Welcome to Doom9's Forum, THE in-place to be for everyone interested in DVD conversion.

Before you start posting please read the forum rules. By posting to this forum you agree to abide by the rules.

 

Go Back   Doom9's Forum > General > Audio encoding

Closed Thread
 
Thread Tools Search this Thread Display Modes
Old 26th December 2003, 22:25   #261  |  Link
kempfand
Registered User
 
kempfand's Avatar
 
Join Date: Oct 2001
Location: Switzerland
Posts: 225
Ratings:
- I like Panorama slightly better than B-Pan & B-Proc
- Aurora gave excellent results, but I only used it on 3 CD-conversions, which for me is not enough to make a statement "it's better" than B-Pan.
- However: Given the speed-advantage of the Bidule-setup, I'd clearly give that more 'ranking-weight' than the "maybe slightly better" of Aurora.
- SIR (Super Impulse Reverb) -VST: I've just started to play with that and need to do more tests before I can make a statement.

Andreas
kempfand is offline  
Old 28th December 2003, 19:19   #262  |  Link
puzio
Registered User
 
Join Date: Nov 2003
Posts: 26
Hello Kempfand.

Thank You for your work.

Do You see that in Farina's guide with Audiomulch Y-signal is connected with 4-input of B_Decoder , NOT 3-input of B_Decoder.
So We have W,X,Z(silence),Y !!!!! in input of B_Decoder, NOT W,X,Y,Z.

In Bidule as You wrote I did convertion with "Bette Midler - Wind Beneath My Wings" I get "Front Left" loud, but "Front Right" is very low. With "Panorama" way they are similar.

?????

THX
puzio is offline  
Old 28th December 2003, 20:04   #263  |  Link
kempfand
Registered User
 
kempfand's Avatar
 
Join Date: Oct 2001
Location: Switzerland
Posts: 225
Hi puzio,
Quote:
Do You see that in Farina's guide with Audiomulch Y-signal is connected with 4-input of B_Decoder , NOT 3-input of B_Decoder.
So We have W,X,Z(silence),Y !!!!! in input of B_Decoder, NOT W,X,Y,Z.
Good catch. I personally believe it is a mistake on A. Farina's page.

As you correctly note, the reverb is done to create W,X,Y. No need for Z (as there is no information on Z in the starting stereo-wav).

On the B_decoder, the pin-sequence is W, X, Y, Z (as you correctly note).

According to my understanding, reverbed W should connect to W-pin of B-decoder, X to X, and Y to Y. This is also how EoH's original guide works.

Quote:
I get "Front Left" loud, but "Front Right" is very low. With "Panorama" way they are similar.
I get perfect reults here (both with SIR & Panorama). I believe that this problem (and similar ones, such as 'wrongly' rotated sound-fields) are related to how the channel-mapping & demuxing is done.

Some suggestions:
a) Make sure B-Proc is configured as in the guide
b) Use BeSweet to demux the 6-ch-wav into the mono-wav's. If you use wav2wav6.exe, you need to either re-name the channels, or use a different channel-mapping.

Good luck,
Andreas
kempfand is offline  
Old 28th December 2003, 20:30   #264  |  Link
Umma
Registered User
 
Join Date: May 2003
Posts: 76
Has anyone every contacted Angelo Farina about us using his guidelines over here?
Umma is offline  
Old 28th December 2003, 20:42   #265  |  Link
kempfand
Registered User
 
kempfand's Avatar
 
Join Date: Oct 2001
Location: Switzerland
Posts: 225
I'm not sure if I read your question ('his' guidelines ?) correctly (bear in mind that my native language is not English).

- If you refer the puzio's question about the wireing of the reverbed W, X & Y, then the answer is no (it's on my todo list).

- If you refer about potential 'ownership' of guidelines, then let me state that Farina's page was just recently updated, and that the bidule I outline above does exactly what what already described here in July 2002 ( A Guide from Stereo to 5 channel Surround ! .

