eac3to - audio conversion tool [Archive] - Page 11

ACrowley

31st July 2007, 09:40

Same here.

The "3dB Attenuation" is also an option in the encoder GUI, AFAIK it's for rear-channel attenuation.

I'm using 1.0.21, so ACrowley's version is probably newer. Or it's a different product than "Surcode DVD Pro".

I use Surcode DVD DTS Pro 1.0.23

I will Test it. But what i know surely is that changing the DynamicCompress Reg.Setting has no Effect!
The wave has same Peaks compared with DRC off encoded

@Madshi
You say in the Screenshot its -27db ? I dont know how to set the DialogN. Values with Registry (Hexa)Dezimal Values ?
Then i can Test it

madshi

31st July 2007, 10:18

You say in the Screenshot its -27db ? I dont know how to set the DialogN. Values with Registry (Hexa)Dezimal Values ?
Then i can Test it
Use Windows calc. Enter "-27", then switch to "Hex". You'll get 16 chars. Cut off the first 8. The last 8 chars is what you need to put into the registry.

d0ORk

31st July 2007, 10:21

When I updated to the newest eac3to.exe I'm getting this:

Q:\test>"D:\_eac3to\eac3to.exe" "Q:\test\sample-long.DD.stream.00.ea
c3" "Q:\test\sample-long.DD.stream.00.ac3" -448
E-AC3, 5.1 channels, 0:01:08, 1536kbit/s, 48khz, dialnorm: -27dB
Remove Dialog Normalization information. Please wait...
Muxing eac3 file to raw. Please wait...
Disabling DRC for Nero E-AC3 decoding...
The file size of the raw file doesn't seem to fit.
The expected file size for 16 bit is 37,50 MB.
The expected file size for 24 bit is 56,25 MB.
The real file size is 0 Bytes.

I evodemuxed the sample-long.evo and two others now.

d0ORk

31st July 2007, 10:37

Is there a download source for the previous version of eac3to.exe ? That one worked for me

ACrowley

31st July 2007, 11:20

I made a Test with Surcode dts pro 1.0.23 and the DialogNorm Registry Settings.

http://img66.imageshack.us/img66/6021/sshot3mi9.th.png (http://img66.imageshack.us/my.php?image=sshot3mi9.png)

All dts are decoded to wave with Tranzcode 0.4b

Theres absolutly no difference when i change the RegistryValues fro Dialog Normalization
And not whne changing Dynamci Compress

So it has no Effect so far i can see

madshi

31st July 2007, 11:28

madshi

31st July 2007, 11:30

tebasuna51

31st July 2007, 12:06

Maybe Tranzcode ignores these values? Does Tranzcode ignore DialNorm for AC3 decoding? I think ffdshow would be a good test decoder cause it executes DialNorm for AC3 decoding, IIRC.

Another useful test would be to check if those DTS files are bit wise identical or whether there are any bytes different. If the DTS files are 100% identical then no decoder on earth can output a different result for the 3 DTS files.

I test with SurCode DVD Pro DTS v1.0.23.

1) Initially i don't have "DIALOG NORMALIZATION" at register

2) With "DIALOG NORMALIZATION"=dword:ffffffe5 (-27)

3) With "DIALOG NORMALIZATION"=dword:ffffffe1 (-31)

the three dts are bit identical.

BTW, in encoder options I can only choose:

-Data Rate:
1536 -> "DTS BITRATE INDEX"=dword:00000001
768 -> "DTS BITRATE INDEX"=dword:00000002

-Sample Rate:
48000 -> "SAMPLING RATE"=dword:00000000
44100 -> "SAMPLING RATE"=dword:00000001

-Attenuate rear channels 3 dB:
Yes -> "SURROUND DOWNMIX"=dword:00000001
Not -> "SURROUND DOWNMIX"=dword:00000000

Exiton

31st July 2007, 12:08

Don't know why that happens. It seems to work for everybody else? Do you have the Blu-Ray/HD DVD plugin correctly licensed/registered? And are you using the latest Nero version?

Err... I updated eac3to and I'm getting the same error messages.

ACrowley

31st July 2007, 12:19

@madshi

Tranzcode applies no DRC..thats sure
But i think DialogNormalization ,as like as ffdhow

But with AC3 Filer the decoded dts are always at bit higher Peaks (a bit), SO ithink AC3 Filter applies no DRC, no Dialog Norm
However, these 3 Test Samples are identical with all decoders..no Effect when changing the Values in Registry

madshi

31st July 2007, 12:32

the three dts are bit identical.
Argh. Ok, thanks for testing!

