eac3to - audio conversion tool [Archive] - Page 10

ACrowley

27th July 2007, 10:00

I got overflows even when set -6ch in besweet ... (decode to 6 mono waves). So the resulted wavs' waveform seems a bit "flat" where they reached the "maximum" ... :(

So far i know it has no Effect. Its only a Warning.
But maybe it can cause Problems when you donwmix 5.1 to stereo.
But when you decode all channles it will not have any Effect

Anway, The Optimum would be to decode AC3 to wave with certified Dolby Decoder.
To Bad Nunedo outputs DRC without any Decoder Options ( 1000$!)
Soft Encode/Digigram has a Dolby Decoder too ( Dolby Engine too)..but as i say above i dont know wich Option means 100% "No" DRC/D.N

Besweet/Foobar is fine so far

tebasuna51

27th July 2007, 13:25

So far i know it has no Effect. Its only a Warning.
But maybe it can cause Problems when you donwmix 5.1 to stereo.
But when you decode all channles it will not have any Effect
The typical azid warning:
W7: Downmix overflow (1: +0dB)
is not related with channels downmix and is produced when the decoder, translating from frequency domain to time domain, reach a value upper to 0 dB and can't be translated exactly.

Is not a problem and must be assumed like others imperfections in the encoder/decoder process with lossy formats.

Soft Encode/Digigram has a Dolby Decoder too ( Dolby Engine too)..but as i say above i dont know wich Option means 100% "No" DRC/D.N
But when RF Remod is similar to heavy(=loud) it should be the Full Dynamic Range, it makes Sense. Lineout= DRC+DN
- Custom analog: DRC only
- Custom digital: DRC + DN
- Line Out: also DRC + DN
- RF Remod Mode: DRC + Gain (+8 dB in my sample but I don't know the reason)

Then RF Remod Mode is not Full Dynamic Range.

ACrowley

27th July 2007, 13:50

The typical azid warning:
W7: Downmix overflow (1: +0dB)
is not related with channels downmix and is produced when the decoder, translating from frequency domain to time domain, reach a value upper to 0 dB and can't be translated exactly.

Is not a problem and must be assumed like others imperfections in the encoder/decoder process with lossy formats.

- Custom analog: DRC only
- Custom digital: DRC + DN
- Line Out: also DRC + DN
- RF Remod Mode: DRC + Gain (+8 dB in my sample but I don't know the reason)

Then RF Remod Mode is not Full Dynamic Range.

Yes, youre Right, thats waht it looks like :

http://nwgat.net/woot/files/57/other/soft_thumb.png
(http://nwgat.net/woot/files/57/other/soft.png)

I made AC3 reencodes from all waves . These Settings should keep alwys Source Volume Level:
http://forum.doom9.org/showthread.php?t=118759&highlight=Sony+Vegas

The AC3 Reencode from Azid Waves with disabled DRC/DN sounds perfect :) Same Dynamics/Wavestructure/RMS Value ,compared with the AC3 Source.

Reencodes from SoftEncode "RF Mode Wave" plays at lower Volume in loud Dynamic Parts but a little bit gained...so youre Right
I cant find any Soft Encode decoding Mode with all disabled and full Dynamic Range..strange

At least you can see in the Screenshot NeroAudioDecoder decodes AC3 with full DRC+DialogNormalization.
No Effect from Showtime Options. Sonic applies DRC and or DialogN. too.

However , the AC3 decoding Thing is cleared for me. THX :)

madshi

27th July 2007, 16:47

@ACrowley, could you make that test ac3 file available, please?

ACrowley

27th July 2007, 17:44

Once again with TrueHD ,This is from the 300 HDDVD :

http://nwgat.net/woot/files/57/other/300_thumb.png (http://nwgat.net/woot/files/57/other/300.png)

The Decoded Wave has less dynamics as like as the DVD AC 448kbps

I tried to change the NeroShowtime Options, but no Effect!

@ACrowley, could you make that test ac3 file available, please?
I used this AC3 file for my last Test.
This AC3 Sample is perfect for a Testrun because it has a lot of dynamic Sounds and a Speech Part

http://rapidshare.com/files/45393728/untouched_AC3_Reference_Sample.ac3.html

EDIT:
...can be true maybe its only DalogNormalization
I will compare it with Azid and DialogNormalization

madshi

27th July 2007, 18:46

Nero applies DRC on TrueHD Decoding. The Decoded Wave has less dynamics as like as the DVD AC 448kbps
Are you sure that it's not just lower volume (dialnorm?)? Looking at the screenshot it seems to me that everything is just not as loud as in the AC3 WAV.