As I said: My native language is not English, so I might mis-read.

Andreas
kempfand is offline  
Old 28th December 2003, 20:59   #266  |  Link
puzio
Registered User
 
Join Date: Nov 2003
Posts: 26
.
puzio is offline  
Old 28th December 2003, 21:35   #267  |  Link
Umma
Registered User
 
Join Date: May 2003
Posts: 76
Quote:
Originally posted by kempfand
I'm not sure if I read your question ('his' guidelines ?) correctly (bear in mind that my native language is not English).

- If you refer the puzio's question about the wireing of the reverbed W, X & Y, then the answer is no (it's on my todo list).
Sorry about the ambiguity. I should have been clearer as the wiring would have been a topic to touch with him, that and...

Quote:

- If you refer about potential 'ownership' of guidelines, then let me state that Farina's page was just recently updated, and that the bidule I outline above does exactly what what already described here in July 2002 ( A Guide from Stereo to 5 channel Surround ! .
I know it was with a couple of others (and you, too, right?) that EoH put together the first stereo to 5.1 guide, but he also referenced Farina's site in the use of Aurora to create the b-wave. I'm not saying Farina 'owns' the method of conversion, but by providing us with directions in using Farina's Aurora program (and with a big nod to Farina's own guide to using Cool Edit and Aurora), EoH indicated that Farina was a big inspiration at least in the beginnins of the conversion process. Of course it is a dynamic thing and it will constantly be changing as newer stuff comes out, but I guess my question should have been: Has Farina ever seen how his suggestions and use of his Aurora software have inspired people? Or something like that...

I didn't mean any offense, Andreas. I was basically wondering if anyone had contacted him about anything over the last year or so.

Pax,

D
Umma is offline  
Old 28th December 2003, 22:20   #268  |  Link
kempfand
Registered User
 
kempfand's Avatar
 
Join Date: Oct 2001
Location: Switzerland
Posts: 225
Quote:
Has Farina ever seen how his suggestions and use of his Aurora software have inspired people?
My guess is no. I'll write him and give the background. I've been a bit sloppy here and should have done it earlier. I agree that he should be given the right credit.

// off-topic on //
Reason I didn't do it yet: I've sent a really nice & kind email to Dave Malham, saying "thank you" for the nice tools he and his team provide for free (B-Pan, B-Proc, etc). Never heard back a single word ... Needless to say that I was very disappointed (fully understanding that he might have personal reasons which make him focus on other things). Similar experience (on different topics) with Mark Taylor on tooLame (exe/dll).
// off-topic off //

Peace and kind regards,
Andreas


Edit/Add: Also wrote an email to Christian Knufinke, the author of the SIR-VST, saying "thank you". Without the effort from people such as him, all this wouldn't be possible.

Last edited by kempfand; 28th December 2003 at 23:48.
kempfand is offline  
Old 29th December 2003, 14:22   #269  |  Link
Eye of Horus
Banned
 
Eye of Horus's Avatar
 
Join Date: Jun 2002
Location: The Netherlands
Posts: 326
Quote:
Originally posted by kempfand
My guess is no. I'll write him and give the background. I've been a bit sloppy here and should have done it earlier. I agree that he should be given the right credit.

// off-topic on //
Reason I didn't do it yet: I've sent a really nice & kind email to Dave Malham, saying "thank you" for the nice tools he and his team provide for free (B-Pan, B-Proc, etc). Never heard back a single word ... Needless to say that I was very disappointed (fully understanding that he might have personal reasons which make him focus on other things). Similar experience (on different topics) with Mark Taylor on tooLame (exe/dll).
// off-topic off //

Peace and kind regards,
Andreas


Edit/Add: Also wrote an email to Christian Knufinke, the author of the SIR-VST, saying "thank you". Without the effort from people such as him, all this wouldn't be possible.