Are you absolutely sure that DTS really knows Dialnorm at all? I rather doubt it right now. After a google search it seems to be that DTS doesn't have a Dialnorm like feature.

madshi

31st July 2007, 12:33

Err... I updated eac3to and I'm getting the same error messages.
Hmmm... Does GraphEdit.exe (renamed to Recode.exe) give out more than 0 bytes when doing "Nero Audio Decoder 2 -> Dump"?

ACrowley

31st July 2007, 12:48

Argh. Ok, thanks for testing!

Are you absolutely sure that DTS really knows Dialnorm at all? I rather doubt it right now. After a google search it seems to be that DTS doesn't have a Dialnorm like feature.

Madshi , im not sure
Look at here, its a Comparion from original dts Source decoded with all Decoders :
http://img123.imageshack.us/img123/42/46819065qo5.th.png (http://img123.imageshack.us/my.php?image=46819065qo5.png)
AC3 Filter has higher Levels..

This is a Comparions from AC3Filter Source with a Surcode reencode with -10db DialogNorm RegistryValue
http://img66.imageshack.us/img66/7116/84926638sa7.th.png (http://img66.imageshack.us/my.php?image=84926638sa7.png)

No difference.
When it would be possible to change DialogNorm via Registry for Surcode ,there must be a visible Difference

Read here (from Hardware Sound Processor)
http://www.hometheaterhifi.com/volume_4_3/meridiandts.html

DTS also does not have flags that support late evening listening (compression) or dialog normalization, while DD does.

Then it makes sense why Surcode output is identical with DRC/DialogN Reg.Values On or Off, and AC3 Filter is only a bit louder

Also i was wondering why honai told us his AC3 Tracks are louder as dts..it should be reversed.

honai

31st July 2007, 13:17

Also i was wondering why honai told us his AC3 Tracks are louder as dts

What I meant was that typically my receiver plays DD louder than DTS, but with the new eac3to the transcoded DTS tracks were louder than the same DD tracks from retail DVDs.

ACrowley

31st July 2007, 14:01

What I meant was that typically my receiver plays DD louder than DTS, but with the new eac3to the transcoded DTS tracks were louder than the same DD tracks from retail DVDs.

yes..i know
And i was only wondering why typically dts is always louder as ac3 :)
Ofcourse the decoding from eac3to 1.10 is "louder" now , because it preserves the full dynamics instead of older Version with full DRC/Dialog.
So your dts reencode has more dynamics/is louder too. I think a AC3 reencode from EAC3 is closer to the Source. Surcode isnt the best Encoder ....

d0ORk

31st July 2007, 14:07

It's working again.
I had to unpack the files all again and made sure that the old ones are overwriiten. Stupid error of me.

Err... I updated eac3to and I'm getting the same error messages.

tebasuna51

31st July 2007, 14:54

Argh. Ok, thanks for testing!

Are you absolutely sure that DTS really knows Dialnorm at all? I rather doubt it right now. After a google search it seems to be that DTS doesn't have a Dialnorm like feature.
Of course DTS have DialNorm and DRC like you can see in DTS Coherent Acoustics; Core and Extensions (http://webapp.etsi.org/action/PU/20021224/ts_102114v010201p.pdf) (pag. 14 and 17)

"Embedded Dynamic Range Flag V DYNF 1 bit
DYNF indicates if embedded dynamic range coefficients are included at the start of each sub-frame. Dynamic range correction may be implemented on all channels using these coefficients for the duration of the sub-frame."

"Dialog Normalisation Parameter/Unspecified V DIALNORM/UNSPEC 4 bits
...
The dialog normalization gain (DNG), in dB, is specified by the encoder operator and is used to directly scale the decoder outputs samples. In the DTS stream the information about the DNG value is transmitted by means of combined data in the VERNUM and DIALNORM fields (see table 5.22)."

Other question is if a encoder use these features.

madshi

31st July 2007, 15:24

Hmmmm... Thanks for the PDF, it's helpful, although a bit confusing. I'll do some more tests later when I have more time.

ACrowley

31st July 2007, 16:00

I Made more Tests.
The Surcode DTS Pro 1.0.23 Registry Settings for DialogNorm and DRC have no Effect

madshi

31st July 2007, 16:24

Ok, the next test would be to check which dialnorm value Surcode writes into its DTS encodes and to manipulate this value and see if it shows any effect with any decoders.

d0ORk

31st July 2007, 17:37

Is it possible to see a delay somewhere?
When I demux the Departed Trailer, Evodemux shows no delay. But when I encode and mux it to .mkv with MeGui the video/audio is async.

madshi

31st July 2007, 17:41

Is it possible to see a delay somewhere?
When I demux the Departed Trailer, Evodemux shows no delay. But when I encode and mux it to .mkv with MeGui the video/audio is async.
EvoDemux shows the timestamps (PTS) of the first video and audio packets. If they don't match, you can calculate the delay from that.

d0ORk

31st July 2007, 18:18

I still got a strange thing.
The source evo is 1min 8seconds. After evodemux the vc1 (I use that ending for the demuxed video) is 3min 10sec...

