Hi, sorry if this question seems trivial, but I could not find an answer anywhere.
So, I have an AC3 file that came from a Dolby TrueHD stream. I didn't encode it, so I don't know how it was extracted.
This is the information output by MediaInfo:
Wasn't AC3 supposed to be a lossy file with a 640 kbps bitrate?Code:General Complete name : C:\Users\XXXX\Desktop\1.ac3 Format : TrueHD / AC-3 File size : 402 MiB Duration : 1h 27mn Overall bit rate : 640 Kbps Audio Format : TrueHD / AC-3 Mode extension : CM (complete main) Format settings, Endianness : Big Muxing mode : After core data Duration : 1h 27mn Bit rate mode : Variable / Constant Bit rate : 640 Kbps Maximum bit rate : 2 916 Kbps Channel(s) : 6 channels Channel positions : Front: L C R, Side: L R, LFE Sampling rate : 48.0 KHz Bit depth : 16 bits Compression mode : Lossless Stream size : 402 MiB (100%)
However, I see there "Lossless" and "2916 kbps".
I really thought that you would need the thd+ac3 file to contain the lossless data.
Once again, sorry if this is a stupid question.
+ Reply to Thread
Results 1 to 16 of 16
-
-
Unless I'm mistaken, an AC3 and a TrueHD file use the same headers & data packetizing, so to something like MediaInfo, it will look like the same stream, and unless it parses BOTH kinds of elements, it could be that it is only reading the AC3 portion.
Most of these kinds of files have a backwards-compatible "core" section, followed by the extended "non-core" section.
One way to tell with your file is this: a standard AC3 file uses Constant Bitrate, so an 87min title that is 402MiB does roughly equal 640kbps (closer to 630, but MI might not be showing exact amounts anyway). A proper TrueHD file (aka MLP), being lossless, must by nature be Variable Bitrate (the lossless residual must vary to match up with what was encoded within the lossy section so that the resulting stream restores to LPCM). And a proper TrueHD file would likely be in the range of the aforementioned 2916kbps*, so 2916 (averaged) x 5220 sec = ~1858MiB or 1.81GiB. That's much more than the file you've got there.
Scott
*(by comparison, an uncompressed 5ch 16bit 48kHz LPCM file would be 3750kbps).
<edit>@pdr, the file could be nominally a TrueHD, but with the extension portion removed (with headers left in). Or it could be that MediaInfo is just showing the filesize for the AC3 portion? (doubt it), or miscalculating other fields...
How else would you account for the bitrate / filesize / running time calculations?Last edited by Cornucopia; 9th May 2013 at 18:25.
-
-
Kelinya - For the record, you don't have to pretend to be a woman here (your avatar) to get help. And please spare me the inevitable "But I really am a woman!" response. The odds of a woman giving a crap about the subject of your post are about 1 in 1000 at best, so I'm just going to take the safe bet here. You're welcome to stay and maybe you'll like it enough to do so, but spare me the fake "hint hint - I'm really a woman - help me guys!" story.
-
-
@poisondeathray
@Cornucopia
Thanks very much for the input.
So this is probably a lossy core with wrong headers, is that it?
But one thing that I haven't pointed out; the file is actually 26 min long when muxed with the other streams.
For some reason, when extracted, it shows 87 min, however if I play it like that, the player indicates 87 min but skips 3 seconds each second, ending indeed after 26 min...
And in that case, I think 402 MiB could well be lossless data for that duration. But again, why does it show an overall bitrate of 640 kbps?
I actually have a flac from the same source, also (supposedly) lossless, and this is what it says:
Code:Audio ID : 2 Format : FLAC Format/Info : Free Lossless Audio Codec Codec ID : A_FLAC Duration : 26mn 13s Bit rate mode : Variable Channel(s) : 6 channels Sampling rate : 48.0 KHz Bit depth : 16 bits Title : FLAC, 6 channels, 48.0 KHz, 16 bits, 1 311 Kbps Writing library : libFLAC 1.2.1 Default : Yes Forced : No
Anyway, I know that DTS HD MA works like that, by using additional data to complement the lossy core in order to deliver the lossless data. But I've read somewhere that TrueHD does not work like that, having actually two different interleaving data streams. Is that how it really works?
@jman98
Geez, what the hell...
Mate, I will not discuss my gender here, since I think is kind of irrelevant.
But I must say, the avatar is from Haruhi Suzumiya, a series which I'm a huge fan of. So, that's not really a woman, is it? And my username is actually my real surname, not even a first name.
And, by your logic, when I see an avatar with a dog, what? that individual is trying to be accepted by the forum's dog pack? Maybe Noahtuck there is trying to pass himself off as a corpse...
