Convert DTS to Dolby and any Lossless to Apple Lossless [BluRay]

[FLX90]

Hello everybody,

I have a large BluRay collection and want it to get into my iTunes library to play on AppleTV.
I have no idea of transcoding and stuff like that, but I read many threads here and on other forums and acquire some knowledge.

First. Your program, madshi, is very great. Thank you for this awesome tool!





I need to convert the DTS-audio-streams to Dolby. And I used eac3to for that.


The Audio-Streams of the BluRay are the following and I need them to put into the mentioned codec:

- DTS --> Dolby Digital (AC-3)
- DTS-HD High Resolution --> Dolby Digital Plus (E-AC-3)
- DTS-HD Master --> Apple Lossless (ALAC)
- Dolby TrueHD --> Apple Lossless (ALAC)



Here are the code snippets and their logs I used for some example BluRay-Rips.



Code-Snippet for DTS-HD Master & Dolby Digital --> Apple Lossless & Dolby Digital:

mkvmerge -i german_dd_to_ac3__english_dtshdma_to_alac.mkv
mkvextract tracks german_dd_to_ac3__english_dtshdma_to_alac.mkv 0:video.h264 1:audio1.ac3 2:audio2.dtshdma 3:sub1.srt
ffmpeg -i german_dd_to_ac3__english_dtshdma_to_alac.mkv -f ffmetadata chapters
eac3to audio2.dtshdma audio2.flac
eac3to audio2.dtshdma audio2_core.ac3
ffmpeg -i audio2.flac -acodec alac audio2.m4a
ffmpeg^
-i video.h264 -i audio1.ac3 -i audio2.m4a -i audio2_core.ac3 -i sub1.srt -i chapters^
-map 0:0 -map 1:0 -map 2:0 -map 3:0 -map 4:0^
-metadata:s:a:0 language=ger -metadata:s:a:0 handler="Dolby Digital" -metadata:s:a:1 language=eng -metadata:s:a:1 handler="Apple Lossless" -metadata:s:a:2 language=eng -metadata:s:a:2 handler="Dolby Digital" -metadata:s:s:0 language=ger -metadata:s:s:0 handler="Deutsch"^
-disposition:a:0 default -disposition:a:1 0 -disposition:a:2 0^
-movflags disable_chpl^
-c:s mov_text -c:v copy -c:a copy^
german_dd_to_ac3__english_dtshdma_to_alac.m4v 2>> ffmpeg.log

ffmpeg1.log



Code-Snippet for DTS-HD Master & DTS --> Apple Lossless & Dolby Digital:

mkvmerge -i german_dts_to_ac3__english_dtshdma_to_alac.mkv
mkvextract tracks german_dts_to_ac3__english_dtshdma_to_alac.mkv 0:video.h264 1:audio1.dts 2:audio2.dtshdma 3:sub1.srt
ffmpeg -i german_dts_to_ac3__english_dtshdma_to_alac.mkv -f ffmetadata chapters
eac3to audio1.dts audio1.ac3
eac3to audio2.dtshdma audio2.flac
eac3to audio2.dtshdma audio2_core.ac3
ffmpeg -i audio2.flac -acodec alac audio2.m4a
ffmpeg^
-i video.h264 -i audio1.ac3 -i audio2.m4a -i audio2_core.ac3 -i sub1.srt -i chapters^
-map 0:0 -map 1:0 -map 2:0 -map 3:0 -map 4:0^
-metadata:s:a:0 language=ger -metadata:s:a:0 handler="Dolby Digital" -metadata:s:a:1 language=eng -metadata:s:a:1 handler="Apple Lossless" -metadata:s:a:2 language=eng -metadata:s:a:2 handler="Dolby Digital" -metadata:s:s:0 language=ger -metadata:s:s:0 handler="Deutsch"^
-disposition:a:0 default -disposition:a:1 0 -disposition:a:2 0^
-movflags disable_chpl^
-c:s mov_text -c:v copy -c:a copy^
german_dts_to_ac3__english_dtshdma_to_alac.m4v 2>> ffmpeg.log

ffmpeg2.log



Code-Snippet for Dolby TrueHD & Dolby Digital --> Apple Lossless & Dolby Digital:

