Voxengo - Sound with a Soul
Professional VST PlugIns and
AU PlugIns

Support Forums Shopping Cart
Equalizer Plugins Compressor Plugins Saturators, Guitar Amps All AU, VST Plugins Reverbs, Spatial Imaging Mastering Plugins Harmonic Enhancers Free VST, AU PlugIns Track Plugins Delay, Echo, Modulation Utility, Denoising, Analysis Drum and Cymbal Samples
Special offer - CurveEQ introductory price
Forums     Plugins     Impulse Modeler and Deconvolver
Deconvolver help -> dc IR does not match the source / optimal sweep length

This topic was last updated 180 days ago, and thus it can be considered old.  Replying is disabled for this topic.

Dear Aleksey,

i am interested in the Voxengo Deconvolver but i have some questions before purchasing:

I ran the Trial but my test-results were not pleasing yet.  Maybe i am doing something wrong.  Well, i am just a musician with limited technical knowledge so I would be glad for some tipps:

My aim is to create specific IRs (EQ-Curve and Cab) to import into a Fractal Audio Axe-FX Hardware Unit.  The deconvoluted files have to be mono, 48Khz, 24 Bit, 1024 Samples.

As an easy starting point i wanted to create IRs of 3 different VST Plugins within my DAW (48Khz and 24 bit)
(a: EQ, b: Reverb, c: Cab-Simulation):

1) created a 6s long Sweep with fades using the Voxengo Deconvolver
2) imported the file into the DAW and added a VST Plugin
3) reduced the Audio-level of the sweep to receive a nearly -0.3 db wet output Level (after VST FX)
4) set the Left marker at the beginning of the Sweep-file and right marker at the end of the sweep +1 sec
5) exported the 100%-wet Signal (length 7 seconds) mono, 24bit, 48Khz
6) deconvoluted with the Voxengo using REVERSED Technique
7) duplicated the original Sweep within my DAW to a second Track and added the SIR2 (even tried Waves IR1)
8) imported the dc-File into SIR2 and compared the output signal of both tracks
=> I thought both tracks should now be nearly identical (Track 1: Sweep + VST; Track 2: Sweep + SIR2 with imported dc-IR)
9) compared the signals acoustically and with some Audioanalyzer and there a big differences in the signals

a) the dc-file of the VST-EQ was nearly inverted and the highs were cutoff, the ir was completely unusable
b) the Reverb was less loud and the predelay was slightly different but it worked okay, i would say 85% match.

Maybe i am doing something wrong, i would be glad for any tipps.

My last question is about sweep-file length.  My final IRs should be max 1024 Samples (21,33 ms).
So what would be the best value for the Voxengo deconvolver to setup the sweep file?  0.02 s with fades, ore am i on the wrong way?  Would be glad for any tipps.

Thank You very much.

cheers - Nick

Your routine seems correct, but I would suggest you to add 1s of pre-silence as well as 1s of post-slience you've added already.  If the VST EQ/VST FX is non-linear the deconvolution will fail - it works with linear filtering only.

The 6s sweep is fine.  But you will have to chop-off the resulting IR file to 1024 samples - this may affect the sonic quality, of course.
This topic was last updated 180 days ago, and thus it can be considered old.  Replying is disabled for this topic.

Copyright © 2002-2012 Aleksey Vaneev - Terms of Use - Privacy Policy - About Voxengo - Questions & Answers