Rectal Prolapse

31st July 2007, 21:46

madshi, is it possible to have eac3to output a new DD+ track with dialnorm (and maybe DRC?) removed? It would be a great option for realtime playback of DD+ tracks without the step of conversion.

madshi

31st July 2007, 21:59

madshi, is it possible to have eac3to output a new DD+ track with dialnorm (and maybe DRC?) removed? It would be a great option for realtime playback of DD+ tracks without the step of conversion.
Removing the DRC information from the DD+ track would be theoretically possible, I think, but much harder than removing DialNorm. Removing DRC would mean parsing through the full DD+ structure, which is relatively complicated. And if there's a parsing bug, the DRC removing would destroy the DD+ track. Because of that I'd rather not do that. Removing Dialnorm is oh so much easier, because it's right at the beginning of each audio frame, so not much parsing is needed and there's no big danger for bugs. I could easily add an option to give out a modified DD+ track without DialNorm. Actually you can do that yourself already: Just ask eac3to to convert the DD+ track to e.g. "raw". Then during the decoding phase you can copy the "something.tmp.ddp" file. This temporary file is a full copy of the original DD+ track, just with DialNorm turned off.

But I have to say that defeating DialNorm alone doesn't make much sense, as long as DRC is still active, because DRC is the much worse problem.

Rectal Prolapse

31st July 2007, 22:01

Ah, thanks madshi! :)

For realtime playback, is Nero Audio Decoder the best solution right now?

Daodan

31st July 2007, 22:21

Sorry to interrupt this lenghty DRC/normalize talk, but does anyone know a method for decoding truhd from mt2s files? No splitter I have sees that (sonic, nero, cyberlink, etc). Only xport can demux but that doesn't help much as I don't think there is a filter source for that.

madshi

31st July 2007, 22:28

For realtime playback, is Nero Audio Decoder the best solution right now?
I think for realtime playback I'd use the Sonic one right now. It seems to be a bit less bad than the Nero one in terms of DRC. But maybe you should better use ShowTime or PowerDVD. That way you could get around DRC.

does anyone know a method for decoding truhd from mt2s files?
Haven't tried that yet cause I haven't run over a TrueHD Blu-Ray yet. But I know they are beginning to appear.

Daodan

31st July 2007, 22:37

Here's a piece: http://d01.megashares.com/?d01=2209572
Track 1 is 5.1 truhd. I found smth intersting with that, will post after a few more tests (I managed to get some sound out of it, but has some crackling in it).
Also Sonic demux sees the demuxed truhd with xport...only problem is that nero audio can't connect to that...

Rectal Prolapse

31st July 2007, 23:26

I read that Blu-ray TrueHD tracks have a DD5.1 640 kbps track interleaved with the TrueHD stream. It is NOT a core track, but a separate track.

madshi: Unfortunately, I have major PQ issues with PowerDVD, and showtime crashes on a lot of blu-ray titles. Oh well. I guess I'll stick with Sonic for now, or actually use eac3to for what you designed it for and mux my own. :)

Daodan

31st July 2007, 23:30

Yeah, that's probably why I manegd to get some sound. Having it interleaved sounds like a lol idea, why would they do that? Just because the crappy players don't support truhd?

ACrowley

1st August 2007, 09:03

Yeah, that's probably why I manegd to get some sound. Having it interleaved sounds like a lol idea, why would they do that? Just because the crappy players don't support truhd?

yeh...can be true

All Decoders decodes the 640kbps 2.0 AC3 so far i can see. Nero cant connect to it

MichalHabart

3rd August 2007, 17:41

Hello Madshi,
i still have the same problem converting pcm file to dts. But i finally find the problem, look:

D:\VOBs\!EAC3toAC3>eac3to.exe "D:\VOBs\Air Force One\bits0001.pcm" E:\English.dts -1536 -16
RAW, 5.1 channels, 2:04:34, 16 bits, 48khz
Convert the multi channel raw file to mono wav files. Please wait...
Convert the mono wav files to dts. Please wait...
Surcode sais/asks: "Invalid Wave File E:\English.L.wav.".
Surcode sais/asks: "Invalid Wave File E:\English.R.wav.".
Surcode sais/asks: "Invalid Wave File E:\English.SL.wav.".
Surcode sais/asks: "Invalid Wave File E:\English.SR.wav.".
Surcode sais/asks: "Invalid Wave File E:\English.C.wav.".
Surcode sais/asks: "Invalid Wave File E:\English.LFE.wav.".
Surcode sais/asks: "At least one valid source file must be specified to encode."
.
Pressing the Surcode "Encode" button didn't seem to work...
Pressing the Surcode "Encode" button didn't seem to work...
Closing Surcode...
Creating the DTS file failed. The multi channel wav files are valid, though.

