AFMG Network Forum

Jessica · Member

My attempts to begin the auralization process have been stopped dead: I can not get passed the creation of the Impact File. It seems as though no matter what combination of time, order and rays per loudspeaker I use I see nothing but yellow and magenta flashing by on my screen. Then when I view the total number of impacts it is just depressing. Following are the stats, what am I doing wrong?

560,000 cu. ft. (approx 15,860 cubic meters) cathedral - very traditional design: cross shape with dome over apex, etc.

First test was run with just 1 speaker and 1 listener to save time

Mean free path time was found as suggested in the tutorial and then multiplied by the order (I just went with the 10 used in the tutorial to start). This eventually led to a 325ms time on the second pop-up of the setups.

200,000 rays per loudspeaker = 95% chance of impact.

This led to 249 total impacts. That seems unacceptably low - or am I asking too much?

I've done several other iterations (including with more speakers and listeners) trying to adjust the time, order and number of rays based on the results, but I feel like I'm trying to hit a moving target.

Any advice would be greatly appreciated.

Jim Mobley · Member

Hi Jessica,

Are you checking the box to save the impact file? Can you pack and send me the project?

Jim Mobley · Member

Hi Jessica,

I've run several .mpc files using your model and I have some observations.

This is a large cruciform church with a shallow dome over the crossing and a half dome over the choir. Sabine RT is in the six second range.

1) The combination of Speaker 10 and Seat 3 that you were running has some quirks. The speaker is just above the platform and the seat is in the seating area near the center of the crossing. The platform itself blocks the floor reflection from the seat, and none of the dome faces return a direct path to the seat, so the first arrival is a sidewall reflection at almost 60 milliseconds after the direct sound. This of course is unrealistic and is due to the limitations of Ray Tracing. To auralize this combination successfully, I'd recommend an AURA Response calculation with Diffuse Rain switched on.

2) Other combinations of source and receiver don't exhibit such problems and could be successfully auralized, but you need to use longer calculations. I was able to start getting reasonable impact files with a length of 1.5 seconds, 40 orders and 2 million rays/loudspeaker. This calculation with all seats and one source ran for about 8 hours on one of the 2.6 GHz, P4 processor EASE classroom computers. Of course a statistical tail would need to be added. I found that the room has a distinct double slope to the decay with the early decay being much faster than the late. Using the Sabine RT values gave a more realistic looking tail than Predicted Impacts.

I do believe that you will have much greater success in this room using AURA to generate the virtual room impulse responses for auralization than you will with Ray Tracing.