I mailed Farina for at least 3 times and never got a response back !

grtz,

EoH
Eye of Horus is offline  
Old 29th December 2003, 14:40   #270  |  Link
Eye of Horus
Banned
 
Eye of Horus's Avatar
 
Join Date: Jun 2002
Location: The Netherlands
Posts: 326
Quote:
Originally posted by Umma
I didn't mean any offense, Andreas. I was basically wondering if anyone had contacted him about anything over the last year or so.

Pax,

D
I try to contact him long before I wrote my FIRST guide with the Aurora plugins. Because he didn't answer any of my emails, I decided to go on without him and luckily I found someone who helped me in the process.

The second guide was written without any of his software.

This year I emailed him for the third time and again got no response.
I gave up :-)

To me it's not quite clear where the problem is...... Guidelines like Farina's are on so many sides, that it's still the question if these were his' in the first place ! I mean.... he didn't invent Ambisonics ! And everyone can write guidelines about how to use tools. (even I !!)

Don't get me wrong : the guy is a genius, but I don't think it's necessary to "ask him permission" to write guidelines !

I would have loved however if any of these guys came on this board and supplied us with more tips and help. But..... after emailing (Andreas did some emailing too !) and no replies, we gave up hope...

What rests is that we have to find it all out by ourselves :-))

CanI add a nasty one ? If not allowed the moderator can remove it : I am not surprised that when people are selling stuff, but don't reply to emails, the potential buyers will look for illegal versions. If I ask if a routine can be explained, before buying and I get no answer, I won't buy it ! But I will try to find it out by myself with a not so legal version ! Seeing the enormous response my 2 guides have on this board, the makers of these software should think again. That's why I will give 10M kudos to David from Plogue ! When he start selling, I want to be one of the first to buy his program !! And...... don't forget it works two ways. We act as his beta-testers :-)

my 2 eurocents !

EoH
Eye of Horus is offline  
Old 29th December 2003, 16:47   #271  |  Link
kempfand
Registered User
 
kempfand's Avatar
 
Join Date: Oct 2001
Location: Switzerland
Posts: 225
Comments from Angelo Farina

Great news: I've just got mail from Angelo Farina , which I share with his permission.
Quote:
Subject: Re: UHJ t B : VST-method using SIR & Bidule

Hi,

I have seen Your postings, and it was evident that in my original AudioMulch
patch there were severe errors (patching errors and improper gain settings
in SIR). So I corrected everything (I hope)...

I also tested the system with Bidule, and added a corresponding image
on my web page:
Also usage of Audiomulch/Bidule and SIR (Super Impulse reverb) are explained

I also saved the Bidule patch on my Public server area: Link
Please, check if everything now is OK....

Regarding the concern about the usage of UHJ-to-Surround conversion for
not-UHJ stereo recordings, I can explain what follows:

1) Most coincident or near-coincident (i.e. ORTF) microphonic techniques
still work reasonably well in a UHJ-decoder, as "artificial" mixing (as
most pop/junk music is) designed for Dolby Surround (as almost anything
found on VHS tapes or stereo DVD-video). Spaced or binaural recordings,
instead, are not optimally treated by UHJ decoding networks.

2) The decoding impulse responses DO NOT CONTAIN any room information: the
surround channels are decoded from the original recording only, and contain
information which appear "out of phase" on the Left-Right channels. This is
very similar to the "Dolby Surround" approach, but here the result is stereo
instead of mono, and is full-band instead of band-limited. This outperforms
original Dolby Surround, Dolby ProLogic and similar 2-to-5 channels decoders
(Circle Surround, SRS, etc.). The only serious competitor which can rival or
even sound better than UHJ decoding is actually Dolby Pro Logic II
(available only in very recent surround processors).

3) For very dry original recordings there is no "surround" to be extracted
at all, and none of the above-mentioned methods (nor UHJ) can work. The
better solution, in these cases, is to employ convolution with room reverb.
I suggest, in these cases, to employ the Ambiophonics decoding method
(www.ambiophonics.org). Similar results, but less mathematically rigorous,
can be obtained with some Lexicon hardware tools (such as the 960L).