It seems to me that he creates mono wav files in different location (in fact in same where source pcm file is) and then he is trying to find them somewhere else (where destination file should be). Can you please change it in eac3to?

madshi

3rd August 2007, 18:53

It seems to me that he creates mono wav files in different location (in fact in same where source pcm file is) and then he is trying to find them somewhere else (where destination file should be). Can you please change it in eac3to?
Yey, I'm glad we found the cause of the problem! :)

madshi

3rd August 2007, 18:58

eac3to 1.11 released

http://madshi.net/eac3to.zip

(1) bugfix: (L)PCM -> DTS encoding automation failed when source and destination folders differed
(2) added: new "-allowDnr" switch allows Nero's audio decoder to apply DNR
(3) added: new "-keepDialnorm" switch disables removal of E-AC3 dialnorm information

MichalHabart

3rd August 2007, 20:32

Yey, I'm glad we found the cause of the problem! :)

Me too and thanks for fast fix. Will try tomorrow :)

Edit: Now everything is working properly. Once again, thank you madshi for excellent work.

ACrowley

5th August 2007, 10:09

madshi

5th August 2007, 16:30

Ouch, why did I write "DNR"? I guess I was sleeping. Of course I meant DRC!! Will fix that in the next build.

madshi

6th August 2007, 00:09

eac3to 1.12 released

http://madshi.net/eac3to.zip

(1) The tools "flac.exe", "aften.exe" and "sox.exe" are now distributed in the eac3to zip to make your life a bit easier.
(2) Correct channel mapping for 7.1 LPCM tracks is default now.
(3) The new option "-down6" allows downconverting of 7.1 tracks to 5.1 (this option is activated by default for AC3 and DTS encoding).
(4) I'm shipping a modded "flac.exe" with eac3to now, which has no problems with 2GB file output, anymore!! As a result the flac encoding by eac3to should now be more or less perfect.
(5) Fixed a number of bugs.

Btw, those Blu-Ray soundtracks which say 6.1 LPCM are actually 7.1. Ok, two channels are identical, that's probably why the cover sais 6.1. But the LPCM track contains full audio data for 8 channels. So use the "-8" switch for 6.1 LPCM tracks.

SvT

6th August 2007, 00:19

I can read 5 more reasons to download this program !!!!!!!!!!!!!!

Thank you very much !!!!!!

ACrowley

6th August 2007, 07:45

@madshi

So is 1.12 a "full Vesrion" of eac3to ?
Or do we need all the older Files like the .pas Files etc ?

Nice work on flac.exe..i will test it

greets

madshi

6th August 2007, 08:08

So is 1.12 a "full Vesrion" of eac3to ?
Or do we need all the older Files like the .pas Files etc ?
It's a "full version". The pas etc files are the source code of eac3to, which is now in the sub folder "sources". You can delete this "sources" folder if you don't need/want to have the source code of eac3to. You should keep the other subfolders and files. Ok, you can delete sox, if you don't plan to downconvert anything from 24bit to 16bit. And you can delete the aften/flac sub folders, if you don't ever plan to use eac3to to encode ac3/flac.

ACrowley

6th August 2007, 10:43

ok, thx

On more suggestion, maybe you should note in eac3to.exe that Sonic Decoder will apply DRC and DialogNormalization, that you can only avoid it with NeroDecoders.
Sure, we here know it, but for other Users its more obvious

And, for bluray lpmc to wave , do i need still the -bluray parameter ?
Or works the Auto CHmapping for lpcm too ,as like as for eac3/TrueHD(when i want standard L=0 R=1 C=2 LFE=3 SL=4 SR=5 ) ?

EDIT:
Nice work!! lpcm24bit to flac 24bit works perfect! With Timeline and no 2gb Restriction!
Older eac3to Version outputs always a 24bit flac which was corrput8 to big, only noise) but now it works !

