Re: [linux-audio-dev] [ann] caps-0.3.0

[Giuseppe Zompatori]
>Speaking about the compile flags you use for caps -O6 does nothing
>more than -O3 and from -O2 upwards there's no need to specify
>"-funroll-loops", also -O3 often bloats the generated code
>unecessairly making it slower than -O2 AFAIK.

To tell the truth, I've never evaluated different -O settings (neither
have I looked up their implication of other optimization flags). I'll
keep it in mind, but it's not the highest priority. Also, the benefits
of different -O and other flags might differ on PPC, x86, x86+sse,
x86+sse2, x86-64, amd or intel etc.

I'd happily be convinced by some hard facts (code size vs. execution
speed) though. :)

>Why introduce PSU emulation (not that it's a bad thing in itself) when
>you're still not simulating what makes a guitar sound sweet, namely
>output valves?

Why yes, AmpV actually models a (granted, _very_ simple but
nonetheless sweet to my ears) approximation of output stage
saturation.

>I've been looking at the source code and it seems you're just
>simulating a single 12AX7 preamp having weird tone stacks... In
>reality I've never seen something like that in other than valve based
>distorion/pedeal boosters...

True, it's a single preamp tube. There used to be two, just like in
the real circuits I've analyzed, but the effects of the first one are
far too subtle to make a palpable difference in tone. Thus, it fell
victim to optimization.

>It would be cool to have amps based on real good sounding amps models
>where you have different preamp stages with the tone stack placed at
>different gain stages, or more simply preamp voicings, like say the
>difference you find between fender, marshall, vox,mesa boggie and
>soldano preamps sections, and then model the output stage with the
>same valves arrangements and data from the actual valves like say EL34
>for marshalls and 6L6 for fender/mesa/vox/soldano/whatever...

You're forgetting the influence of output transformers on the tone,
which is just as important as the power amplification tube stage, plus
the non-linear effects of the final output speakers! :)

Not that I have anything against modelling all these effects better.
Yet, we're talking about a few months of work here, mostly analysis
of real circuits and speakers in theory as well as in operation. And,
see below ..

>It is in the output section valves that the real "juice" comes out in
>a real amp, and it's the place where the different valve types and
>arrangements can make a huge difference, like class a vs cass ab,
>single ended vs parallel and push pull...

My opinion is that these effects only become a major tone-forming
factor if the amp in question is driven _real_ hard (master gain close
or at maximum, up to this the data I have gathered from my real amp
indicates it's all fairly linear).

I've always been playing small amps at moderate volume in real life,
and the only time I ever played one in the vicinity of maximum output
power, the output transformer melted down within minutes. Not my cup
of tea, thank you. (I didn't even like the sound while it lasted. Go
figure :)

>I quickly gave a try to AmpV and AmpIV using a DI box, and had issues
>with the gain settings of those plugins, the more I rise the knob the
>more it gives me a farty distortion up to the point where the sounds
>disappears completely. I've tried this with gcc 3.2 and gcc 4.0.2.

Have you tried raising the AmpV 'watts' parameter to its maximum
setting? The effect you describe should be completely gone by then
with AmpV. Also, AmpV has never been designed with high-gain
scenarios in mind.

With AmpIV however, what you describe is not the expected behaviour.
Does the AmpIV from caps-0.2.3 give the same effect? In any event, can
you send me your AmpIV settings (and other plugins' settings making up
the FX chain) and a short input signal sample so I can recreate the
problem?

>Due to frustration with free as in freedom guitar amp sims, I recently
>bought one of those
>http://www.behringer.com/LX1-PRO/index.cfm?lang=ENG, it's a pity
>there's nothing even coming close to it in the free world up to now :(

Well, glad you found an escape route, and sorry for wasting your time
with my poor efforts! :)

In the end, it seems the two of us are after a different kind of
guitar tone. I've been playing AmpV in a jam lately and not once over
the course of the session have I felt the need to fiddle with the
sound (and that's a first). I'm more than happy with it, sorry to hear
you're not.

Thanks for your feedback, I appreciate it,

Cheers, Tim
Received on Thu Jan 12 16:15:08 2006