Hi guys, so I made a thread about this a few months ago, but I thought I might start a new one as I have much more to show now !

The idea of the reverb is to go beyond traditional convolution reverbs sampled from real rooms, and use wave-based simulation to create a more natural sounding reverb without the artifacts from the speaker, and with features that are not really possible to do in the real world like instrument specific directivity and advanced field separation.

I'm looking for suggestions as to what kind of room, microphone and instrument placement I should consider in the plugin I'd like to develop in the future. The big downside with this technology is that it is much longer to do than conventional convolution reverbs, as not only does setting up the simulation is quite complicated, but the render times also last for hours and hours. So I want to be sure that whatever room and setup I'm starting to model is worth it, and will be useful for most people.

Here's a little video presentation I made about it :

And if you want to hear it in action in a full orchestral context, you can check out my midi mockup of the Tears of the Kingdom trailer :

Hope you guys find this interesting

 

Last edited: Dec 7, 2024

Reinventing SPAT again, I see.
Oh well, why not. It's not like they should have a monopoly.

 

I have absolutely no idea what it has to do with SPAT, but okay

 

Great video. A few things that popped into my mind:
1. I am extremely surprised that the de-facto approach to acoustics simulation is actually NOT solving the wave equation for the space in question. That geometric acoustics thing just seems so extremely sloppy.
2. Can I summarize your approach as creating IR's from solving the PDE's instead of acoustically capturing it in the room? Real rooms and air fluid are also not perfectly described by the PDE, much less by its numerical solution. This implies that your approach migt not be more realistic, just unrealistic in a different way. By the way, I'm nitpicking, don't take this particular comment too seriously.
3. The piano example was very impressive to me.

What I'm left wondering is how you're planning to give people a way of using these. Are you planning to simulate a few spaces that you like and just having users select the IR? I did not understand how you are adding the directionality in your examples, is it just panning in the input before the IR loader you are using? For example, imagine we want to add reverb with your technique to two different sources, one panned to the left and another to the right. Do you use one IR for each? That also eliminates the interaction between them.

If I had to give you one single personal opinion about preference on a potential plugin it would be that would prefer one single VERY good sounding space with all the different IR's needed for the interactivity about directionality and frequency response instead of having multiple different spaces with one or two IR's, which would make it just another convolution reverb but with IR's that come from solving PDE's instead of capturing in a room.

 

Oh, so you're the guy who made the video? I watched it yesterday because YouTube was pushing it hard for the last few days. Great stuff and actually pretty impressive on solo piano. I'm actually interested in trying it out once you have more to share. It could be actually a game changer since it might be able to achieve a lifelike performance with Pianoteq, and that could save me money in the long run (not having to get studio time to record on a Bechstein for some soundtracks where the piano is just too upfront).

If I had to suggest one place and positions, I would suggest to mimic the Teldex in Altiverb. Or maybe the Sydney Opera House. Those are my two most used IRs in Altiverb.