AFMG Network Forum

hawk · Member

Hi

I am trying to calculate the reverberation time of a room using ray tracing and Schroeder's backward integration method. The room is rather big (V=2455.56 m3) and the geometry is asymmetric.

I have tried different ray tracing calculations with different 'Impact Control Time's (ITC) and 'Trace Control Order's (TCO):
-First I used ITC=200ms & TCO=10 (mean free path time is 0.02s),
then ITC=300ms & TCO=15,
then ITC=500ms & TCO=25,
and finally, ITC=1000ms & TCO=50;

Schroeder Reverberation Times keep increasing as I increase the ITC & TCO (and as a result, rays per loudspeaker and of course calculation time!)

I'm curious to know which one is more accurate and how far should I go to attain enough accuracy?

AFMG Pedro Lima · Forum Moderator

Hawk,

How many rays have you set?

In the Probe window, can you see if you are already inside the diffuse field, from the Reflectogram or the ETC and the Schroeder curve? The integral must include energy from the diffuse field only.
Is the reflectogram dense enough?

Of course, you must have a room where the Reverberation definition is not questionable anyway, for starters... Wink

In general the results must converge as you increase the the time span and the order control. However, you must start from a minimum accepted.

Yes, calculation times for ray tracing are a factor that must be carefully dealt with whenever modeling the room. Having that in mind will help you on decisions about simplifications for the model. How many Faces does your room has? What are the avg Sabine/Eyring RT times?

Regards,
Pedro

hawk · Member

Dear Pedro,

First thing, thanks for your timely response. Smile

The number of rays were different for each calculation, I tried to keep the Impact Chance above 80% though, according to the EASE 4.3 user's guide. For instance, for the 500 ms time limit & 25 order limit, I used 250,000 rays with an impact chance of 83.5%

I think the field is diffuse and the Reflectogram is pretty dense, here it is for the aforementioned calculation:

As you can see, the Schroeder RT is around 1.5 s, but Sabine & Eyring RTs are much higher, 3~4 seconds on the average. I think 1.5s is closer to the actual value, and as the room is a workshop with machines placed randomly all over it (I have included the machines in the model), absorption coefficients are not uniformly distributed so Sabine & Eyring approximations are not accurate.

The number of faces: 26 faces [main surfaces of the room] + 7 faces [windows and doors, coated and 2-fold] + 162 faces [27 machines modeled as 6-faced cubes] = 195 faces in total

For the record, I have carefully followed the instructions given in the user's manual when building the model, the room is closed and I have double checked everything about the model itself.

Best Regards,
Hawk

AFMG Pedro Lima · Forum Moderator

Hawk,

Can you please send the Packed Project to support@amfg.eu, assigned to me? This way I can have a closer look.

Please include a description of the problem and the Ray Tracing files you generated.

Best Regards,
Pedro

hawk · Member

I can send the packed project, but the Ray Tracing (Impact) files are too large, like at least 100+ MB and some even more than 1 GB!

AFMG Pedro Lima · Forum Moderator

hawk,

So far I haven't received the Project. Were you able to send it (without the tracing files)?

Best Regards,
Pedro