mkvmerge -i german_dd_to_ac3__english_truehd_to_alac.mkv
mkvextract tracks german_dd_to_ac3__english_truehd_to_alac.mkv 0:video.h264 1:audio1.ac3 2:audio2.truehd 3:audio2_core.ac3 4:sub1.srt
ffmpeg -i german_dd_to_ac3__english_truehd_to_alac.mkv -f ffmetadata chapters
eac3to audio2.truehd audio2.flac
ffmpeg -i audio2.flac -acodec alac -af "aformat=channel_layouts=7.1(wide)" audio2.m4a
ffmpeg^
-i video.h264 -i audio1.ac3 -i audio2.m4a -i audio2_core.ac3 -i sub1.srt -i chapters^
-map 0:0 -map 1:0 -map 2:0 -map 3:0 -map 4:0^
-metadata:s:a:0 language=ger -metadata:s:a:0 handler="Dolby Digital" -metadata:s:a:1 language=eng -metadata:s:a:1 handler="Apple Lossless" -metadata:s:a:2 language=eng -metadata:s:a:2 handler="Dolby Digital" -metadata:s:s:0 language=ger -metadata:s:s:0 handler="Deutsch"^
-disposition:a:0 default -disposition:a:1 0 -disposition:a:2 0^
-movflags disable_chpl^
-c:s mov_text -c:v copy -c:a copy^
german_dd_to_ac3__english_truehd_to_alac.m4v 2>> ffmpeg.log

ffmpeg3.log



Code-Snippet for DTS-HD Master & Dolby Digital Plus --> Apple Lossless & Dolby Digital Plus:

mkvmerge -i german_dd+_to_eac3__english_dtshdma_to_alac.mkv
mkvextract tracks german_dd+_to_eac3__english_dtshdma_to_alac.mkv 0:video.h264 1:audio1.eac3 2:audio2.dtshdma 3:sub1.srt 4:sub2.srt
ffmpeg -i german_dd+_to_eac3__english_dtshdma_to_alac.mkv -f ffmetadata chapters
eac3to audio2.dtshdma audio2.flac
eac3to audio2.dtshdma audio2_core.ac3
ffmpeg -i audio2.flac -acodec alac -af "aformat=channel_layouts=7.1(wide)" audio2.m4a
ffmpeg^
-i video.h264 -i audio1.eac3 -i audio2.m4a -i audio2_core.ac3 -i sub1.srt -i sub2.srt -i chapters^
-map 0:0 -map 1:0 -map 2:0 -map 3:0 -map 4:0 -map 5:0^
-metadata:s:a:0 language=ger -metadata:s:a:0 handler="Dolby Digital Plus" -metadata:s:a:1 language=eng -metadata:s:a:1 handler="Apple Lossless" -metadata:s:a:2 language=eng -metadata:s:a:2 handler="Dolby Digital" -metadata:s:s:0 language=ger -metadata:s:s:0 handler="Deutsch" -metadata:s:s:1 language=eng -metadata:s:s:1 handler="Englisch"^
-disposition:a:0 default -disposition:a:1 0 -disposition:a:2 0 -disposition:s:0 default^
-movflags disable_chpl^
-c:s mov_text -c:v copy -c:a copy^
german_dd+_to_eac3__english_dtshdma_to_alac.m4v 2>> ffmpeg.log

ffmpeg4.log

ffmpeg5.log (also tried with mp4-container)



Code-Snippet for DTS-HD Master & DTS-HD High Resolution --> Apple Lossless & Dolby Digital Plus:

mkvmerge -i german_dtshr_to_eac3__english_dtshdma_to_alac.mkv
mkvextract tracks german_dtshr_to_eac3__english_dtshdma_to_alac.mkv 0:video.h264 1:audio1.dtshr 2:audio2.dtshdma
ffmpeg -i german_dtshr_to_eac3__english_dtshdma_to_alac.mkv -f ffmetadata chapters
eac3to audio1.dtshr audio1.eac3
eac3to audio2.dtshdma audio2.flac
eac3to audio2.dtshdma audio2_core.ac3
ffmpeg -i audio2.flac -acodec alac -af "aformat=channel_layouts=7.1(wide)" audio2.m4a
ffmpeg^
-i video.h264 -i audio1.eac3 -i audio2.m4a -i audio2_core.ac3 -i chapters^
-map 0:0 -map 1:0 -map 2:0 -map 3:0^
-metadata:s:a:0 language=ger -metadata:s:a:0 handler="Dolby Digital Plus" -metadata:s:a:1 language=eng -metadata:s:a:1 handler="Apple Lossless" -metadata:s:a:2 language=eng -metadata:s:a:2 handler="Dolby Digital"^
-disposition:a:0 default -disposition:a:1 0 -disposition:a:2 0^
-movflags disable_chpl^
-c:v copy -c:a copy^
german_dtshr_to_eac3__english_dtshdma_to_alac.m4v 2>> ffmpeg.log



