Old 09-17-2021, 06:50 PM   #1
sguyader
Human being with feelings
 
Join Date: Dec 2020
Posts: 70
Default Examples of binaural export using different plugins

I wanted to explore binaural audio, with the aim of converting my stereo mixes to something more immersive, to be listened on headphones. To this end, I started looking at different options, and I found that mixing plugins from different brands sound better to me.

Below, I put 4 clips using the same stereo track (a stem of recorded acoustic drum kit, with reverb baked in it, panned -/+ 90°) so you can compare different plugins and associations of plugins:

https://soundcloud.com/sebastien-guy...l-plugins-test

1. Full Sparta chain, ambiENC followed by ambBIN (both set to 7th order ambisonics)

2. IEM stereo encoder (set to 7th order ambisonics), followed by the Sparta ambiBIN binaural decoder (set to 7th order ambisonics)

3. DearVR Pro, set to encode AmbiX TOA, followed by the Sparta ambiBIN binaural decoder (set to TOA)

4. DearVR Pro, direct export from its binaural export setting

What's your opinion? I really dislike 4., but I like 3. a lot. Options 1. and 2. sound a bit different because the encoders are different, but they are quite close because the binaural decoder is the same.

Last edited by sguyader; 09-17-2021 at 06:58 PM.
sguyader is offline   Reply With Quote
Old 09-17-2021, 11:26 PM   #2
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 13,660
Default

Great idea.

I'll give it a listen later when in the studio and let you know!
__________________
subproject FRs click here
note: don't search for my pseudonym on the web. The "musicbynumbers" you find is not me or the name I use for my own music.
musicbynumbers is online now   Reply With Quote
Old 09-19-2021, 11:03 AM   #3
Kewl
Human being with feelings
 
Join Date: Jan 2009
Location: Montreal, Canada
Posts: 89
Default

Quote:
Originally Posted by sguyader View Post
Options 1. and 2. sound a bit different because the encoders are different, but they are quite close because the binaural decoder is the same.
This is the surprising part. On a mono source, the encoding should really be the same for a given azimuth and elevation.

So here, it boils down to how the encoders deal with stereo. Slight gain differences.
Kewl is offline   Reply With Quote
Old 09-19-2021, 07:41 PM   #4
sguyader
Human being with feelings
 
Join Date: Dec 2020
Posts: 70
Default

Quote:
Originally Posted by Kewl View Post
This is the surprising part. On a mono source, the encoding should really be the same for a given azimuth and elevation.

So here, it boils down to how the encoders deal with stereo. Slight gain differences.
Well, you made a point. The placement of the instrument was not exactly the same between both encoders, particularly in terms of elevation. I did the test againbut with the exact same placement (+/- 88° azimuth, 30° elevation) and now the difference is inaudible. I did a null test, and there's only a slight difference between them, that is not explained by the difference in gain (the full Sparta chain sounds fuller to me). But that difference in really negligible.

I still puzzled by the results from DearVR Pro: the encoder/panner is really nice and gives me a stronger sense of depth and spatial separation (the instrument sounds less in your face). However its binaural decoder really sucks (at least to my ears).

There's a lot to explore with the Sparta ambiBIN decoder. I found a source of HRIR's, and the following clip is the same drum loop, with a KU100 HRIR instead of the default one, and "phase simplification" preprocessing instead of the default "diffuse-field EQ" setting:

https://soundcloud.com/sebastien-guy...0-phasesimplif

That one sound really good to me: deep, full, spacious, "alive"
sguyader is offline   Reply With Quote
Old 09-20-2021, 10:14 AM   #5
Kewl
Human being with feelings
 
Join Date: Jan 2009
Location: Montreal, Canada
Posts: 89
Default



http://www.radio.uqam.ca/divers/ambi_encoding.png

Mono encoding, 30° azimuth, 30° elevation. The encoding is identical. There are subtle differences, related probably to the math library used in each encoder and how it deals with degree/radian conversion, precision of PI, precision of sine and cosine, etc.
Kewl is offline   Reply With Quote
Old 09-20-2021, 10:32 AM   #6
sguyader
Human being with feelings
 
Join Date: Dec 2020
Posts: 70
Default

Quote:
Originally Posted by Kewl View Post


http://www.radio.uqam.ca/divers/ambi_encoding.png

Mono encoding, 30° azimuth, 30° elevation. The encoding is identical. There are subtle differences, related probably to the math library used in each encoder and how it deals with degree/radian conversion, precision of PI, precision of sine and cosine, etc.
Really very close! but any idea why the last value has opposite sign between both encoders?
sguyader is offline   Reply With Quote
Old 09-21-2021, 07:49 AM   #7
Kewl
Human being with feelings
 
Join Date: Jan 2009
Location: Montreal, Canada
Posts: 89
Default

Quote:
Originally Posted by sguyader View Post
Really very close! but any idea why the last value has opposite sign between both encoders?
One is 0.00000002 and the other is -0.00000002. So, basically, noise around zero.

The formula for that channel is sqrt(5/8)*cos(3*azimuth)*(cos(elevation))^3. http://pcfarina.eng.unipr.it/Aurora/...t_formulas.htm

So difference probably due to different math libraries and how they deal with degree/radian conversion, precision of PI, precision of cosine, etc.