You can, of course, post this info also on the doom9 forum....
Bye!

Angelo Farina
The mentioned Bidule is correct. Based on my and EoH's listening tests, which compared different decoding-layouts (Square, ITU 5.1 loudspeaker, Pentagon) and different VST-decoders (B-Dec, Emigrator, VVMic), we prefer the Pentagon-layout, but feel free to make your own tests.

Kind regards,

Andreas
kempfand is offline  
Old 30th December 2003, 16:50   #272  |  Link
Umma
Registered User
 
Join Date: May 2003
Posts: 76
Wow. What a response. At least we know he read these posts. Strange timing on my part to ask about contacting him...

I did some comparisons with 5.1 and Pentagon, and Pentagon wins out, too. Not just with me, but to about three others who listened to some Simon and Garfunkel that I did. Pentagon won.

I really don't know what is meant by "dry" recordings, though.

And it is still not clear why the ambiophonics method should be used. I read through some of the stuff on the web site, but then I think "ambiophonics," I think about what Jonas was saying about using a room impulse created in the listening environment...I don't remember seeing anything about that at the ambiophonics web site.
Umma is offline  
Old 30th December 2003, 18:36   #273  |  Link
specise_8472
Guest
 
Posts: n/a
Have just uploaded to the FTP site a VST plugin to do ambiophonics.
At the moment just does the six mentioned encodings.

On a lot of recordings there is a big difference from ambiosonics decoding. One I tried it out on was The Music Of The X_Files. Track 6 contains chiming bells. The reverb through the rears was spectacular. Sounded good in Emigrator too, but was not quite as pronounced, and dropped off sooner.

Ambiophonics was produced as a way of producing true 5.1 from the information in an WXYZ file. This file contains the inforamtion needed for decoding
X = Front back
y = Left right
z = Up down (when included)
w = pressure signal (basicaly the input sound at -3db.

Any suggestions improvements etc. Let me know.

On another note, I have just got Roomverb M2 from www.spinaudio.com
Putting it between the input signal and B-Proc produces very good results. This is a very good reverb engine that as heaps of presets and full control over every setting. Some recordings are really brought to life. There is a demo avaliable.
 
Old 30th December 2003, 22:01   #274  |  Link
kempfand
Registered User
 
kempfand's Avatar
 
Join Date: Oct 2001
Location: Switzerland
Posts: 225
specise_8472: Thanks for sharing this. This looks really

It's funny: Yesterday, I was just reading the corresponding paper and thought this would be great to try. I don't habe the necessary coding experience though, to put this into dll.

I will try it tomorrow :-)

Thanks again and kind regards,

Andreas
kempfand is offline  
Old 31st December 2003, 01:10   #275  |  Link
kempfand
Registered User
 
kempfand's Avatar
 
Join Date: Oct 2001
Location: Switzerland
Posts: 225
Update: Specise's ambiophonics-VST ROCKS ! I just did a testing series with one song (Armik: Rubia; guitar music).

Created 9 tracks as follows:
Code:
1) B-Pan, B-Proc, Emigrator (Pentagon)
2) Panorama, Emigrator (Pentagon)
3) SIR, Emigrator (Pentagon)
4) Ambio I
5) Ambio I~
6) Ambio J
7) Ambio J~
8) Ambio K
9) Ambio K~
All were batch-BeSweet'ed to dts with normalization (@DSPGuru: BeSweet working in batch-mode like a charm).

Based on this one-song-only-test, B-Pan is very good, but for my ears, "Ambio J" wins (more discrete impression, nicer & broader sound-image). In general, I preferred the I , J, K -type over the "~"-type, but this is my personal impression and based on 1 initial song only.

In summary @ All: Highly recommend you give it a try.