First of all.
The procedure is always the same (make mkv from BD --> extract mkv --> convert audiostreams --> put them into m4v container).
Is this a proper way to do that?



The first three are working very well and the last two have problems with E-AC-3.

So here are the questions I have:


1. My chapters from the mkv I extract with ffmpeg (ffmpeg -i input.mkv -f ffmetadata chapters) are in text-format. How can I convert them to tx3g (or extract tx3g from mkv).
2. I can set a subtitle to default (-disposition:s:0 default), I need it to be default and forced. Trying it the same way (-disposition:s:0 forced) did not work. How can I do that?
3. When I have to transcode an audio stream with eac3to (so except lossless -> lossless convert), there is always a difference of 0.03 sec to source audio stream (transcode: 01:40:05.79; source: 01:40:05.76). How can I avoid this?
4. I have some warnings about timestamps and durations in my log:
[ac3 @ 00000000005989c0] Estimating duration from bitrate, this may be inaccurate
[ipod @ 0000000002ca4940] Timestamps are unset in a packet for stream 0. This is deprecated and will stop working in the future. Fix your code to set the timestamps properly
[ipod @ 0000000002ca4940] pts has no value
Will they disappear when I put timestamps into pts-format in the container?
I can extract them from mkv (mkvextract timecodes_v2 input.mkv 0:video.timecodes.txt 1:audio1.timecodes.txt 2:audio2.timecodes.txt 3:sub1.timecodes.txt 4:sub2.timecodes.txt 5:sub3.timecodes.txt 6:sub4.timecodes.txt) but only in timecodes_v2-format. So how can I convert timecodes_v2 in pts and put them into the container?
4. How to find out the frame size and tell it to ffmpeg (because of this warning: [ipod @ 0000000002ca4940] track 1: codec frame size is not set)?
5. What does this warning message mean: [h264 @ 0000000000444240] Stream #0: not enough frames to estimate rate; consider increasing probesize. And how to avoid it?
6. I will buy the necessary software to get the best results from eac3to (Nero, ArcSoft, Sonic, SurCode, ...). In theoretic, will I get better quality results when I use DTS:X Content Creator Suite and Dolby Media Producer Suite?
7. The conversion of DTS-HD Master and Dolby TrueHD to Apple Lossless takes very long time and CPU usage compared to FLAC to ALAC. Why, when it should be only a lossless to lossess "copy"?


Now the conversions I could not handle.

DTS-HD Master & Dolby Digital Plus --> Apple Lossless & Dolby Digital Plus:
I could not put the E-AC-3 stream into the m4v-Container ([ipod @ 000000000328e140] Could not find tag for codec eac3 in stream #1, codec not currently supported in container; Could not write header for output file #0 (incorrect codec parameters ?): Invalid argument). So I tried the mp4-container.
The conversion is almost finished, than ffmpeg crashes with the following error message: Assertion s->buf_ptr < s->buf_end failed at src/libavcodec/put_bits.h:108.
How can I put E-AC-3 into the m4v-Container?


DTS-HD Master & DTS-HD High Resolution --> Apple Lossless & Dolby Digital Plus:

Eac3to was not able to convert DTS-HD High Resolution to Dolby Digital Plus (eac3to audio1.dtshr audio1.eac3 --> This audio conversion is not supported.). How can I handle this?





It would be really nice if you can help me with my questions.


Thank you very much!


Best regards,
Felix

[hubblec4]

Welcome to Doom9.
1. My chapters from the mkv I extract with ffmpeg (ffmpeg -i input.mkv -f ffmetadata chapters) are in text-format. How can I convert them to tx3g (or extract tx3g from mkv).

