I've been using Deconvolver for years to capture the IRs of churches and halls where I make recordings - it's been extremely useful, thank you.
If you are planning on doing any further development, I wonder if you might consider the possibility of modifying the the deconvolution routines (and the test tone creation), so that IR capture would be the result of multiple passes of a sweep, then averaging the result to create a single impulse response. I believe that this could be of great benefit when dealing with relatively noisy environments, where intrusive noises like creaks and thunks (short duration) as well as passing cars and trucks (longer duration, but not steady state) would introduce less pollution to the outcome.
What got me thinking along this line was remembering my experiences with a MLSSA system in the early '90s, which used a test tone of "pseudo-random" noise based on a maximum length sequence, which audibly cycles through several times. At the time, I was impressed by its resistance to spurious noises in that we were taking impulses of a hall undergoing renovation, with the attendant hammers and saws at work while we were collecting.
I attempted to create my own test tones as an experiment with Deconvolver: four 3 second ascending tones in succession, and another which had the second and forth sweeps reversed, so that it went up, down, up, down. Upon deconvolution, the four ascending sweep test yielded an IR with three IR clicks/reverb tails; the up-down test yielded an IR click/tail, followed by what sounds like an aliasing sweep (simultaneously ascending and descending), then another IR click/tail, then the mirrored sweep sound. Clearly, in its current state, Deconvolver does not know what to do with a multiple pass test tone, and has even less success with a descending tone.
Lockwood ARS - http://LockwoodARS.com
Quality music recording services for classical and acoustic music.
Toronto, ON, Canada
Host DAW: Merging Technologies' Pyramix 11.0.4 (64 bit) AND 9.1.10 (32 bit)
System: Windows 7 SP1 64 bit on Mac Mini (6.2) via Bootcamp 5.1.5621
Windows 7 SP1 32 bit on MacBook Pro (3.1) computer via Bootcamp 4.0.4033.
Curve EQ, Deft Compressor, Elephant, GlissEQ, HarmoniEQ, LF Max Punch, Polysquasher2, Polysquasher3, Soniformer, TransGainer, Voxformer; also Deconvolver.
What you are asking for can be achieved by using a longer sweep, it's mathematically similar and will likewise remove spurious noises.
I do not see a reason to use reversed sweeps, no benefit here.
This topic was last updated 180 days ago, and thus it was archived. Replying is disabled for this topic.