Voxengo Premium Membership - All Voxengo Plugins For a Fixed One-Time Fee
Winter Holidays 2017-2018 Discount
Forums     Plugins     Impulse Modeler and Deconvolver unexpected behaviour


Last

Next

Previous

First



With new patterns is possible to emulate common stereo micing tech./XY,NOS, Blumlein.../Soon I discover a problem:-why front left cardioid rec.wall is picking sound same way as back right/for normal I should invert it/-and why there is almost no direct sound from bidirectional characteristic/other char.give normal direct spike at 8/ I suppose you should also better decribe what is exactly emulated by various patterns-is it volume only or proximity effect too, or phase invert in bidirect.pattern too...?Or it has nothig common with real microphones characteristics?Thanks,Karel.

It should be said that microphone patterns found in Impulse Modeler do not work exactly like in real microphones - these are just models and may not work good in all cases.  They also do not model proximity effect.

Karel is right.  This looks like a bug.  If the cardioid pattern is selected, any receptor that is assigned to the left channel is drawn properly: the little pointer in the middle points to the same direction as a real microphone would.  However, for the receptors assigned to the right channel, the pointer is drawn backwards.

To get the cardioids right, the receptors in Impulse Modeler have to point like this:

-\ -/

R L

while, I believe, it should look like this:

\- -/

R L

(assuming the sound comes from above)


You may try moving points around to get the right pointer direction.  It's really simple.

The pointer direction is definitely inconsistent for left and right with regards to phase at least.  Try setting up a Blumlein type pattern with two figure-8s at 90 degrees - it would appear that the pointer indicates direction of positive phase in one channel, but negative phase in the other channel.  To get round this you have to orient the two recepetors as if the are pointing to the side rather than the front.  Easy to work around, but very confusing.


Yes, I know, it all become pretty confusing when mic patterns were implemented.  Moreover, model I use to calculate impulses does not work really well with mic patterns other than omnidirectional one.

I do not think I can do much here in the near future.


Well, once I have figured out that I always need to flip the right channel around, all the cardioid flavours sound quite okay to me, and the width of the images changes exactly as I would expect for the different characteristics.  I tend to keep the recepting walls very, very short (grid off), and use lots and lots of rays.

Aleksey wrote:

Moreover, model I use to calculate impulses does not work really well
with mic patterns other than omnidirectional one.
What do you mean, simulating two figure-8s at 90 degrees works fine here (other than the pointer being inconsistent between left and right, which is easy to work around)?


The model I use is called 'monte carlo' method.  So, it can give some errors - though, I do not remember exactly, but there are some situations where IM simply returns wrong results for given mic polar patterns.  But maybe that does not give much of a trouble at all in practical situations.

It should be also said that I *could* use something like 'recepting spheres' (speculating here) instead of 'recepting walls' and so, results could be different.


For the future then...

I would love to see an updated version of IM, with more control over mic configurations, but you know that ;-)

Monte Carlo method sometimes gets it wrong - a bit of a gamble then :-)

This topic was last updated 180 days ago, and thus it was archived.  Replying is disabled for this topic.

Last

Next

Previous

First