Re: [linux-audio-user] virtual sampler

New Message Reply About this list Date view Thread view Subject view Author view Other groups

Subject: Re: [linux-audio-user] virtual sampler
From: Juan Linietsky (
Date: Fri Mar 22 2002 - 03:41:36 EET

On Fri, 22 Mar 2002 03:15:23 +0200
"xk" <> wrote:

> > ok, lets get to my point, i'm basically talking about mixer speed, but
> seems that my point was "changed" along this thread. What i'm talking about
> is that VOICE MIXERS (aka, the critical zone of any synth -oscilator or
> wavetable based-). _that_ is what it should be written
> > using fixed point since it is what it uses the least CPU.
> >
> > in MIDI, and usually, any kind of wavetable synthesis, the mixing consists
> of:
> >
> > 1-find nearer(s) val
> > -simply done with a right shift
> > 2-interpolate
> > -ONLY multiply, division is made by shifting
> > 3-filter
> > -yet again, only multiply, and then shift
> > 4-mix into left/right buffers - only multiply
> > 5-increment sample iterator
> >
> With your approach you always have to check for overflowing. This is far
> more worse than clipping. The user will have to reprocess an entire track
> for example just because one overflowed/clipped kick. This is why I said
> using floats is a time saver.

No, no overflow check at all, because you read samples 16 bits and mix into a 32 bits buffer. This is specially nice for tables of basic waves
or mixtrures (such a sine/wav/square/etc) because you can get the iterator
to wrap around naturally.

> > this IS slower using floats, I did made sure of that myself, i've written
> > mixers in both fixed point and floating point, and the SAME MIXER using
> fixed point WAS FASTER in machines ranging from old pentiums to
> Like I said in the first place, I don't know this for a fact. So you could
> be right.
> > >
> > > 23 bits is more than enough precision for digital audio, so this is not
> an
> > > issue.
> >
> > Allow me to prove you wrong, lets say we've got 23 bits, now.. 23 bits
> means numbers ranging from 0->8388608.
> > now let's say we have a sample iterator, in other words the offset to a
> sample being played. in each sample mixed, we must do step 5 of what i have
> described above, this is:
> >
> You are kidding, right?
> When I said 23 bits are enough I meant that they are enough for holding
> audio data in them, and I assumed you were also reffering to this.
> No sane person would use a float for a sample iterator. It's obviuos that
> the sample iterator has to be an 32 bit integer.

oh! my apoligies then. I made this point because iiwusynth uses float iterators, which is what i tried to point out earlier in this thread,
but i guess i didnt get to explain myself properly. I suppose the thread
went into the wrong direction.

I'm sorry about all this.


Juan Linietsky

New Message Reply About this list Date view Thread view Subject view Author view Other groups

This archive was generated by hypermail 2b28 : Fri Mar 22 2002 - 03:27:50 EET