In your previous tests with the Serenity AC3, the first half of the AC3 was significantly compressed with the DRC decoders, but the 2nd half was more or less identical to the non DRC decoders. So there it was clearly visible that the DRC compresses some parts of the track but not everything. In your 300 test I don't see this behaviour. It seems to me that the whole TrueHD decode is just of lower volume.

Or am I wrong?

I used this AC3 file for my last Test.
This AC3 Sample is perfect for a Testrun because it has a lot of dynamic Sounds and a Speech Part
Yeah, that's a very good test AC3. Thanks!

madshi

27th July 2007, 19:14

Here are my test results with Nero. I've tried the Nero filters through GraphEdit. And I've also checked what Nero outputs if you use Nero "ShowTime". It seems that the DRC setting in ShowTime does affect AC3 decoding, but only inside of ShowTime. When using the Nero filter in GraphEdit, AC3 decoding seems to have very strong DRC applied. Here are the results:

http://madshi.net/TrueHD/neroac3_.png (http://madshi.net/TrueHD/neroac3.png)

First wav is "Nero filter in GraphEdit". Second wav is "ShowTime with DRC turned on to 50%". Third wav is "ShowTime with DRC turned off". The third wav looks fine to me. It seems to have dialnorm applied, but no DRC.

Wilbert

27th July 2007, 19:27

Try changing the Sonic filter settings (see Control Panel -> CinePlayer).
Any idea what the environment setting is doing in the audio tab?

madshi

27th July 2007, 20:58

Any idea what the environment setting is doing in the audio tab?
Nope.

ACrowley

28th July 2007, 04:42

Are you sure that it's not just lower volume (dialnorm?)? Looking at the screenshot it seems to me that everything is just not as loud as in the AC3 WAV.
Or am I wrong?

Yeah..youre Right ! Nero applies "only" DilaogNormalization on theTrueHD ! No DRC
However, the Difference is big compared with azid Full dynmics without DCR/DialogN

http://nwgat.net/woot/files/57/azid2_thumb.png
(http://nwgat.net/woot/files/57/azid2.png)

I will make more Test if i can get a difference when i change Showtime Options. But in my last Tests i cant

We could get a final Result when somebody compare 300 HDDVD True HD it with the 300 Bluray LPCM
---
@Madhsi
I ask you again.. ? How do you dump the wave from my AC3 Test sample in your Test without DRC ? Your Graph etc ?
I cant get any effect when changing Showtime Options.
I mean its a essetial Information...

madshi

28th July 2007, 09:55

ACrowley

28th July 2007, 10:30

Ok, it seems that Nero does do DialNorm in the TrueHD decoder. That explains nicely why Nero outputs a 16bit TrueHD track with full 24bit of information!

Two questions for Dolby experts:

(1) Can DialNorm be properly undone on the final WAV?

(2) Does the (E-)AC3 / TrueHD data contain information about whether (and how much) DialNorm shall be executed? If so, maybe we can manipulate the (E-)AC3 / TrueHD data before sending them to the decoder?

Mh, Im not sure.
But i think not after decoding to wave.
When the wave is compressed at lower Volume its bad. What will you do ?
The best is to avoid any DRC/DialogN...

But i think this is a good Conclusion :
Heres a Comparions with 300 HDDBD/Bluray (AC3,TrueHD,EAC3) Audio and it looks like :

1. "AC3" BluRay 640kbps untouched / to Compare it with : No DRC ,No DialogNorm
2. NERO "TrueHD" / applies DialogNorm, No DRC
3. NERO "EAC3" / applies DRC and DialogNormalization, even with changed Showtime Settings
4. SONICinemaster4.3 "EAC3" /looks like applies DRC and DialoN. + Gain
5. "AC3" Bluray 640kbps untouched / to compare it with : only DialogNorm.

http://nwgat.net/woot/files/57/other/sshot-3_thumb.png
(http://nwgat.net/woot/files/57/other/sshot-3.png)

Conclusion is

NeroAudioDecoder2 :
- not good for EAC3 cause it applies full DRC and DIalogNormalization
- Better then nothing for TrueHD cause its "only" DialogNormalization
- I find no Way to change Nero Settings via Showtime or Registry to change it

SonicCinemasterAudioDecoder4.2
- not so low Volume as Nero, but looks like DRC and or DialogNormalization with "DolbyAuidoGain" buuhh
- TrueHD decoding still not work without Backroundnoise with 4.3
- cant see any Changes with different Settings in Sonic System Panel

About the AC3 to wave decoding :
-Azid/Foobar/NiscAudio can decode it fully without DRC/DialogN, so i use it to keep Full Dynamics
-Steinberg Nuendos cert. Dolby Decoder applies DRC and DialogN. No Way to change Dolby Decoder Options
-SoftEncode cert Dolby Decoder works very strange. Cant find any untouched Output.