Last edited by Kewl; 09-21-2021 at 03:39 PM.
Kewl is offline   Reply With Quote
Old 09-21-2021, 07:58 AM   #8
Kewl
Human being with feelings
 
Join Date: Jan 2009
Location: Montreal, Canada
Posts: 89
Default

That being said, maybe the perceptual result would be closer if you compared SPARTA AmbiENC to IEM MultiEncoder, rather than the IEM StereoEncoder. The StereoEncoder has to deal with Left and Right summing: the AmbiENC and MultiEncoder, I believe, do not.
Kewl is offline   Reply With Quote
Old 09-21-2021, 01:38 PM   #9
sguyader
Human being with feelings
 
Join Date: Dec 2020
Posts: 70
Default

I think I'm done with comparisons. For now I'll stick with Sparta ambiENC and ambiBIN, except maybe when I want to automate some movement of the tracks, as I found that ambiENC doesn't write the automation envelopes when I click and move the object in ambiENC (it works with the IEM encoder).
sguyader is offline   Reply With Quote
Old 09-22-2021, 03:45 AM   #10
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 13,660
Default

3 seems best to me
__________________
subproject FRs click here
note: don't search for my pseudonym on the web. The "musicbynumbers" you find is not me or the name I use for my own music.
musicbynumbers is online now   Reply With Quote
Old 10-15-2021, 03:30 AM   #11
Dear Reality
Human being with feelings
 
Dear Reality's Avatar
 
Join Date: Jun 2018
Location: Düsseldorf
Posts: 5
Default

Quote:
Originally Posted by sguyader View Post

I still puzzled by the results from DearVR Pro: the encoder/panner is really nice and gives me a stronger sense of depth and spatial separation (the instrument sounds less in your face). However its binaural decoder really sucks (at least to my ears).
Hi everyone,

There must be something wrong with the setup in example 4. This is not our binaural sound

You can listen to some examples on Plugin Alliance and also get a free 14-day trial there, if you want to check dearVR PRO:
https://www.plugin-alliance.com/en/p...earvr_pro.html

Best,
Peter
Dear Reality is offline   Reply With Quote
Old 10-18-2021, 06:57 AM   #12
Dannii
Human being with feelings
 
Dannii's Avatar
 
Join Date: Mar 2010
Location: Adelaide, South Australia (originally from Geelong)
Posts: 5,591
Default

I prefer 3 most. It seems to sound most natural to me.
Having said that, I have yet to find any binaural encoder that convinces me. Perhaps I have a very unusual HRIR profile.
I've tried many encoders from many sources but the only thing I've found convincing are actual recordings done with a KU100. Even the encoders using KU100 impulse responses don't sound convincing to me.
__________________
Dannii is offline   Reply With Quote
Old 10-19-2021, 10:06 PM   #13
BPBaker
Human being with feelings
 
BPBaker's Avatar
 
Join Date: Oct 2013
Location: Brooklyn, NY
Posts: 54
Default

Quote:
Originally Posted by Dear Reality View Post
Hi everyone,

There must be something wrong with the setup in example 4. This is not our binaural sound

You can listen to some examples on Plugin Alliance and also get a free 14-day trial there, if you want to check dearVR PRO:
https://www.plugin-alliance.com/en/p...earvr_pro.html

Best,
Peter
Hello Peter,

Would Dear VR be open to adding ambisonic inputs to the DearVR Pro plugin? I was surprised to find out (after purchasing) that it sums the input signals to mono--I had expected to be able to use it as a true ambisonic reverb.

Having inputs for a diversity of formats (stereo, ambix b-format, fuma, surround, cube, etc.) would make this tool considerably more powerful.

Could I also request support for other VBAP spatial formats like Mach 1/SPS?
__________________
www.brendanpatrickbaker.com

Last edited by BPBaker; Yesterday at 06:28 PM.
BPBaker is offline   Reply With Quote
Old Yesterday, 01:24 PM   #14
sguyader
Human being with feelings
 
Join Date: Dec 2020
Posts: 70
Default

Quote:
Originally Posted by Dear Reality View Post
Hi everyone,

There must be something wrong with the setup in example 4. This is not our binaural sound

You can listen to some examples on Plugin Alliance and also get a free 14-day trial there, if you want to check dearVR PRO:
https://www.plugin-alliance.com/en/p...earvr_pro.html

Best,
Peter
Hi Peter,
Thanks for chiming in. I actually already own a license for DearVR Pro (and DearVR Monitor as well).
I'm surprised too as I really that the demo sounds on your website sound good, but on my system it doesn't. And I don't know what could be wrong on my side.
By the way, I saw a video on Michael Wagner's channel today where he uses DearVR monitor for monitoring a Dolby Atmos mix, and again the DearVR render sounds more coloured than the others.

Finally while you're here, it would be great if the DearVR plugins could be updated. Not only they are still not native for Mac M1 chips, but also on my Big Sur système there are bad graphical glitches on buttons and sliders.
sguyader is offline   Reply With Quote
Old Yesterday, 01:33 PM   #15
sguyader
Human being with feelings
 
Join Date: Dec 2020
Posts: 70
Default

Quote:
Originally Posted by Dannii View Post
I prefer 3 most. It seems to sound most natural to me.
Having said that, I have yet to find any binaural encoder that convinces me. Perhaps I have a very unusual HRIR profile.
I've tried many encoders from many sources but the only thing I've found convincing are actual recordings done with a KU100. Even the encoders using KU100 impulse responses don't sound convincing to me.
I agree that sounds coming from behind don't really give the impression of coming from behind. Same for above and below. But still with ambisonics encoding/binaural decoding the sound is more immersive, less flat. Not full 3D yet, but ~2.4D 😂
sguyader is offline   Reply With Quote
Reply

Thread Tools
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 -7. The time now is 05:14 AM.


Powered by vBulletin® Version 3.8.11
Copyright ©2000 - 2021, vBulletin Solutions Inc.