What is tx3g? Could you share a sample with me?

[tebasuna51]

Many questions, I can write some answers but we need a ffmpeg expert (I'm not) to others.

3. When I have to transcode an audio stream with eac3to (so except lossless -> lossless convert), there is always a difference of 0.03 sec to source audio stream (transcode: 01:40:05.79; source: 01:40:05.76). How can I avoid this?

All encoders insert a initial silence delay, also the lossy frames (minim blocks of samples, from 11ms-DTS to 32 ms-AC3) have different sizes and sometimes the last frame have silence at end.

6. I will buy the necessary software to get the best results from eac3to (Nero, ArcSoft, Sonic, SurCode, ...). In theoretic, will I get better quality results when I use DTS:X Content Creator Suite and Dolby Media Producer Suite?

You don't need buy nothing all can be do with free software.

How can I put E-AC-3 into the m4v-Container?...
Eac3to was not able to convert DTS-HD High Resolution to Dolby Digital Plus (eac3to audio1.dtshr audio1.eac3 --> This audio conversion is not supported.). How can I handle this?

You can't put EAC3 or DTS in mp4 container, see audio codecs allowed in https://en.wikipedia.org/wiki/Comparison_of_video_container_formats

Please use the preferred mp4 audio codec AAC for lossy audio, the size/quality is comparable to EAC3 and better than AC3.

[SeeMoreDigital]

Hi @FLX90,

Out of interest... What playback device(s) are you using?

[FLX90]

Thank you for the answers so far.


What is tx3g? Could you share a sample with me?


I can only present you the meta information I get from a m4v-File. It is DRM-protected so I cannot extract the chapters:


Menu
ID : 34
Codec ID : tx3g
Duration : 2 h 48 min
Title : Core Media Text
Language : German
Encoded date : UTC 2017-06-10 01:38:06
Tagged date : UTC 2017-06-10 01:39:07
Menu For : 2,36,37,38,39
00:00:00.000 : 2
00:09:33.000 : 3
00:17:09.000 : 4
00:24:58.000 : 5
00:35:43.000 : 6
00:42:15.000 : 7
00:50:18.000 : 8
00:54:46.000 : 9
00:58:57.000 : 10
01:06:12.000 : 11
01:17:26.000 : 12
01:25:30.000 : 13
01:34:30.000 : 14
01:46:13.000 : 15
01:55:32.000 : 16
02:02:12.000 : 17
02:12:55.000 : 18
02:20:52.000 : 19
02:35:10.000 : 20
Bit rate mode : VBR




You don't need buy nothing all can be do with free software.


Can be done, yes, but not with best quality results.
According to this https://en.wikibooks.org/wiki/Eac3to/How_to_Use#Audio_Decoders I would need Nero, ArcSoft and SurCode at least to get the best sound quality.


You can't put EAC3 or DTS in mp4 container, see audio codecs allowed in https://en.wikipedia.org/wiki/Comparison_of_video_container_formats

Please use the preferred mp4 audio codec AAC for lossy audio, the size/quality is comparable to EAC3 and better than AC3.

I tried it and DTS and DTS-HD Master Audio are doing fine in mp4-Container. Only Dolby Digital Plus has problems with the mp4-container.
But the container should support it, because there is no error message like I get when I try it with m4v-container, only a total crash of ffmpeg at the end of the conversion.

I would prefer to put DD+ back in the container without transcoding (losing information). For lossy transcoding that have to be done (DTS) I will use AAC.



Out of interest... What playback device(s) are you using?


I will use Apple TV 4K.

[sneaker_ger]

If ffmpeg crashes:
https://ffmpeg.org/bugreports.html


In general I find your solutions to be overly complicated. ffmpeg can read from mkv, convert the audio and then output mp4 in a single step. It's not necessary to split it into several steps with demuxing, converting, remuxing. Note that eac3to's encoders and decoders are outdated or sometimes non-existant compared to ffmpeg.

[tebasuna51]

Can be done, yes, but not with best quality results.
According to this https://en.wikibooks.org/wiki/Eac3to/How_to_Use#Audio_Decoders I would need Nero, ArcSoft and SurCode at least to get the best sound quality.

That page are obsolete.

You need Nero7+plugins only to decode AAC with eac3to, and AAC don't be recoded to preserve the max quality.