Ich wüsste nicht wo die encoder ohne deine Arbeit wären im HDDVD/Bluray Audio Sektor...weiter so!

madshi

6th August 2007, 10:58

On more suggestion, maybe you should note in eac3to.exe that Sonic Decoder will apply DRC and DialogNormalization, that you can only avoid it with NeroDecoders.
Sure, we here know it, but for other Users its more obvious
E-AC3 DialNorm is fully defeated by eac3to, doesn't matter which decoder you're using. But I should probably add a warning about DRC with Sonic's decoder, I agree.

And, for bluray lpmc to wave , do i need still the -bluray parameter ?
Or works the Auto CHmapping for lpcm too ,as like as for eac3/TrueHD(when i want standard L=0 R=1 C=2 LFE=3 SL=4 SR=5 ) ?
Channels should automatically be mapped correctly for every kind of input, unless you have a strange input. For Blu-Ray PCM I expect the data to be demuxed by xport or TsRemux, though. If you use the GraphEdit dump method to demux a Blu-Ray PCM track, you'll have to set different parameters in eac3to, e.g. endian will be "little" instead of big etc...

ACrowley

6th August 2007, 18:42

E-AC3 DialNorm is fully defeated by eac3to, doesn't matter which decoder you're using. But I should probably add a warning about DRC with Sonic's decoder, I agree.

Ah, Ok....sure ea3to "reverse" DialogN itselft...
Sonic Decoder outsinde eac3to applies starnge Mix of DRC and or DialogNorm+gain

Channels should automatically be mapped correctly for every kind of input, unless you have a strange input. For Blu-Ray PCM I expect the data to be demuxed by xport or TsRemux, though. If you use the GraphEdit dump method to demux a Blu-Ray PCM track, you'll have to set different parameters in eac3to, e.g. endian will be "little" instead of big etc...

Yes..works perfect

But thats not my Experience....All LPCM Tracks i have demuxed with SonicHDDemxuer/graphedit are working with BigEndian.
Shouldt eac3to detects autom wich Endian it is ?
When i set no Endian Switch the Output is ok

And your tweaked flac.exe...is it working only eac3to intern or can i use the 2gb workarounds outside eac3to ?

madshi

6th August 2007, 19:18

Sonic Decoder outsinde eac3to applies starnge Mix of DRC and or DialogNorm+gain
I'm quite sure it's no DialNorm with the later eac3to versions. But probably DRC + gain.

But thats not my Experience....All LPCM Tracks i have demuxed with SonicHDDemxuer/graphedit are working with BigEndian.
Strange. Well, that's just fine.

Shouldt eac3to detects autom wich Endian it is ?
I can try to add auto Endian detection.

And your tweaked flac.exe...is it working only eac3to intern or can i use the 2gb workarounds outside eac3to ?
Of course that tweaked flac.exe works just fine on its own, but I will not tell you that, because I want eac3to to gain world domination... :)

P.S: That said, I've fixed the *output* 2GB limitation. The *input* limitations are still valid. You can work around them by feeding flac.exe though stdin (that's what eac3to is doing), but it's not very comfortable to do that manually.

ACrowley

6th August 2007, 19:29

Ok thx
Yes ,youre Right. Sonic Output looks more like DRC+Gain, very bad

Ah, Ok...It works everytime for me because Big Endian is default and all my LCPM are Big Endian as i say even demxued with Graphedit
Autodetection would be nice.
I know that SonicHDDemxuer output should be little Endian..But its reversed, with little Endian i get garbage on 24bit lpcm
Maybe the older 4.2 Demxuer outputs little E.

And i was talking about flac "output limitation"

yonta

6th August 2007, 19:44

-down6 downconvert raw data from 8 to 6 channels
-down16 downconvert raw data to 16 bit
-96000 TrueHD output or PCM file is 96 kHz
-8 TrueHD output or PCM file is 8 channels
-16 TrueHD output or PCM file is 16 bit
-24 TrueHD output or PCM file is 24 bit (default)
-little PCM file is little endian
-big PCM file is big endian (default)

I have a 24bit-5.1ch-48kHz-big.endian lpcm file and I tried converting to a flac.
The info about the lpcm file is from the output of xport

LPCM Audio Mode = 3/2+lfe
LPCM Audio Bits/sample = 24
LPCM Audio Sample Rate = 48000

So, I figured I don't have to use any option above and used this simple command line.

eac3to "the.wild.LPCM24.pcm" "The.Wild.flac"

This is what eac3to says;
Tried to find out the bitdepth automatically, but the algorithm failed.
Please use either "-16" or "-24" to tell eac3to which bitdepth the source has.

probably auto-detection is default instead of "-16" or "-24"?