So the Target is to find a Way to change Decoding Options for Nero and Sonic. to get EAC3 /TrueHD without DRC/DialogN.
And ofcourse hoping for a fast ffmpeg EAC3 Decoder/Ac3Filter Release

madshi

28th July 2007, 10:59

Conclusion is

NeroAudioDecoder2 :
- not good for EAC3 cause it applies full DRC and DIalogNormalization
- Better then nothing for TrueHD cause its "only" DialogNormalization
- I find no Way to change Nero Settings via Showtime or Registry to change it

SonicCinemasterAudioDecoder4.2
- not so low Volume as Nero, but looks like DRC and or DialogNormalization with "DolbyAuidoGain" buuhh
- TrueHD decoding still not work without Backroundnoise with 4.3
- cant see any Changes with different Settings in Sonic System Panel

About the AC3 to wave decoding :
-Azid/Foobar/NiscAudio can decode it fully without DRC/DialogN, so i use it to keep Full Dynamics
-Steinberg Nuendos cert. Dolby Decoder applies DRC and DialogN. No Way to change Dolby Decoder Options
-SoftEncode cert Dolby Decoder works very strange. Cant find any untouched Output.
Yes, I think you've nailed it. However, there's one more option: Namely playing AC3 and E-AC3 inside of ShowTime and capturing the output (possible with a little hacking). This should probably help getting rid of DRC for AC3 and E-AC3 with Nero. However, I don't know yet how to get rid of DialNorm.

-----

I'm still wondering whether DialNorm can't be undone afterwards. I mean it's "only" a volume change. Can't we turn up volume again in the PCM/WAV file to undo that change? I'm not sure how much harm would be done by this "volume down + volume up"?

ACrowley

28th July 2007, 12:11

Yes, I think you've nailed it. However, there's one more option: Namely playing AC3 and E-AC3 inside of ShowTime and capturing the output (possible with a little hacking). This should probably help getting rid of DRC for AC3 and E-AC3 with Nero. However, I don't know yet how to get rid of DialNorm.

-----
I'm still wondering whether DialNorm can't be undone afterwards. I mean it's "only" a volume change. Can't we turn up volume again in the PCM/WAV file to undo that change? I'm not sure how much harm would be done by this "volume down + volume up"?

yep....

However. AC3 decoding is not necesarry because azid/Foobar works optimal.
It was only own Target to clear it because i encode a lot of
AC3s.

Hmm...The DialogNormalization is not only a simple overall Levelchange ,isnt it ? Its a dynamic Thing too?
So a simple Volume + would not help..dont know

heres a good EAC3 Sample for Testing
http://rapidshare.com/files/45531448/300_EAC3_640kbps.rar.html

Can you try Showtime Output Setting "Passtrough" with your Tweak?
"Passes Audio to your Soundcard without further processing"

And
1. What is your Showtime Capture/Dump Tweak ?
2.in your other Test you use PowerDVD for decoding TruehD ? how ??

tebasuna51

28th July 2007, 13:21

Hmm...The DialogNormalization is not only a simple overall Levelchange ,isnt it ? Its a dynamic Thing too?
So a simple Volume + would not help..dont know

Apply the Dialog Normalization is always a fix attenuation for all the stream. Each ac3 frame (32 ms. for 48 KHz.) have a Dialog Normalization value but must be the same for all frames in the stream.

The DRC attenuation/gain value can vary for each block (5.33 ms., 6 blocks per frame).

A gain of (31 - DialogNorm) dB can restore the original volume and must be used to reencode to ac3 using the same DialogNorm and DRC settings.

madshi

28th July 2007, 13:27

Apply the Dialog Normalization is always a fix attenuation for all the stream. Each ac3 frame (32 ms. for 48 KHz.) have a Dialog Normalization value but must be the same for all frames in the stream.
Do you happen to know which bits in the AC3 header tell the decoder how much Dialog Normalization is needed? Do you think we can just modify those bits to get rid of DialNorm?

A gain of (31 - DialogNorm) dB can restore the original volume and must be used to reencode to ac3 using the same DialogNorm and DRC settings.
So DialNorm 31 means no change at all, right?

