and 24bits per sample. So I’m not sure what that was showing above. It is 3GB. Audacity is reading it now. Still I’m not sure what this will gain me over just reading the m2ts by itself. We’ll see if it sounds better when I cut and edit pieces.
I just embroiled my ears to hear the differences between a couple of different cuts I made from different versions. The ‘flac’ version suffered the same issue. I needed to remix it so that the rear channels were a bit lower and not over-powering the fronts. I hear virtually no difference in the Flac version as opposed to the PCM version created by Audacity and then reencoded (the same piece) to Flac from audacity.
I also felt it needed a tiny bit of treble. Maybe because I amplified the fronts. This was because of the remixing. But it sounds just incredible… the best sound almost I’ve ever heard in like in my life… I think it’s time I reveal what I’m listening to. It’s Daft Punk Tron-Legacy.
I would not expect it to sound better than WAV, and you say it doesn’t.
But it is a smaller file, and so Audacity can import if without running foul of the 4 GB limit for WAV.
You may do better on the specialist forums like http://www.afterdawn.com/ or http://www.videohelp.com/. They may know better tools, especially better analysis tools to show you what is really in the stream.
For example, have a look at Extract Content From A Blu-ray Disc - AfterDawn: Guides.
Gale