< Back to IRCAM Forum

Questions about spat.oper settings

Hi everyone,
I’m working on a Max/MSP patch that allows me to automatize the spatialization of the sound from different sources. I’m using spat5.oper and spat5.spat~ for this, everything works fine and I’m happy about it. But as a beginner on the Spat environment I have a question about spat.oper.
What are the main parameters I should care of for the room settings? I made some presets that create the environments of the rooms where I usually work; I’m sending “/room/1/reverb/roomsize $1” and right now I input a value which Is between 20 and 50m³ (following the dimension of the real room), then I send the speakers’ positions and start working on the sources’ xyz. Is this the correct workflow? Should I check some other parameters to get a realistic spatialization? Why the default value for Room size Is 2000 m³ (which seems a pretty big number for it)?
Sorry if I’m not explaining well but I’m confused right now and I’m trying to understand.

Thanks in advance!

Hi Francesco,

In spat5.oper, the “main” parameters are represented by large dial buttons – such as “reverberance” and “room presence”.

Your workflow is fine. The “room size” is a handy meta-parameter (it affects all other delay-related parameters of the reverb engine) to quickly/roughly tune the reverb. Yet, it doesn’t have to precisely reflect a realistic space.

Spat is not based on a geometric model / simulation, anyway; it relies on a perceptual model that mimics certain attributes of real spaces. This perceptual model has been designed for relatively large spaces (concert halls, auditoriums, etc.); therefore the 2000 m3 default setting. And the reverb engine is not ideal for very small rooms (20m3 is really small !).
Finally, your reverb settings dont have to necessarily match the “environments of the rooms where you usually work”.

Hope this helps,

Hi T,
thank you for the quick response! In the meanwhile, I discovered a lot of parameters I wasn’t thinking of, such as radius and speaker distance in general correlated to the perception of the sound. I’ve read here that the optimal setup is always keeping the radius at 1m and keeping the speakers at a similar distance; I’ll put an example of how my presets look right now. These simulates the distance and position of the speakers that will perform my project (it is a work for University). Assuming that I have set this wrong, I’m now wondering if I should change the radius to 3m then use spat5.normalize, or change the speakers’ position to fit the 1m radius. Another solution could be changing spat5.spat~ to spat5.align~? (Which I also think has less impact on CPU load)

Thank you again!



One important precision : spat5.align~ is not a replacement for spat5.spat~.
spat5.align~ can be used in addition to spat5.spat~, in order to align/compensate the loudspeaker setup (in terms of propagation delays and gains).
For instance, your preset2 is kind of ellipse-shaped; spat5.align~ will compensate the different speaker distances so that it behaves as if they were on a circle.
See spat5.tuto-alignment-1.maxpat for further information.

Hope this helps,