But you can use ffmpeg to decode AAC, AC3, EAC3 (also 7.1), TrueHD, DTS (also DTS-HD) and don't need Nero or ArcSoft at all.

Surcode is a DTS standard (lossy) encoder not recomended at all, use AAC or even AC3 to lossy encode.

[SeeMoreDigital]

I will use Apple TV 4K.

Does the Apple TV 4K support playback of multi-channel ALAC in .mp4?

[FLX90]

After careful thought I will convert all Audio-Streams to ALAC.
Space is no problem and so I can preserve the best quality.
Lossy --> lossless: No transcode and so no lose of information. Am I right?

I solved question #2 with a program called subler (only for mac). It is doing the whole tagging very well.

Question #1, #4, #4 (my fault ...) and #5 are still present.

Does the Apple TV 4K support playback of multi-channel ALAC in .mp4?

Unfortunately up to now I don't have an Apple TV 4K.
I tested with Apple TV 4 and the audio playback was fine. (Video playback was choppy, but I think the 4K ATV will solve this problem.)

[SeeMoreDigital]

Hi again FLX90,

What make and model of 4K display do you have?

[FLX90]

What make and model of 4K display do you have?

I don't have a 4K-Display and won't buy one.

All my BluRays are 1080p, so I don't need 4K.

[SeeMoreDigital]

I don't have a 4K-Display and won't buy one.

All my BluRays are 1080p, so I don't need 4K.So why bother buying an Apple 4K TV STB?

[FLX90]

So why bother buying an Apple 4K TV STB?

Support of Gigabit Ethernet and A10X instead of A8 (hope the better CPU can handle the high bitrate-video-streams better).

[FLX90]

Solved #4, #4 and #5.

#4a & #5:
Warning had something to do with variable frame rate and timecodes.
Timecodes get lost during extract h264 from mkv.
Avoid this problem by directly converting from one container to another:
ffmpeg -i german_dd_to_alac__english_dtshdma_to_alac.mkv -map 0:0 -c copy temp.m4v

Choppy video playback mentioned here (https://forum.doom9.org/showthread.php?p=1858325#post1858325) had also something to do with converting VFR to CFR (MPC-HC handled that anyway, ATV doesn't).

With this, probesize error also disappeared.

#4b:
Solved this by specifying audio codec -c:a alac.
Don't know why but worked.


Another question:
Like I said (https://forum.doom9.org/showthread.php?p=1858325#post1858325) I'll convert everything to ALAC.
But I don't know which bit depth for the lossy audio streams.
To lose as less information as possible I do the following:
Convert lossy audio to 24 bit flac.
Check if last 8 bits are used.
If so I convert to 24 bit ALAC.
If not I convert to 16 bit ALAC.

Is this the best way to do?

I asked in another forum and they said I should convert to 16 bit and don't waste much thought on the 8 bits.
But couldn't explain me why.
My thinking was that 48 dB possible difference in dynamic range (144 db <-> 96 dB) is essential to preserve, especially for movie audio.

What would you say?

[tebasuna51]

Like I said I'll convert everything to ALAC.
But I don't know which bit depth for the lossy audio streams.
To lose as less information as possible I do the following:
Convert lossy audio to 24 bit flac.
Check if last 8 bits are used.

All lossy decoders output at least 32 bits float, with a precision, more or less, equivalent to 24 bits int (24 bits of float mantissa).
Always the last 8 bits are used (or they are truncated/rounded before).

Even if source was 16 bits, for instance audio CD, a lossy encoder can't preserve the 16 bits exact, for that is lossy.

I asked in another forum and they said I should convert to 16 bit and don't waste much thought on the 8 bits.
But couldn't explain me why.

The medium human ear can distinguish until a precision of 20 bits, then always is better 24 bits than16 when source is lossless.

But with lossy sources these 8 bits are always unexact, the precision is always less than 16 bits, that is lossy encoders.

[FLX90]

The medium human ear can distinguish until a precision of 20 bits, then always is better 24 bits than16 when source is lossless.

But with lossy sources these 8 bits are always unexact, the precision is always less than 16 bits, that is lossy encoders.

But when "lossy decoders output at least 32 bits float" why is the precision "always less than 16 bits"?