ACrowley

28th July 2007, 14:02

Do you happen to know which bits in the AC3 header tell the decoder how much Dialog Normalization is needed? Do you think we can just modify those bits to get rid of DialNorm?
So DialNorm 31 means no change at all, right?

@tebasuna51

When i "reencode" a AC3 this should be correct : (?)

1. AC3 to wave with AZID , no DRC/DialogN.
BeSweet.exe -core( -input "E:\test.ac3" -output "E:\test.wav" -6ch )

2. Reencode with Sony Vegas Prof Dolby Encoder v7 ( or a other cert. Dolby Encoder) and these Settings :
http://forum.doom9.org/showthread.php?t=118759&highlight=Sony+Vegas

-31db DialogNormalization, no DynamicRangeComp ,no other Settings
It will guarantee you the volume level of the output AC3 will be the same as the original Wave?

So i dont have to care about Source AC3 Values because i get the Source Wave Levels.
Must be correct because i get the same Wavestructure and RMS Value as like as the AC3 Source (based on azid decoding)

The other,full Dolby Complaint, Method (when i want produce AC3) is to use a DRC Profile in Combi with a proper DialogNom. Value.
To get this correct DialogNormalization Value, i have to run a RMS Calculation and set it for the DialogN...right ?

tebasuna51

28th July 2007, 18:58

Do you happen to know which bits in the AC3 header tell the decoder how much Dialog Normalization is needed?
Bits 55-59 for 2/0 (stereo) ac3
Bits 57-61 for 3/2 (surround) ac3
Five bits make a value between 0 and 31
Do you think we can just modify those bits to get rid of DialNorm?
Yes, but you need recalculate also the CRC values (bytes 3-4 at the begin of the frame, I don't remember now if we need recalculate the ending CRC frame also) if you want the decoder accept the frame.
So DialNorm 31 means no change at all, right?
Ok, all these bits to 1.

Maybe you can read this post (http://forum.doom9.org/showthread.php?p=964460#post964460) and related.

madshi

28th July 2007, 19:03

Bits 55-59 for 2/0 (stereo) ac3
Bits 57-61 for 3/2 (surround) ac3
Five bits make a value between 0 and 31

Yes, but you need recalculate also the CRC values (bytes 3-4 at the begin of the frame, I don't remember now if we need recalculate the ending CRC frame also) if you want the decoder accept the frame.

Ok, all these bits to 1.

Maybe you can read this post (http://forum.doom9.org/showthread.php?p=964460#post964460) and related.
Thank you!! :)

How about mono AC3? And do you know the bits for E-AC3 and TrueHD, too?

tebasuna51

28th July 2007, 19:33

When i "reencode" a AC3 this should be correct : (?)

1. AC3 to wave with AZID , no DRC/DialogN.
BeSweet.exe -core( -input "E:\test.ac3" -output "E:\test.wav" -6ch )

Correct, now you have the full dynamic range of original source.

2. Reencode with Sony Vegas Prof Dolby Encoder v7 ( or a other cert. Dolby Encoder) and these Settings :
http://forum.doom9.org/showthread.php?t=118759&highlight=Sony+Vegas

Like Skelsgard say in the post:
"This project will guarantee you the volume level of the output AC3 will be the same as the original WAV"

But is not Dolby compliant, with the settings DialNorm=31, DRC=None you obtain a kind of encode similar to ogg, mp3, ... without the 'advantages' of Dolby normalization (the dialogs must be at -31 dB, great dynamic range but low RMS ...).

The problem is listen an ac3 and after a CDA or mp3 with modern music (low dynamic range with high RMS). This is the origin of the 'low volume' problem of ac3 streams.

-31db DialogNormalization, no DynamicRangeComp ,no other Settings
It will guarantee you the volume level of the output AC3 will be the same as the original Wave?

There are decoders that can be instructed to ignore the BSI values but we can't do anything about that. For instance the Ac3Filter have a DRC function than ignore the values in the ac3 stream and analyze the signal to make a 'dynamic' DRC.

The other,full Dolby Complaint, Method (when i want produce AC3) is to use a DRC Profile in Combi with a proper DialogNom. Value.
To get this correct DialogNormalization Value, i have to run a RMS Calculation and set it for the DialogN...right ?
Ok. You can read the Sticky: GUIDE: How To Properly Encode Dolby Digital Audio (AC3) (http://forum.doom9.org/showthread.php?t=56020)
and remember select only a fragment with representative dialog volume (maybe only in central channel) to get the rms value for DialNorm

tebasuna51

28th July 2007, 20:01

How about mono AC3?
For mono AC3 the DialNorm bits are 51-55
And do you know the bits for E-AC3 and TrueHD, too?
About E-AC3 you can read the Anex E of Digital Audio Compression Standard (AC-3, E-AC-3) Revision B (http://www.atsc.org/standards/a_52b.pdf) (still in my 'todo' list)

About TrueHD, I don't know documentation.

