[linux-audio-dev] Fwd: RE: [music-dsp] Transformer emulation revisited - Linux-audio-dev

7 Nov 2002

This is from the music-dsp list, it seems relevant to what people were
discussing with SPICE emualtion of tube circuits.

- Steve

----- Forwarded message from "Sergio R. Caprile" &lt;yogurthu(a)arnet.com.ar&gt;
-----

...
  Date: Thu, 07 Nov 2002 11:05:32 -0300 (ART)
 From: "Sergio R. Caprile" &lt;yogurthu(a)arnet.com.ar&gt;
 Subject: RE: [music-dsp] Transformer emulation revisited
 To: music-dsp(a)shoko.calarts.edu
 X-Mailer: XFMail 1.5.2 on Linux

 On 06-Nov-2002 music-dsp-digest wrote:

  I've been too lazy to sit down and calculate
the typical Preamp hi-freq
 boosting from schematic component values, need to do that sometime.  
 I did some Spice simulations some time ago, as far as I remember it was more a
 sort of low-freq roll-off: single pole hi-pass between stages and poor resistor
 decoupling.

  Even if the preamp has hi-freq boost that
emphasizes higher-frequency
 distortion, a downstream saturating transformer might "sound good" by
 selectively adding extra harmonics to the low notes? From one view, the
 transformer might be "undoing" the preamp's hi-freq emphasized distortion.
 But the two might work synergystically to make a more complex sound than a
 simple broadband fuzz?

 Since there is non-linear processing involved, the transformer is unable to
 undo the preamp hi-freq distortion, but it adds a different color. The main
 reason to hi-pass before distorting stages was to avoid that cutting rumbling
 sound or excessive bass, well, that was before grunge and Metallica and...

  There was no tone control setting that would
completely straighten out the
 lines in the tri wave, even with the amp running clean. There were
 frequency-dependent effects in distorted tones which made the waves even
 more difficult to interpret.  
 The Fender tone network (later adopted and modified by Marshall and almost
 everyone, including Roland) is designed to color the signal, there is no
"clean"
 position as in a regular Baxandall network. There is an intentional notch at
 250/500 Hz that can be softened by the mid control and two bumps at 100Hz and
 5KHz.
 I am not much into tubes, but most of the preamp stages where intentionally
 biased so the tubes could easily be overdriven. Some Marshall have "feature"
 capacitors running across different stages, I mean, the schems say: for model A
 this capacitor, model B no capacitor, model C, change to this value. Pretty much
 of the sound seems to have come out of sorcery ;^), or trial and error btw.

 James:
   You said you simulated a simplified tone control
   Did you simulate the Fender/Marshall network ? I tried to get the transfer
 function but it gets a bit complicated and I always give up before bilinear
 transform, and don't want to split the circuit as that might remove all the
 interactions along the controls.

 Regards

 -- 
 --------------------------------------------------------------
 Sergio R. Caprile, Electronics Engineer, Bs.As., Argentina
 http://www.geocities.com/scaprile
 --------------------------------------------------------------

 dupswapdrop -- the music-dsp mailing list and website: subscription info,
 FAQ, source code archive, list archive, book reviews, dsp links
 http://shoko.calarts.edu/musicdsp/ 
----- End forwarded message -----