Hello

I would like to have some advice, please!

The story:

I'm an acoustic guitar player. (Amongst other things).  I'm very picky about amplified sound, and experimented many many transducer systems.  None of them are perfect.  Having nevertheless found a system that's not bad at all, I still was unhappy about what we call "warmth".  Warmth is added by the body of the guitar, and even a soundboard transducer doesn't add it realisticly, and even causes problems at stage volume.  Tweaking frequencies with an EQ doesn't solve this problem: you can't add "warmth" to strings that doesn't contain low frequencies by eq-ing.  So all that does is unbalancing low strings.

So I had the following idea: I made a recording of the sound generated by the guitar body without strings, by tapping on it at the most sonorous location, and used this file in Pristine Space, and routed the guitar's sound through PS.  Tweaking the wet-dry balance gave exactly the result I expected: "warmth" was added to all notes, including the high strings, and the amplified sound was really closer to the natural sound than anything I tried beforehand.

I still am not satisfied entirely: there's still something a bit to "rough" in the high frequency range, and the dynamics are slightly wrong compared to unamplified playing, so I would like to go further than that, and imagined the following: recording the guitar both by a stereo couple of mic's, (or only one, if needed) and simultaneously from the transducer's output.  Then use sofware such as the deconvolver to obtain the sonic signature, so that I can use this in Pristine Space to recreate a sound as if the guitar was mic'ed, but using the transducer.  I mean, sort of comparing both signals, and determine how to process one to obtain the other.

Am I right in believing this can be done, and if yes, please tell me how exactly, I would be very grateful!

Thanks many
Dirk

I'm a bit sceptical this can be done with transducer/mic combo.  Indeed, you can then use deconvolution to extract guitar body sound from the mic'ed recording, but I think it may sound phasy/adjusted, etc...

What I *think* should work is placing a very small speaker in front of guitar (maybe one inch from the hole and directed into it); playing a sine sweep through it and simultaneously recording resulting sound with a mic placed at the desired acoustic location.

Then you can use this recording to deconvolve with Deconvolver.

Speaker you are using does not have to be high fidelity one - you may correct the frequency response later.  Something like a small omni speaker would be the best option (I believe it's possible to make a custom-built omni speaker from eight small tweeter-like speakers).

I suspect there can be other approaches to capturing guitar's body resonance.  But the one I've described is most easy and should not produce much problems (at least, theoretically).

Here's how Prof.  Angelo Farina form the University of Parma did it (seemingly very successful) with violins:

http://pcfarina.eng.unipr.it/Public/Papers/068-Ciarm95.PDF
http://pcfarina.eng.unipr.it/Public/Papers/069-Ciarm95.PDF
http://pcfarina.eng.unipr.it/Public/IMP-RESP/Violins/