Edit: For E-AC3 DialNorm is always bits 46-50, don't have CRC at the beginning but yes at the end.

madshi

28th July 2007, 23:03

@tebasuna51, thanks very much again for the helpful information! Do you happen to have a link or maybe even some code about how to recalculate the (E-)AC3 CRC?

tebasuna51

29th July 2007, 01:43

@madshi
Sorry but today is my first approach to eac3.

honai

29th July 2007, 07:07

@madshi

Could you please update your first post with the findings concerning DRC/Dialog Normalization in Nero Audio Decoder? Thanks.

ACrowley

29th July 2007, 10:24

Correct, now you have the full dynamic range of original source.
Like Skelsgard say in the post:
"This project will guarantee you the volume level of the output AC3 will be the same as the original WAV"

But is not Dolby compliant, with the settings DialNorm=31, DRC=None you obtain a kind of encode similar to ogg, mp3, ... without the 'advantages' of Dolby normalization (the dialogs must be at -31 dB, great dynamic range but low RMS ...).

The problem is listen an ac3 and after a CDA or mp3 with modern music (low dynamic range with high RMS). This is the origin of the 'low volume' problem of ac3 streams.

There are decoders that can be instructed to ignore the BSI values but we can't do anything about that. For instance the Ac3Filter have a DRC function than ignore the values in the ac3 stream and analyze the signal to make a 'dynamic' DRC.

Ok. You can read the Sticky: GUIDE: How To Properly Encode Dolby Digital Audio (AC3) (http://forum.doom9.org/showthread.php?t=56020)
and remember select only a fragment with representative dialog volume (maybe only in central channel) to get the rms value for DialNorm
yep.
I use AdobeAudition 2 for RMS calculation to get DialogNormalization Value .
I use Adobe Audition.

Wich Method would "you" pers recommend ?
- Measure RMS ,set it for DialogNormalization with DCR Profile "Film".
-Or all disabled , -31db always for DialogNorm

1 AC3/EAC3 to AC3 "re"encoding
2. TrueHD/LPCM/DTS encoding to AC3 ?

I mean most guys i know use -31db Dialog Norm with no DRC.
I think eac3to/aften use it too.
I can remember there where some heavy discussions in other Forums.
Ofcourse, as i say the -31db/no DRC Method isnt DolbyCompliant, but it will guarantee Source Wave Volume Level.

Comparison from Source/Dolby Conform/ -31db not Conform
http://img509.imageshack.us/img509/6803/dolbykj2.th.png (http://img509.imageshack.us/my.php?image=dolbykj2.png)
they all have the same RMS Values and Wavestructure

I pers use -31db with no DRC when i reencode from AC3/EAC3.
When i compare the wave from Source AC3 with the reencode. it has the same RMS Values and Wavestructure
And i play AC3 on PC via AC3Filter (no DRC enabled) or SPDIF to my Surround Receiver/Speaker

For channel upmixes( more a "producing" as reencodig) RMS for DialogN+DRC
But i think maybe its better to encode Dolby Compliant from lossless HD Audio ?
So, what is your Advice ?

THX

tebasuna51

29th July 2007, 13:22

@ACrowley
There are 3 options:

1) Encode to ac3 with DialNorm 31, DRC None.
Recommended for users with good audio equipment and without problems with childs or neighbors.

2) Encode with the proper DialNorm and DRC.
Recommended for users with control over the decoder (PC, good receivers).

3) Apply the DRC, Maximize (Peak method) and encode to ac3 with DialNorm 31, DRC None.
Users with low cost audio equipments and with the "low volume ac3 problem"

Obviously the second method is more flexible but...

- The Dialog Normalization concept from Dolby is useless if you listen TV audio, CDA, mp3 with the same equipment. Actually the -31 dB ac3 volume is far off the typical volume of others sources.

- The Dynamic Range Compression concept is very interesting with sources with wide dynamic range, is useless for modern music with high volume but short dynamic range.

- If the encoder to use is Aften, remember: the actual DRC routines are in beta stage with differences (http://forum.doom9.org/showthread.php?p=893111#post893111) with the behavior of commercial encoders.

