Subject: Re: [linux-audio-dev] questions to be resolved
From: Jay Ts (jay_AT_toltec.metran.cx)
Date: Tue May 22 2001 - 22:50:51 EEST
Paul Davis wrote:
> I (Jay Ts) probably should not have written(!):
> >Give me a minute or two, and I can
> >probably think up some reason why someone will want to do that without
> >clicks.
>
> My stopwatch is running.
12:00 Arizona time zone (=PDT) :)
Seriously, I started to come up with something within a few minutes
of sending my last message, but I admit, it's taken me a lot longer
to develop the ideas and write them (like, 50 minutes)!
> I'll adjust for email latency :)
How kind. :)
This is going to be long-winded, which leads me to think that
maybe there's a much simpler and better example waiting to be discovered.
But, I thought I would put this out anyway, and see what you think.
One thing that I thought of was that in real hardware effects, clicks
when changing programs can be totally unacceptable.
For example, (and Paul, I know you will be able to relate to this, because
you have one too) my Quadraverb 2 is not able to change programs without
clicking if the audio is not silent during the program (preset) change.
I use it as a guitar (and other instrument) effects box, although it was
*really* designed as a studio rackmount effect box. (And the designers
might have thought, "oh, *nobody* will expect it to change without making
clicks... :)
So, I have a MIDI footpedal controller connected to the Q2, and loads
of great presets that I'd love to switch between during tunes, and the
only way to do it is to quiet the instrument, stomp on the foot controller
to change programs, and then start playing again. It isn't really a very
good solution.
(As an aside, just to show that Alesis is not alone in this limitation,
t.c. electronic produced the G-Force, which can switch between programs
click-free, and then later came out with their "improved" model, the
Fireworx, which cannot do that! So the G-Force is for live playing,
while the Fireworx is a "studio" product. Simply meaning that studio
users are not allowed to change presets midway through tracks!)
So, if a soundbox is being used for the software equivalent of a
"preset" in my digital effects box, and it is necessary to disconnect
one and connect another to change effects, then yep, a change in
the graph is necessary while audio is passing through.
Here is your cue to tell me that it is possible through other methods
to implement the same functionality without needing to change the
graph! And right off, I can say that if we have support for control
signals flowing along with audio, then the switch to turn one effect
level off (actually, ramp it down) while ramping up the level of another,
would have an equal function.
Now here's the kicker, as I see it. Although my Q2 does produce clicks
when changing presets, those clicks are caused by the DSP programs
suddenly being swapped. (E.g. from a reverb to a chorus - how could
that be done smoothly, without a crossfade? Which would maybe require
double-buffering.) However, that kind of click is *much* quieter than
the kind that would be produced by physically disconnecting the inputs
and outputs, and reconnecting them to another box. Or, the kind that
the Q2 would produce if it had a buffer underrun internally. And by
quieting the input, it is at least possible (although not ideal) to
work around the limitation.
> Oh, your
> answer should address why its a problem for them to carry out the
> operation before starting playback/recording. Thanks :))
Remember that I'm thinking about using the audio software as a
performance instrument. The changes will necessarily happen while
the audio stream is running. It's just a matter of whether or
not anyone would *want* to change the graph at that time.
An example might be a DJ who is playing with lots of different
effects during a long (hours long?) set of continual, unbroken
music. (Consider the Alesis AirFX, which seems to be designed
for DJs.) He/she may not be able to think ahead of time what
effects are going to be needed, and loading all of them might
not be allowed by the application or hardware's ability to support
that many simultaneous effects. Tens or even hundreds of effects
might not fit into the GUI windows very well, either. It would
"handle" much more easily if the DJ or musician could keep the
system simple, adding or deleting effects (and/or instruments or
other functions) at will, as he/she feels like it at the time.
Similar to the DJ, imagine a performing musician who is using
the LAAGA-enabled audio software to 'sculpt' the sound during
a concert, improvising in the moment. It would be similar to a
musician playing a modular synthesizer, and plugging and unplugging
patch cords during performance to change the sound of the instrument.
Wouldn't it be cool to have a modular synth (with real patch cords)
that could do that? Again, it *can* be done with a real modular
synth, and maybe there will be little clicks. (Caveat: I haven't
actually been able to try it.) But it could be managed such that
the clicks would not be ear-breakingly loud, and maybe not objectionable.
Well, I had to think pretty hard to develop this argument,
and honestly, I don't really care about the issue very much.
I get along pretty well using just one preset per song from
my Q2! Although I'd *like* the ability to switch smoothly.
- Jay Ts
jayts_AT_bigfoot.com
This archive was generated by hypermail 2b28 : Wed May 23 2001 - 18:41:55 EEST