Voxengo Premium Membership - All Voxengo Plugins For a Fixed One-Time Fee
Forums     Plugins     Impulse Modeler and Deconvolver Direct Signal Level


Last

Next

Previous

First



Can someone please explain how "Direct Signal Level" works?

Specifically,

1) What is the meaning of the number, as in "Normal value is 8"?  Is it an arbitrary scale or some ratio or what?

2) What does "normal" mean?  Does it mean that the direct rays are weighted exactly the same as the reflected rays?

3) Does the direct ray follow the frequency-dependent attenuation as set in the properties of ** AIR **?  If the direct ray passes through some transparent curtains, is it also correctly attenuated?  Or do I need to manually decrease the Direct Signal Level to emulate the receptors distance from the emitors?

Thanks for any help.

(Needless to say, an absolutely amazing product!)


This 'Level' is a relative value.  Since with room model I've used in Impulse Modeler all signal levels are relative, I had no means to automatically set the direct signal level you get in the calculated file.

This value is something like a scale parameter, which is not directly connected to any other parameters.

By 'Normal' I meant estimation on how loud a direct signal should be (but this is still not precise).

Direct ray is also affected by frequency-dependent air absorbtion parameters.

I should also say that I may review this 'direct signal level' concept in the future, especially if I can find some unbiased formula to calculate it automatically.


Thanks Aleksey.  I suspected this was something not easily describable.

I have found a trick to balance the levels to taste: it requires creating two impulses.

Start with the full desing with all the walls, and postion the emitor and the receptor where you want them and capture the impulse (without the direct ray).  Then, remove all the walls, such that only the direct ray can hit, (since there are no reflections), and capture the impulse for that as well (with the direct ray, any level).  Finally, create two new tracks by convolving the original track once with each impulse.  Treat the track convolved "with the walls" as wet and the one with "no walls" as dry, and mix a desired relative balance.  No guessing about the direct signal level there.  Note that there is no additional panning involved since the no-walls-convolved channel is already coming from the "correct" location and distance.

It has also occured to me that you can have designs where there is no direct ray (e.g., there's an obstacle between the source and the receptor.) I am not sure how that should behave, that is, what 'direct signal level' might possibly mean in that case.


tonyd wrote:

Thanks Aleksey.  I suspected this was something not easily
describable.
I have found a trick to balance the levels to taste: it
requires creating two impulses....
Interesting idea, but why so complicated?  Do you find that the "direct signal level" parameter does not give you enough control over the wet-dry balance?


Hi Andrew,

Yes, the direct level does provide the same control, however, the results are far from immediate. ("Set the level, generate the impulse, load into Sonar, listen, hmmm... a touch too dry, back to IM, decrease the level a bit, etc. etc." ) What I described can be tweaked in real time, even automated.  I am still only beginning to experiment with this particular width-depth placement technique, but to my ears the results sound more convincing than anything else I have tried.

Tony


tony, do not forget that Pristine Space automatically reloads impulses.  So, after recalculation in IM Pristine Space will automatically load that updated impulse after several seconds.

Tony, try what Aleksey suggests - the way that Pristine Space automatically reloads an impulse response when it's been altered by Impulse Modeler makes the "direct signal level" changes effectively real-time.  It works really well.

Well, some of the impulses take several minutes to generate.  I guess I need to find a better way of working, sort of quick preview mode (random, fewer rays, small FFT window) and switch to the hi-res mode only at the end.  As I said, I'm just beginning to explore this new paradigm of creating spatializations.

What amazes me about going the IM way is I can eliminate most of the other processing (like low- and high-pass I'd normally apply on the send to a conventional reverb, or an additional delay on a separate bus).  If the materials and ** AIR ** are well tuned, all this just "comes out", the impulse already contains this!

There is a theorem in digital signal processing that any linear, causal and time invariant filter is a convolution.


tonyd wrote:

Well, some of the impulses take several minutes to generate.  I
guess I need to find a better way of working, sort of quick
preview mode (random, fewer rays, small FFT window) and switch
to the hi-res mode only at the end.  As I said, I'm just
beginning to explore this new paradigm of creating
spatializations.
I agree that is an issue - it would be a nice time saver if the number of rays could be set globally rather than for each individual emitter.


Should I implement this as a global setting maybe?
This topic was last updated 180 days ago, and thus it was archived.  Replying is disabled for this topic.

Last

Next

Previous

First