ACrowley

29th July 2007, 13:41

Thank you

- No, i encode not with Aften, i use SonyVegas Dolby Digital Pro Encoder v7 or STeinberg Nuendos DD encoder.
Both are certified Dolby Encoder

The most AC3 Tracks i use are from movies has a wide Dynamic (HDDVD,Bluray)
Currently and in the Past i use Method 1.

Actually the -31 dB ac3 volume is far off the typical volume of others sources.

When i compare the Source (always azid decoded no DRC;DialogN) its looks exactly as like as encodes from both Methods and with the same RMS Values
http://img509.imageshack.us/img509/6803/dolbykj2.th.png (http://img509.imageshack.us/my.php?image=dolbykj2.png)

So far i can say , for my Receiver via SPDIF ( AC3Filter)

Source :sounds great, high Dynamics, big Differences between DRC Modes
Method 1 :sounds as good as the Source with DRC off. No big Differences between DRC Modes, a bit louder as the Source
Method 2 :sounds by far lower Volume with or without DRC, mmmhhh
-On PC via AC3Filter i cant hear any real differences

One last Question : Can Method 1 cause clipping ?
THX for all your help!

madshi

29th July 2007, 20:17

InorganicMatter

29th July 2007, 20:46

SWEET!!

I think I shall reconvert the audio of the HDDVD I just finished converting to 720p AVC.

honai

29th July 2007, 23:34

Wow, that's amazing! Thank you so much!

Rectal Prolapse

30th July 2007, 05:39

Nice work madshi and the rest of you!

ACrowley

30th July 2007, 08:03

Good news!!

http://madshi.net/eac3to.zip

The latest eac3to version 1.10 now detects if E-AC3 tracks have Dialog Normalization activated (most tracks have) and removes it automatically! As a result by using eac3to Dialog Normalization with E-AC3 decoding is no longer a problem. (Thanks to tebasuna51 for the helpful tips!)

Furthermore I've found a way to talk the Nero E-AC3 decoder into not applying DRC, even though we're using it outside of ShowTime. That means: With the help of the Nero decoder we now have near perfect E-AC3 decoding. Enjoy!!

Finally, the Surcode automation was slightly improved. It now outputs some more messages. If it still fails for some people, please post the full eac3to output here. Thanks.

Because of the latest information about DRC and Dialog Normalization and the latest improvements in eac3to for the Nero decoder, I've now made Nero the default E-AC3 and TrueHD decoder. So the flag "/nero" is no more. There is a flag "/sonic" now to force the use of the Sonic filters. I do not recommend the Sonic filters, though, cause they still apply DRC and I don't know how to disable that.

@ACrowley, I hope you'll find some time to test the new version to make sure that E-AC3 decoding is really without faults now. Thanks!

Test Sample EAC3 640kbps 5.1 (300HDDVD)
http://img528.imageshack.us/img528/4598/sshot3he5.th.png (http://img528.imageshack.us/my.php?image=sshot3he5.png)

you can see ec3to 1.10 works now without DRC AND without DialogNorm (for EAC3). TrueHD is still with DialogNorm.
Madshi....nice work!:thanks:

SO the new Conclusion is :

TrueHD ,always, with NeroDecoder : DRC OFF ,Dialog ON
EAC3 "only" via "eac3to" with NeroDecoder : DRC OFF DialogN OFF

Later Today when im have Time i will show you that the Dynamics are as like as the AC3 640kbps from BluRay ! Nice

madshi

30th July 2007, 08:25

Test Sample EAC3 640kbps 5.1 (300HDDVD)
http://img528.imageshack.us/img528/4598/sshot3he5.th.png (http://img528.imageshack.us/my.php?image=sshot3he5.png)

you can see ec3to 1.10 works now without DRC AND without DialogNorm (for EAC3).
Thanks for confirming!

ACrowley

30th July 2007, 08:51

Heres a Comparison from AC3 640kbps Bluray and decoded EAC3 640kbps HDDVD (Nero eac3to 1.10) from 300

http://img257.imageshack.us/img257/8132/sshot3ih8.th.png (http://img257.imageshack.us/my.php?image=sshot3ih8.png)

Its obvious.
Thew 1.10 Version works perfect . No DRC ,No DialogNorm. anymore with NeroAudioDecoder

honai

30th July 2007, 09:15

Excellent!

I have observed an interesting effect.