If you want more technical background, read the corresponding papers from Robin Miller , especially
- Transforming Ambiophonic + Ambisonic 3D Surround Sound to & from ITU 5.1/6.1
- Scalable Tri-play Recording for Stereo, ITU 5.1/6.1 2D, and Periphonic 3D (with Height) Compatible Surround Sound Reproduction

@specise_8472:

Quote:
Any suggestions improvements etc. Let me know.
I'm just thinking on howto achieve a workable & affordable solution to include the SC (surround Center) speaker. My amp (Yamaha 630) has a speaker-connection for SC.

If I understand this correctly, SC with dts is matrixed into the SL&SR. But this is currently probably a 'dead-end' as I'd guess that no affordable encoders are around.

Any chance or idea on howto 'achieve' this packaging of SC (into SL&SR) already within the VST ? Just thinking lound ...

Kind regards,

Andreas
kempfand is offline  
Old 31st December 2003, 04:46   #276  |  Link
specise_8472
Guest
 
Posts: n/a
Quote:
Originally posted by kempfand
Update: Specise's ambiophonics-VST ROCKS ! I just did a testing series with one song (Armik: Rubia; guitar music).


I'm just thinking on howto achieve a workable & affordable solution to include the SC (surround Center) speaker. My amp (Yamaha 630) has a speaker-connection for SC.

If I understand this correctly, SC with dts is matrixed into the SL&SR. But this is currently probably a 'dead-end' as I'd guess that no affordable encoders are around.

Any chance or idea on howto 'achieve' this packaging of SC (into SL&SR) already within the VST ? Just thinking lound ...

Kind regards,

Andreas
As is mentioned in the papers, 6.1 AC3 is matrixed center surround, and so does not work properly. At the moment I have done as suggested and mixed the SC channel into the LS RS at -3db. The only way to do a proper SC is to use DTS Discrete 6.1 which there is currenly no software for that I know of. Nuendo and Surcode only support 5.1.

I can output the SC if wanted, if you have a reason for it?
I am looking at writing a PLI/II encoder in the future to use. Also I am looking now into a filewriter that thakes the output from the Steinberg VST AC3 encoder and puts it to disk as a valid AC3 stream. Just think from Source to AC3 in one hit.

Also in the papers are mentioned that you can recover the original WXYZ file from the 5.1 track. Would there be a use for this facility to turn original AC3 tracks (and DTS) into WXYZ files for processing through emigrator to more or less speakers? Just need to reverse the plugin and change the co-efficients.

About the difference between ijk ijk~, the ~ are basically the same but with the microphone tilted -30degrees.

Just thinking out loud

Last edited by specise_8472; 31st December 2003 at 05:49.
 
Old 31st December 2003, 22:16   #277  |  Link
Ace_V
Registered User
 
Join Date: Dec 2003
Location: New Zealand
Posts: 5
Hi,

I think I've set up every thing correctly but my front and rear channels sound very similar. There shouldn't be as much vocals in the rear channels as in the front, right?

Can anyone help me? I must have done something wrong.

Also just tried it with a mono file which I would have expected to only come through the center channel or at the most the front channels, but it has pretty much the same sound in all channels except the LFE.

thanks

Last edited by Ace_V; 1st January 2004 at 07:25.
Ace_V is offline  
Old 2nd January 2004, 02:36   #278  |  Link
kempfand
Registered User
 
kempfand's Avatar
 
Join Date: Oct 2001
Location: Switzerland
Posts: 225
Quote:
The only way to do a proper SC is to use DTS Discrete 6.1 which there is currenly no software for that I know of.
I had to do some reading an am still getting 'up to speed' on the topic. Regarding dts ES, I understand there are 2 flavours: DTS-ES Discrete , for which I also understand there is currently no encoder, and DTS ES matrix. See e.g. technical_literature.

