The Line 6 "Pod" has been the DSP guitar-amp-modelling standard in
recent years:
http://www.musiciansfriend.com/srs7/sid=040711183006068154015079244725/ g=home/search/detail/base_pid/482197/
This has been the popular substitute whenever an real amp can't be
used, but I haven't heard of it putting a dent into real amplifier
sales.
I heard it a year or two ago. I did not think it sounded all that
great.
I have played around in this area, and I have noticed a few things:
1. The desirable sound results heavily from a combination of soft
clipping and hard clipping. Other nonlinearities are much less important.
2. It is often good for the signal should have a fairly spiky or "bright"
waveform prior to soft clipping. One desirable feature of "Fender" amps
is "darkening", in which a spiky waveform is squashed in a way that
reduces treble content when soft clipping occurs. This requires a soft
clipping milder than that achieved with silicon diode clipers.
Silicon diode clippers, in combination with hard clipping, make a good
sound like that of some "Marshalls". The soft clipping is harsher there
and produces more noticeable harmonic content, so "darkening" is not as
apparent. Many diode clippers have their insufficiency being a failure to
include hard clipping. And my best results were from having a soft
clipper follow a hard clipper rather than precede a hard clipper.
3. Much of the sound is filtering characteristics of the loudspeaker.
Modern DSP should be capable of convoluting the signal with an impulse
response of an actual guitar amp loudspeaker (fed the impulse through an
appropriate impedance), cabinet, and even a room and microphone.
4. Part of the sound of "Marshall" amps comes from the amp's output
impedance getting fairly high during distortion. When the amp is running
clean, the output impedance is lower. Frequency response of a loudspeaker
varies with the amp's output impedance.
When attempting "Marshall" simulation (including the closed-back
cabinet) without doing anything about varying output impedance, I have
found a good approximation by using (or modeling) the loudspeaker being
fed through a resistor of value somewhat more than the loudspeaker's
nominal impedance. A "Marshall" cabinet has less bass content when fed by
a low impedance source.
This was less of a factor with open back cabinets.
5. Although simulating nonlinearities of the loudspeaker are necessary
for a totally realistic sound, I did not find this necessary for a good
sound. As part of this, I did try a tube amplifier output stage (combined
with op-amps) with ridiculously low supply voltages (as low as 12 volts)
but normal heater voltages. This does work, although maybe not always.
- Don Klipstein (
[email protected])