eac3to - audio conversion tool

[madshi]

Originally Posted by Inventive Software

Small feature request if possible: merging 6 separate FLACs to a 5.1 FLAC. Currently, eac3to doesn't do this, nor does FLAC.exe.

Merging 6 files to one is generally possible but it would cost me some programming time. Because of that I don't really have it on my to do list right now, I'm sorry. I think there are more important things that I need to add first.

[madshi]

Originally Posted by tjf

This mlp track ripped from DVD-A tubular bells gives an error

[mlp @ 68A4E2E0]Lossless check failed - expected 0, calculated e7

Is this a bug in libav?

Yes, this seems to be a bug in the libav decoder. I've forwarded your sample to the decoder programmer. Hopefully we'll get a fix sooner or later. For now you can use the Nero decoder. Or you can keep using libav. The bug seems to affect only one audio block (just a few milliseconds in the audio file) and the difference between faulty decoding (libav) and correct decoding (Nero) isn't big. At least it's not visible in a WAV form editor.

[madshi]

Originally Posted by yfed

actually, my mistake, it does consist of two files..

BDMV/STREAM/00000.m2ts 19.36 GB
BDMV/STREAM/00001.m2ts 22.51 GB

I joined them into one big file using copy /b command prompt. Maybe that could've caused delay?

The delay could be caused by this. Would it be possible for you to upload the last 20MB of 00000.m2ts and the first 20MB of 00001.m2ts (as separate files, please)? There will not be any benefit for you doing that right now cause eac3to does not handle m2ts files yet. But I'd like to put those samples on storage for later. They might help me figuring out how to handle audio/video well in situations like yours.

[madshi]

Originally Posted by Thunderbolt8

havent had a trueHD track with the latest version for a while, so I assume this is normal now (the mlp@ end of stream message)?

[code][mlp @ 68A4E2E0]End of stream indicated
[mlp @ 68A4E2E0]End of stream indicated

This is normal and correct.

[madshi]

Originally Posted by nincollector

just a general question. will or does an ac3 made from a truehd source sound better than an ac3 made from the eac3 source or is it the same?

Using a TrueHD source should normally sound a little bit better.

[madshi]

Originally Posted by sparknburn

I love this app but has it gotten slower processing streams? I took over 2.5 hrs to convert an eac3 stream to ac3. Or am I crazy.

As nautilus7 already hinted, the latest eac3to version is supposed to be very little bit slower with E-AC3 decoding. But 2.5 hours is really a lot. Do you have a single or dual core CPU? The performance hit of the latest eac3to version is only very little on dual core CPUs, but it will double processing time on single core CPUs.

[madshi]

Originally Posted by jhoff80

I'm getting the same issue with both discs I've tried ripping so far, the unencrypted NIN "The Downward Spiral" and the encrypted NIN "With Teeth" dualdisc DVD-a discs.

Can you please upload a sample of each track? Please make sure that the sample shows the same problem before uploading them. I'll then forward the samples to the MLP/TrueHD decoder programmer. The more samples we collect the better.

[Thunderbolt8]

Originally Posted by madshi

The delay could be caused by this. Would it be possible for you to upload the last 20MB of 00000.m2ts and the first 20MB of 00001.m2ts (as separate files, please)? There will not be any benefit for you doing that right now cause eac3to does not handle m2ts files yet. But I'd like to put those samples on storage for later. They might help me figuring out how to handle audio/video well in situations like yours.

as I said, the delay should actually only be 32ms max, in case of DD+ and if case of lpcm, dtshd or truehd even less and not noticeable at all, because its only 1 part which was joined to the 1st one.

[Snowknight26]

eac3to v2.14
eac3to.exe  FEATURE_1.EVO+FEATURE_2.EVO 4: kingkong.eac3
EVO, 2 video tracks, 4 audio tracks, 3:07:23
1: Joined EVO file
2: VC-1, 1080p24 /1.001
3: VC-1, 480p30 /1.001
4: E-AC3, 5.1 channels, 1536kbit/s, 48khz, dialnorm: -27dB, 1001ms
5: E-AC3, 5.1 channels, 768kbit/s, 48khz, dialnorm: -27dB, 1013ms
6: E-AC3, 5.1 channels, 768kbit/s, 48khz, dialnorm: -27dB, 1013ms
7: E-AC3, 2.0 channels, 192kbit/s, 48khz, dialnorm: -27dB
Extracting audio track number 4...
Removing dialog normalization...
Applying (E-)AC3 delay...
Creating/writing file "kingkong.eac3"...
Video track 2 contains 269566 frames.
Video track 3 contains 324001 frames.
eac3to processing took 1 hour, 12 minutes.
Done.

C2Q and a fast HDD, shouldn't take that long. I could have sworn it was much faster before the parallel reencoding.

[nautilus7]

This is another thing (evo demuxing). We are talking about e-ac3 decoding.

[K-rnivoro]

Weird thing, maybe madshi can help me out.
eac3to seems to work ok for me, it was able to convert EAC3 stream form Troy to AC3 (with Nero of course..). so that's tell me everything is in order.
Thing is, Eac3to can't detect eac3 stream from a rip of "Fearless" movie, it says "The format of the source file could not be detected"
I though i was doing something wrong BUT Sonic Cinemaster is able to play back that .eac3 and also detects it as a DD+ 6 Channles (at the Sonic HD Demuxer window) so the file it has the right format.
Any clue? Want me to upload a fragment of it?

[nautilus7]

I believe there is some garbage in the start of your track.

The first 5 MB should be ok.

[jhoff80]

Originally Posted by madshi

Can you please upload a sample of each track? Please make sure that the sample shows the same problem before uploading them. I'll then forward the samples to the MLP/TrueHD decoder programmer. The more samples we collect the better.

I had the issues with basically every track on each of the discs with some tracks having a long string of them, maybe 30 or 40 of those lossless check failed messages. I was going to attempt to upload them but didn't know of any way to copy only a few megs of each file. I thought when I had XP that I could just copy a file and while it was being copied over make a copy of the beginning of that file, but in Vista this doesn't seem to work.

[nautilus7]

Use a hex editor like HxD or see this post.

[K-rnivoro]

Originally Posted by nautilus7

I believe there is some garbage in the start of your track.

The first 5 MB should be ok.

OK, just tell me how to do it...where do I have to upload the sample ?

[nautilus7]

Instructions on how to cut a sample, look at my previous post.

You can upload the sample here.

[K-rnivoro]

Originally Posted by nautilus7

Instructions on how to cut a sample, look at my previous post.

You can upload the sample here.

There you go:
http://www.sendspace.com/file/7t735g

and just in case:
http://www.sendspace.com/file/7t735g

I've tried woth the "tail section" of the track, same luck... check it out, I'll be waiting for news..

[nautilus7]

Both links are the same... Is this the start of the track?

[K-rnivoro]

Originally Posted by nautilus7

Both links are the same... Is this the start of the track?

Yes, is the same link and it's the start of the file. I just put it twice to be sure (never paste a link before)
I meant I tried to feed eac3to with the tail of the track and it keeps not to recognize it as a eac3 track.

[nautilus7]

Never mind. I fixed it. As i suspected there were some garbage in front of the track (the real audio info).



Open the original track with a hex editor (you can use the free HxD) and delete the 1st 416 bytes. Then you 'll be ok. Make a back up of the track before anything!

EDIT: Just to explain what i did. In your track every frame (eac3 consists of frames) is 1024 bytes long and starts with 0B 77 01 FF. You can easily see that the 1st one is not 1024 bytes, while the rest of them is. Why is it garbage and not just a cut/half frame? Because is completely different from the next frames (the text on the right).