So my understanding:
16 bit lossless --> lossy (24 bit) --> last 8 bits are guessed/interpolated (?)
24 bit lossless --> lossy (24 bit) --> last 8 bits are taken approximated from the lossless

conclusion:
(1) 24 bit lossless --> 24 bit lossy
(2) 24 bit lossless --> 16 bit lossless --> 24 bit lossy

(1) is closer to the originally source and so a 24 bit lossless from a lossy is closer to the originally source (if source was 24 bit).

Is that right?

[nevcairiel]

No, thats not right.

When dealing with lossy audio formats, stop thinking about bitdepth entirely, because its not a concept it tries to maintain at all. You can write a lossy decoder that decodes to 16-bit integer, or you can write one that decodes to 32-bit floating point. Or any other data format if you're so inclined.
There are no guarantees that any amount of bits from the source are maintained accurately, and there is no way to generally tell how many bits of actual "precision" a lossy decoder outputs. This can and will vary between different formats, decoder implementations, and even the material in question.

[FLX90]

Sorry guys.
I really don't understand this.

The lossy formts are:
name depth
u8 8
s16 16
s32 32
flt 32
dbl 64
u8p 8
s16p 16
s32p 32
>>>fltp 32
dblp 64
s64 64
s64p 64

Why wouldn't the most exact image be the s32p?

Or I asked the other way round:
What would be the most exact way to image something stored in fltp?

EDIT:
Tebasuna51, almost ten years ago you wrote that (https://forum.doom9.org/showthread.php?p=1251299#post1251299):


- A sample in time domain have a bitdepth, but the samples are traslated to frequency domain when encoded, and the source bitdepth is lost.
The samples in frequency domain are stored with a precission equivalent to 20-24 bits in time domain (AC3,DTS).
The best option when decode an ac3/dts is use, at least, a bitdepth of 24 bits not matter what is the source precission.

Doesn't this answer contradicts your answer from above?

[nevcairiel]

Plain and simple: Stop thinking about bitdepth with lossy codecs! It makes no sense, its merely a property of how a decoder works, not the format itself.

[SeeMoreDigital]

Out of interest...

Doesn't the Apple TV 4 transcode all audio formats muxed within the .mp4 container, such as: AAC, AC-3, E-AC-3, etc, to a (multi-channel) PCM stream?

[tebasuna51]

...
- A sample in time domain have a bitdepth, but the samples are traslated to frequency domain when encoded, and the source bitdepth is lost.
The samples in frequency domain are stored (*) with a precission equivalent to 20-24 bits in time domain (AC3,DTS).
The best option when decode an ac3/dts is use, at least, a bitdepth of 24 bits not matter what is the source precission.

Doesn't this answer contradicts your answer from above?

(*) But not all components fit in the limited bitrate and some of them are discarded (less amplitude, high frequency) and not stored.

Then we can't recover the exact original samples in time domain, but the best approach is use 24 bits thinking in a new lossy encode.

[FLX90]

Out of interest...

Doesn't the Apple TV 4 transcode all audio formats muxed within the .mp4 container, such as: AAC, AC-3, E-AC-3, etc, to a (multi-channel) PCM stream?

Yes, that's right (but cannot handle DTS etc.).

(*) But not all components fit in the limited bitrate and some of them are discarded (less amplitude, high frequency) and not stored.

Then we can't recover the exact original samples in time domain, but the best approach is use 24 bits thinking in a new lossy encode.

Thank you for the clarification.

[SeeMoreDigital]

Yes, that's right (but cannot handle DTS etc.).
So just re-encode any file that contains DTS to AC-3 (Dolby Digital @ 640Kbps) and place the new AC-3 stream into the .mp4 container.... Job done!

EDIT: I've just found out about an Apple app called Infuse (https://firecore.com/infuse), which supports playback of all current popular audio, video and container formats!


Cheers

[FLX90]

EDIT: I've just found out about an Apple app called Infuse (https://firecore.com/infuse), which supports playback of all current popular audio, video and container formats!


I want to use iTunes to manage my library (smart playlist and many more nice features)
And I want to use iTunes Search API to do the tagging get 3Kx3K covers etc.


So I solved all of my problems and everything is fine.
25 % of my library are already digitized.


Thank you all for your help and your detailled answers!