Old 07-28-2019, 02:44 PM   #1
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 14,214
Default Need some ambisonic help! :)

Just getting finally around to doing some ambisonics (via a dolby atmos setup) and excited but got to my first hurdle which is finding plugins that can work at 64 samples or.. can at least let me use the js time adjustment delay (or similar to add enough buffer to them.

Was hoping to have it at 64 buffers (my usual) but the Sparta plugins need 128 so tempted to buy into another set if needed or if there's a workaround? Or is 128 just the lowest we can go (if we want to keep things linear?)

Thanks..
__________________
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 offline   Reply With Quote
Old 07-28-2019, 11:03 PM   #2
RDan
Human being with feelings
 
Join Date: Jul 2017
Posts: 24
Default

High there,

In this forum section you’ll find a big list of Ambisonic tools.
E.g. the IEM plug-ins support any buffersize: https://plugins.iem.at/

Best
Daniel
RDan is offline   Reply With Quote
Old 07-29-2019, 01:48 AM   #3
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 14,214
Default

Thanks.

I will try the iem ones, must have missed them before.

Will see how comprehensive they are.
__________________
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 offline   Reply With Quote
Old 07-31-2019, 01:58 AM   #4
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 14,214
Default

Hi all.

Started getting somewhere the other day with 7th order ambisonics mostly via the iem plugins so thanks for that.

So far, have a few other questions if any one could have a go at helping that would be much appreciated thanks!

1. Compared to vr where its very easy to localise above and behind sounds in games, so I'm not getting very good headphone positional cues for above or behind me with the binaural plugins I've tried so far (the iem one) and the js one that has many impulses in it.

I'm guessing this is because so far I've only panned using the iem panner which is strictly a volume based pan so no time based cues. My question is, are there pan plugins that will also add some positional delay between ears (haas panning) so that we have proper cues for positioning or does the binaural plugin actually try to add in the head separation anyway if a sound is above the head and to the left? I did add the iem reverb and it definitely helped me localise the direct sound via having the additional cues all around me. Any thoughts?

2. Is there a panner that can do straight front to back, top to bottom or side to side panning without first going through the height or floor speakers first (or the sides/front if it's top to bottom)?

I noticed "hemisphere panners" in the blue ripple plugins but they read like they will still utilitise the top speakers if I pan back to front? Or is this just not an effective pan in Ambisonics anyway? In dolby atmos, it works fine to pan via just the surrounds and not use the heights so I'm hoping there's a plugin to do that

Thanks guys!
__________________
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.

Last edited by musicbynumbers; 07-31-2019 at 02:59 AM.
musicbynumbers is offline   Reply With Quote
Old 07-31-2019, 03:50 AM   #5
jm duchenne
Human being with feelings
 
jm duchenne's Avatar
 
Join Date: Feb 2006
Location: France
Posts: 914
Default

Hi,

IEM binaural plugin, as every binaural ones, uses HRTF filtering to simulate spatial cues, but you know that without personalized HRTF and/or head tracking it can be more or less acceptable...
Other ambisonics to binaural panners or converters may give different results since they may use different HRTF sets.
Doing binaural from ambisonics you combine two potentially problematic techniques ;-)

On the other hand what you describe as "not going through" but passing "above or below" is a limit of ambisonics that is not really "3D" but a hollow sphere. We would need Azimut, Elevation AND Distance coding, the later not being taken into account by the ambisonics principle.
Some algos try to create nearfield simulations to compensate from the fixed radius (and of course distance simulations for farther positions), but to my knowledge it is not yet very effective.
Other spatializations methods have not this limit, but nevertheless the HRTF are also generally not accurate for nearfield, even if you use DBAP.

You can try the free GRMTools SpaceVR that can behave differently :
https://inagrm.com/en/showcase/news/598/nouveaux-spaces

By the way, what Dolby Atmos tool do you use ?

Last edited by jm duchenne; 07-31-2019 at 03:56 AM.
jm duchenne is offline   Reply With Quote
Old 07-31-2019, 05:36 AM   #6
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 14,214
Default

Thanks for the detailed response.

I've been looking today at some tools like "dearVR pro" and the waves one and there are definitely some examples on the videos for DearVR that definitely feel behind or above me. I'm not sure yet without trying them out though whether that's the visual cues or hopefully a better use of acoustic cues from the different ear delays mixed in with later wall reflections.

I've got most GRM plugins so will check those too thanks

I've gone through quite a through of the HRTF impulses and they don't work that well for me yet but maybe it's because of using headphones (HD-600) rather than in ear ones but I thought it wouldn't be affected by the ear/Pinna shape anyway at that close range.

Regards to direct speaker to speaker panning. I guess I could try crossfading (equal power) between the same sound at the back then to the front so it's just volume based but was hoping for something that took into account some Haas delay too.