I just encoded the English DD+ track from Batman Begins to DTS (w/ Surcode) and muxed the track alongside the DD track from the R1 DVD. With the DTS track dialog is considerably "louder" than the DD version. I route audio to my receiver over SPDIF (bit-perfect), and the receiver displays "DIAL. NORM +4" for the DTS track (Pioneer 1014). So it seems that DD+ -> DTS yields dialog loudness that is stronger than the "reference" from the DVD. Not that it sounds unpleasant. I'm just wondering if we aren't running into unintended consequences by ignoring the Dialog Normalization in the source DD+ track ...

madshi

30th July 2007, 09:40

I just encoded the English DD+ track from Batman Begins to DTS (w/ Surcode) and muxed the track alongside the DD track from the R1 DVD. With the DTS track dialog is considerably "louder" than the DD version. I route audio to my receiver over SPDIF (bit-perfect), and the receiver displays "DIAL. NORM +4" for the DTS track (Pioneer 1014). So it seems that DD+ -> DTS yields dialog loudness that is stronger than the "reference" from the DVD. Not that it sounds unpleasant. I'm just wondering if we aren't running into unintended consequences by ignoring the Dialog Normalization in the source DD+ track ...
As far as I understand it, the DD+ dialnorm stuff changes the volume of the whole track (all 6 channels). It doesn't just change the center channel. So I don't think that ignoring the DD+ dialnorm should have any specific effect on the center channel. But that's just AFAIK. Are you sure that it's not the whole track which is louder? That could be easily explained, because the original DD track will have dialnorm in it, but DTS has not. The the original DD track volume will be lowered by your receiver, while the DTS track should not be lowered. Maybe the receiver even raises volume for DTS tracks. That would result in even more loudness. But all this should affect all channels and not just the dialog (center).

honai

30th July 2007, 09:59

You're right, picked a few more listening spots, all channels are louder. The receiver doesn't do anything other than adhere to the standard. Maybe Surcode embeds some additional data?

In comparison I can say that the transcoded DTS is indeed louder than reference DTS tracks from retail DVDs.

EDIT:

On my receiver DTS tracks are typically less loud than DD tracks from the same DVD. I suspect that's because due to the algorithmic nature of DTS vs DD the DTS version allows for a greater dynamic range and thus needs more headroom, resulting in an overall lower volume. However, with Batman Begins the transcoded DTS is much louder than the original DD track.

Maybe the receiver even raises volume for DTS tracks.

Yes, that's what seems to be happening as indicated by the "+4" in "DIAL. NORM +4". For DD tracks the receiver doesn't indicate any volume adjustment.

madshi

30th July 2007, 10:08

So, should I add a switch to allow dialnorm to happen instead of always defeating it? Probably it makes sense to defeat dialnorm if you reencode in AC3, because then you can tell the AC3 encoder to reuse the original dialnorm value (eac3to 1.10 output the original dialnorm value for your information). But maybe when reencoding to DTS it would be better to let dialnorm happen during decoding? However, I think when decoding 16bit TrueHD tracks it would be better to defeat dialnorm to make lossless reencoding easier without quality loss and with small file sizes.

honai

30th July 2007, 10:11

ACrowley

30th July 2007, 12:31

I havent any Problems with reencoded AC3 when the DialogNorm was ignored through Source decoding

Always reencoded with -31db ,no DRC Profile

Should be no Problem

madshi

30th July 2007, 12:37

Here's a thread about DialNorm which I find worth reading:

http://www.hometheaterforum.com/htf/showthread.php?t=230926

tebasuna51

30th July 2007, 15:23

Yes, make it an option, that'd be perfect!

EDIT:

Hm, come to think of it - maybe there is indeed something in the DTS header (from the Surcode encoder) that tells the DTS encoder in the receiver if Dialog Normalization should be applied. Perhaps we should ask someone who's familiar with the DTS standard. Using Google I found conflicting information, i.e. that DTS has/has no facilities for Dial. Norm. information in the stream.

EDIT2:

http://www.laaudiofile.com/images/dtsencoder.jpg

Seems to follow the DD standard of 27..31dB Dialog Normalization. Surcode doesn't expose that parameter, but the resulting DTS stream should definitely receive the Dial. Norm. value obtained from the original EAC3 stream.

Of course DTS format have also a DialNorm field in the header (also have DRC support) but the solution isn't suply the original attenuated to the DTS encoder, the original must be at full dynamic and you can set the DialNorm in your DTS encoder if you want.