For the "DTS ES matrix", I understand all that is needed is:
(a) 45 degree phase shift applied to LS and RS signals
(b) CS at -3db
See above source, as well as dts-es.pdf

Update / Edit added below

Now:
Quote:
As is mentioned in the papers, 6.1 AC3 is matrixed center surround, and so does not work properly
I didn't find that concrete recommendation (can you provide a link, maybe I was just having bricks in front of my eyes). If it that's what is stated, we should test. I'm stating this in analogy to the 5.0-decoding rigs (Pentagon vs. ITU 5.1 layout), where 'recommendations go for 5.1, and personal tests often go for Pentagon.

---
Quote:
Would there be a use for this facility to turn original AC3 tracks (and DTS) into WXYZ files for processing through emigrator to more or less speakers?
Still thinking about this. I could see a potential application for 'down-mixing' into 2.0. Would be keen to see how this compares to other methods (e.g. azid).

Don't want to 'make you work', but I'd be very keen and happ to test, if it's done without too much effort (realising here that nothing is 'easy').

---

Regarding PanAmbio 4.1: Robin Miller mentions this as
Quote:
For a single listener, PanAmbio is superior to 5.1 in accurate 360° localization, spatial impression, and envelopment - a benchmark of excellence
Any idea on howto achieve this witha bidule ?

---

All this is also thinking out loud

Kind regards,

Andreas

P.S. Have done more tests with your Ambi-rc1.dll. It still 'rocks' and wins !

Edit: I'm assuming that "DTS ES matrix" can be encoded with current SurCode version, if conditions (a) & (b) above are fullfilled.

I did a test as follows; BeSweet/azidts-demux a "DTS ES matrix" encoded audio (had to use "-ota( -fs 44100"), and SurCode-re-encoded. FOrcing "DTS ES Matrix" on the amp resuited in the same audible siganl in each speaker (also SC)

Last edited by kempfand; 2nd January 2004 at 03:06.
kempfand is offline  
Old 2nd January 2004, 03:59   #279  |  Link
specise_8472
Guest
 
Posts: n/a
Okay, one at a time

Quote:
For the "DTS ES matrix", I understand all that is needed is:
(a) 45 degree phase shift applied to LS and RS signals
(b) CS at -3db
True, but Surcode does not allow a true implementation of this, as in it will encode okay, but will not set the flags needed in the frame headers to indicate that the front or/and back left-right pair/s are matrixed encoded with the appropriate center signal. Writing a program to go through and do this after could be done. But I have found that the published specs and the actual specs are just different enouth to cause trouble on this point. There are unknown flags that are being used

Quote:
For a single listener, PanAmbio is superior to 5.1 in accurate 360° localization, spatial impression, and envelopment - a benchmark of excellence
I have not found anything majorly useful at the moment to look a adding this feature (yet). I am still testing/playing around. A lot of this centers around crosstalk cancellation, which I am investigating.

Quote:
P.S. Have done more tests with your Ambi-rc1.dll. It still 'rocks' and wins !
As Robin Miller points out, Ambisonics suffers from lack of spaciousness. But a hybrid system of the two prtoduces the best results. Basically what I am trying to acheive. Per-Ambio from Ambisonic WXYZ. Also Ambiophonics really only caters for the single listener sitting in the "sweet" spot. Whereas Ambisonics has a large "sweet" spot.
 
Old 3rd January 2004, 02:22   #280  |  Link
specise_8472
Guest
 
Posts: n/a
Have just uploaded Version 1 of ambiophonics decoder.
Have added 6 new presets, and deleated the 3 that did not work good.

The new presets are based upon proper Ambisonic decode equations, the two versions of each are s = spherical decode and p = controled opposites.

Attention Daphy - just noticed that my old DLL was gone from the site. Was this you, or are you going to have the age old problem of people deleating stuff for the hell of it
 
Closed Thread

Thread Tools Search this Thread
Search this Thread:

Advanced Search
Display Modes

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off

Forum Jump


All times are GMT +1. The time now is 11:46.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2024, vBulletin Solutions Inc.