For dolby Atmos. This is mostly for games (my main job) whereby it's encoded via middleware like Fmod but I've also set up both the reasurround plugin for simple atmos decoding and then the IEM "ALLRA" decoder for going from ambisonic to my dolby atmos array (7th order for now).

My aim overall is for my personal music to be written in such a way that it can simultaneously be output to 3 formats without too much fuss. Dolby atmos, binaural and raw ambisonic for use in a vr animation/unity project at some point too. Oh and try and get a decent mixdown for stereo speakers too without too much trouble.

That might be too much to ask for without a lot of tweaking though but Readave has mentioned a few times that ambisonic mixes can work well in dolby atmos/7.1 etc. Just wish panning was a bit more specific (even at high order).

Maybe without visual cues from say a 360 video and subtle head movements you subconsciously do in VR, it's just hard to get convincing top/below and back positioning on binaural encoded headphones that are "static" without also tracking head position so that the subtle movements we make with our heads to help position sounds would be included?

Hmm.. it's exciting stuff though!

Here's a photo of the new studio build too

https://drive.google.com/open?id=1OH...Drrx9PbbUQR5RI

(although it's missing the acoustic treatment that's going up soon and apart from the 4 Kii three's which are my mains/retirement fund if I go deaf for 5.1 surround and all 4 handle the subwoofer output too) the other speakers are a hodgepodge with my old battered BM15As for rears and cheap speakers for the height ones but might get some floor ones if all goes well too). Also, half the speakers and tv are on rails so I can slide them into couch atmos mode or for gaming too. Just needs some aesthetic love too at some point

Note: I've done the delay and volume normalisation on the speakers too as not ideal positions yet but working on that and started balancing out the timbral differences in the height speakers but not buying another 6 Kii three's for sure! also, got a Denon decoder for atmos from hdmi sources too.
__________________
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.

Last edited by musicbynumbers; 07-31-2019 at 01:38 PM.
musicbynumbers is offline   Reply With Quote
Old 08-29-2019, 10:38 AM   #7
Dannii
Human being with feelings
 
Dannii's Avatar
 
Join Date: Mar 2010
Location: Adelaide, South Australia (originally from Geelong)
Posts: 5,598
Default

I haven't installed it yet but I just downloaded the Facebook Ambisonic VST plugins to try out their Binaural encoding.
Like you, I've not had much success with any of the Binaural plugins I've used so far as far as creating a solid front/back and top/bottom convincing image goes. Left/right is less of a problem.

I'll post here again after I've tried the FB plugins.
__________________
Dannii is offline   Reply With Quote
Old 08-29-2019, 10:30 PM   #8
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 14,214
Default

Thanks.

I found it helped a bit to have some kind of reverb like the Ambisonic iem one on the sound too so it had some delay difference between the dry and wet part of it. Helped to localise it.

I'll try myself too set something up in vr at some point and use the steamvr binaural stuff or Google or the Facebook one.

It might be that it's so easy to localise in vr because of the constant slight head movements we do without being conscious of them as I can even close my eyes in vr and still know exactly where a sound is above/below/behind etc.

Either that or their default ear model works much better anyway.

My dream is to be able to take these ambisonic and atmos mixes I'm doing and easily out them into binaural so that people could get the effect in headphones but not sure if that's the case yet.

Hopefully so!
__________________
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 offline   Reply With Quote
Old 06-12-2020, 02:33 AM   #9
junh1024
Human being with feelings
 
Join Date: Feb 2014
Posts: 240
Default

I can't fix your current problem of multiple parameter linking https://forum.cockos.com/showthread.php?t=237746 , but if you use my tools https://github.com/junh1024/Reaper-Surround (or my template below), you can potentially get a better workflow. By the way that you talk about HOA & channels I thought that you had 20 speakers so my system wouldn't be appropriate for you, but you have less so it might work. Best to read the readme https://github.com/junh1024/Reaper-S...d#introduction in full to get to grips with the FX & my 15.1 system.

I'm only promoting this on a personal basis rn since it's WIP & I'm working on some more FX, mainly upscalers. But reading the text in this thread & the other one you might want something sooner. The 151 GUI panner is a few months old/alpha. The non-GUI panners are 1-2 years old. System is 3 years old.



15.1 is a traditional surround system so there's little unnecessary bleed unlike ambisonics. & you can use traditional surround or stereo FX. For 16ch reverb & EQ, use the IEM or MCFX suite. IEM has a compressor too. I have a few downmixers, so there's probably a downmixer(s) for you. So instead of parameter linking, you can just have FX on the sub master, if it works for you.