@madshi
Instead a option to apply or not the patch to override the DialNorm attenuation I think honai need a way to pass the original ac3 DialNorm value to the DTS encoder like parameter.

madshi

30th July 2007, 15:33

@madshi
Instead a option to apply or not the patch to override the DialNorm attenuation I think honai need a way to pass the original ac3 DialNorm value to the DTS encoder like parameter.
Unfortunately the Surcode DTS encoder doesn't seem to have an option to specify the DialNorm value.

Maybe I should manipulate the final DTS file to fill in the right DialNorm value? :)

ACrowley

30th July 2007, 17:07

Unfortunately the Surcode DTS encoder doesn't seem to have an option to specify the DialNorm value.

Maybe I should manipulate the final DTS file to fill in the right DialNorm value? :)

HKEY_CURRENT_USER\Software\Minnetonka Audio Software\Surcode DTS\OPTIONS

Theres Value for DialogNorm and DRC

Looks like default applies DialogN, maybe you should leave its on default...i never had Problems with Surcode encoded dts and DialogNorm and Dynamic Compress etc. Not with full Dynamic Sources and not with pre Dynam. compressed Sources

.mhh

@honai
I think DTS should sound louder as like as DD on most Receivers..not lower Volume.
On my Pioneer dts is always louder

hristoff2

30th July 2007, 17:09

Thanks once again, awesome. :)

Source: Shooter eac3 5.1 (1536kbps), English

top: eac3to 1.9 / Nero / Center CH
bottom: eac3to 1.10 / Nero / Center CH

http://www.imagebanana.com/img/ajezqffy/thumb/drc_no_drc_comp.png (http://www.imagebanana.com/view/ajezqffy/drc_no_drc_comp.png)

madshi

30th July 2007, 17:20

HKEY_CURRENT_USER\Software\Minnetonka Audio Software\Surcode DTS\OPTIONS

Theres Value for DialogNorm and DRC
I don't have these values there? How are they named exactly? Thanks...

Looks like default applies DialogN, maybe you should leave its on default...
I agree with tebasuna51 that I should forward the original E-AC3's DialNorm value to the DTS encoder. That makes more sense to me than letting the encoder use a default DialNorm value.

Why are you saying "applies"? There's nothing about DialNorm that the encoder "applies". The encoder does not do any DialNorm processing. All it does is writing a specific DialNorm number to the header of every encoded frame. DialNorm is "applied" by the decoder, not by the encoder!

ACrowley

30th July 2007, 17:32

I don't have these values there? How are they named exactly? Thanks...

I agree with tebasuna51 that I should forward the original E-AC3's DialNorm value to the DTS encoder. That makes more sense to me than letting the encoder use a default DialNorm value.

Why are you saying "applies"? There's nothing about DialNorm that the encoder "applies". The encoder does not do any DialNorm processing. All it does is writing a specific DialNorm number to the header of every encoded frame. DialNorm is "applied" by the decoder, not by the encoder!

Thats what i mean.. i told you my English is bad. I should better say the encoder applies the DialogNorm Info into the encode..better ?:)

Do you know how a DTS encoder handles Dialog Norm value ?
I dont if the dts encoder handles the encoding/decoding off DialogN as like as AC3,when -27db is in EAC3/AC3 if -27db is correct for a reencoded dts too? I really dont know the DTS tech Specs

But when you plan to set the DialogNorm from EAC3 into dts,it should be done for a TrueHD Source DialogNorm too ?
Mhhh i never thought about..all my dts reencodes are at default without any special DialogNorm Setting because theres simply no Option in Surcode/Nuendo

HKEY_CURRENT_USER\Software\Minnetonka Audio Software\Surcode DTS\OPTIONS
http://img252.imageshack.us/img252/4904/sshot3ge8.th.png (http://img252.imageshack.us/my.php?image=sshot3ge8.png)

madshi

30th July 2007, 17:43

Ok...Do you knw how a "Studio" DTS encoder handles Dialog Norm value ? Surcode is a certified DTS Encoder too..however, but more semi Pro
Maybe the Manual from DTS (HD) Master Suite shows how to do. I dont if the dts encoder handles the encoding/decoding with DialogN s like as AC3,when -27db is in EAC3 if -27db in dts is correct too? I really dont know
I don't know, either, but I would guess so... :)

P.S: You can test it out. Just set different DialNorm values in the Surcode registry keys. Then test encode in DTS. Then decode again and check the WAV.

But when you plan to set the DialogNorm from EAC3 ,it should be done for a TrueHD Source DialogNorm too ?
For TrueHD sources I would set DialNorm to -31dB in the DTS encoder because we currently can't turn DialNorm off during decoding.

So, how is the exact names of these registry values? :)