Now we're having a nice discussion here and you come to the thread to say that... Please, reread your post. Don't you think that you are maybe, just maybe, being a little childish, sexist and even paranoid?
Okay, back on topic... -
If mediainfo is wrong about the duration (and it often is about many things , because it doesn't scan or parse the file) , and it's only 26min, then it's more likely TrueHD .
mediainfo always shows 640kbps for TrueHD streams - of course this is only the core
You can calculate the actual average bitrate yourself: filesize = bitrate x running time
Anyway, I know that DTS HD MA works like that, by using additional data to complement the lossy core in order to deliver the lossless data. But I've read somewhere that TrueHD does not work like that, having actually two different interleaving data streams. Is that how it really works? -
That's correct for DTS-HD Monster Audio, but it's incorrect for TrueHD.
The Blu-ray specs are a different subject though.
In the Blu-ray specification, .........
Since TrueHD is an optional codec, a separate (primary) AC-3 audio track must be present on Blu-ray discs that contain a TrueHD track.
In the HD DVD format, Dolby TrueHD was a mandatory codec. The HD DVD specification permitted the sole (primary) audiotrack of a movie to be encoded in Dolby TrueHD, -
-
I said and meant the TrueHD stream is lossless, and it doesn't depend on the (so-called) core to be decompressed. Simply because TrueHD is a superset of MLP, and MLP has nothing to do with AC3.
Whereas DTS-HD Master Audio is made of a lossy substream (the DCA "core") PLUS the difference between the core and the original uncompressed audio.Last edited by El Heggunte; 10th May 2013 at 00:12.
-
ahh ok thanks for the clarification
more info from madshi
http://forum.doom9.org/showthread.php?p=1415246#post1415246
There are 2 different TrueHD streams:
(1) TrueHD+AC3. This is a mixed format. There are both (alternating) TrueHD and AC3 frames in the stream.
(2) TrueHD. This stream only contains TrueHD frames and no AC3 frames.
All Blu-Ray TrueHD streams are (1). All HD DVD TrueHD streams are (2). eac3to can convert between both stream types. Converting from (1) to (2) is very easy: The AC3 frames are simply removed. Converting from (2) to (1) is more difficult: The TrueHD stream must be decoded and reencoded to AC3 and then the AC3 frames must be injected into the TrueHD stream in a specific way.
When muxing to MKV, the AC3 frames are usually removed, so MKV TrueHD streams are usually type (2).
There's no special metadata which tells anyone how to convert TrueHD to AC3, AFAIK. -
Yes, these last 2 posts provide the most correct info. DTS-MA is really the only BD legal stream that has a backwards-compatible lossy core stream packed with an extended (non-bc) lossless residual.
The TrueHD stream v2 is truly composed of 2 independent substreams (1 lossy + 1 lossless).
Your REVISED file detail sounds more like what I would expect: 402MiB / 26min13sec = 2093kbps.
Scott -
Fascinating, thanks so much. I don't really care about the Blu-Ray specs, since I remux my BD's in Matroska with only 3 streams (video, highest quality audio and subtitles), so the smaller, the better. Thus, I should delete the lossy data.
I'm sorry, the flac is actually 245 MiB, so it's possibly the same as the 402 MiB lossless trueHD right? I mean the flac compression being a lot more efficient. -
The only problem with deleting the lossy data is that it might have/be an optimized 2.0 downmix, whereas the lossless section might be ONLY a full 5.1/7.1 mix which may not provide an optimized 2.0 downmix (should you require a 2.0 output).
Flac MIGHT be more efficient, but it is unfortunately probably not as compatible with many STBs and other CE devices. I hope that changes in time, but presently that seems to be the case. Pity. Flac has built-in support for additional codec optimizations, and its agnostic nature and opensource make it very attractive.
If you are ONLY using those files on an HTPC of your own making, then maybe Flac would be for you.
Scott
Similar Threads
-
Is AC3 passthru completely lossless?
By KyleMadrid in forum Newbie / General discussionsReplies: 8Last Post: 4th Apr 2013, 08:58 -
Lossless AAC file editor?
By sldvd in forum AudioReplies: 1Last Post: 26th Oct 2010, 18:28 -
Does an .avi file stay lossless...?
By Slowburn in forum Video ConversionReplies: 1Last Post: 18th Mar 2009, 05:56 -
Problem with x264 lossless file creation
By algee2k8 in forum Video ConversionReplies: 4Last Post: 28th Oct 2008, 10:55 -
Lossless rotate of MJPEG file?
By kmlee in forum Video ConversionReplies: 18Last Post: 10th Oct 2008, 18:54