I've attached a mixing template for quad delivery (Dolby atmos, binaural and raw ambisonic, stereo) from 15.1. You will need the ambix suite cuz my output is Fuma, unless you want fuma. @M is vague since there are many possible setups, the official format for the channel part of @M is 9.1 in cubendo & PT. So insert whatever downmixers you like on S_proc. You can insert commercial or IEM downmixers for 20HRTF. You will need to do your own ouput mapping for master output.

You can also take my idea of a sub-master + export buses & make a similar template for an ambisonics project.

What speaker layout do you have (it's not clear from the photo whether you have C & LFE) & what @M layout do you want?
Attached Files
File Type: rpp BorapDownmix-16-mbn-templ.RPP (22.5 KB, 317 views)

Last edited by junh1024; 06-12-2020 at 04:02 AM.
junh1024 is offline   Reply With Quote
Old 06-12-2020, 05:29 AM   #10
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 14,214
Default

Thanks!

I've not been able to read through this properly today but will aim to on the weekend and get to grips with it.

I'm set up for Atmos but I use a phantom centre as couldn't stretch the budget to buy another Kii three speaker lol plus it's fine for just one person monitoring which is normal for me. The LFE is managed by utilising all 4 Kii three's (as they go down to 20hz) with time alignment/filters etc as I found it gives me fantastically accurate LFE overall but for music I don't touch it anyway as intend to also convert the music into ambisonic/binaural (for VR multimedia use).

Anyway, I'll aim to look at it soon. I've got a few stereo projects to finish first but will get back to this asap and thanks for sharing. I'm hoping MPE will be able to modify that script when he has time, I just have to be patient lol It should be useful anyway but yours looks great to and will try it out.

Lastly though.. From experiments, I have found that I prefer panning so far via something like reasurround in normal dolby atmos as I find it seems to do things like back to front translations better than even 7th order ambisonics for me so far (without having to go above or below on the way) but I did find that trying to convert to Ambisonics afterwards caused me to not have the right positions for the above speakers so I need to work through the angles and set that up better.
__________________
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.

Last edited by musicbynumbers; 06-12-2020 at 05:36 AM.
musicbynumbers is offline   Reply With Quote
Old 06-13-2020, 02:14 AM   #11
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 14,214
Default

Hi junh1024

I had a read over the read me file.

I'll try and give this a go properly at some point soon. After testing ambisonic panning vs just the typical reasurround one. I'm leaning towards mixing in dolby atmos 7.1.4 (as my room has that speaker config) and then converting to ambisonic/binarual after and using head tracking on headphones to finalise that. Especially as I don't have any speakers below head height.

Also, I tend to still use stereo inputs or sometimes quad inputs so that I can do additional processing like Haas panning, EQ panning or modulated reverb/filters that can sit in different pan positions to the dry and allow me to psycho-acoustically pan stuff more effectively than just volume based panning.

I'm still experimenting with it all though and if I end up going back to using ambisonic as the primary format I'll definitely look at this as a method of working

Thanks for taking the time to share though!
__________________
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 offline   Reply With Quote
Old 06-13-2020, 03:08 AM   #12
junh1024
Human being with feelings
 
Join Date: Feb 2014
Posts: 240
Default

Quote:
Originally Posted by musicbynumbers View Post
Hi junh1024
I'm leaning towards mixing in dolby atmos 7.1.4 (as my room has that speaker config)

Also, I tend to still use stereo inputs or sometimes quad inputs

OK, so you can use "1.0 to 15.1 Panner GUI (L).txt" (width 100%) then "15.1 to 7.1.4 Downmix (S).txt". Or direct routing via parent channel if you've memorized the order. Using the GUI panner *might* be a more enjoyable experience than reasurround for height.

I have only a few speakers & My speaker array is in a state of dis- array atm cuz cheap rubber bands sortof liquefy.

You can maybe use "5.1 Mix Control (M).txt" as a non-GUI quad panner/manipulator. NB: it expects SMPTE 5.1 so please pad/route your quad stuff to 51.

I know psycho- corrections / modifications like EQ/delay can reinforce the image, but they might be downmix incompatible. I'm looking at alternatives to EQ/delay to increase separation in other contexts.

BTW, might be better to drop the "dolby atmos" monkier & just use 7.1.4 or "3D surround" instead since atmos might imply a codec & official dolby tools/DAWs unless you have those.

Last edited by junh1024; 06-13-2020 at 04:36 AM.
junh1024 is offline   Reply With Quote
Old 06-13-2020, 06:31 AM   #13
musicbynumbers
Human being with feelings
 
musicbynumbers's Avatar
 
Join Date: Jun 2009
Location: South, UK
Posts: 14,214
Default

Thanks.

I'll take a look at that when I get back to it soon.

Definitely at the stage of trying stuff out still so will see how that goes
__________________
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 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 04:47 AM.


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