Do hardware VA's alias as badly as software?

[hakey]

Try Oxium, pure crispy sound in the higher octaves

Even with audio rate FM? I doubt it.

...

[Aroused by JarJar]

nothing stopping Harware VA makers to filter out potential aliasing frequencies of the Oscillator waveforms.

...except that whole "sounding like an analog thing". Wide-open analog waveforms can have high harmonic content at extreme frequencies. 48k is insufficient for capturing the waveform output of my old Vermona discrete analog, and that's no Cwejman or MOTU. The first most obvious way to try to avoid aliasing is to generate waveforms that are more "dull" (fewer partials) than analog waveforms.

The possible ways of avoiding or reducing aliasing in "hardware" are the same methods as in "software". A "hardware VA" is a "software synthesizer" running on a little dedicated computer. 20+ years ago manufacturers and salesmen used to brag about this fact, nowadays they rely on people's ignorance to avoid this fact.

and then claim its aliasing free. just sayin'

You can claim honestly that oscillators are "alias free" but still have aliasing in actual performance, because an alias-free oscillator modulating another alias-free oscillator can generate a deluge of aliasing. Frequency modulation just with sine waves can easily generate aliasing.

is a Virus in 44.1khz or 96k hz ?

Dunno- I had an A, which I think ran at 44.1 or 48. It aliased and I suspect also had wavetable reading errors from tiny wavetables and probably not interpolating indexing, another factor giving it that thin harsh "plastic" sound.

Keep in mind that it is internal rates that matter- a synth might have 384k digital output, but if internally it is running at 48k it's still going to alias every time a frequency gets past 24k, and, less talked about, it's going to have other gritty digital artifacts if it is not interpolating, not running control rates at audio rate, etc. etc.

There's a lot to it.

[david.beholder]

There's actually huge and massive thread on gearslutz about software plugins aliasing, which nobody could read from beginning to the end, because of it's size, but... it provides lot of interesting info with occasional comments of andy cytomic and voxengo guy.

http://www.gearslutz.com/board/music-co ... hread.html

Not all software aliases badly

New buzzword for u-he old-school hardware va emulation technology: alias impersonation

[hakey]

48k is insufficient for capturing the waveform output of my old Vermona discrete analo

It's more than sufficient for capturing the harmonics that you can hear.

[Aroused by JarJar]

48k is insufficient for capturing the waveform output of my old Vermona discrete analo

It's more than sufficient for capturing the harmonics that you can hear.

You're not getting the point. Let's say we have a saw at 400 Hz, which is what, a about a soprano F#. Our sampling rate is 48k, so we generate waveforms which have no content above 24k to avoid aliasing. A 400 Hz saw has partials at 400, 800, 1200, 1600, and so on. Our waveform therefore can have 60 partials. At 400 Hz it's going to sound rich and "analog" when wide open. But lets say we go higher. Whoops, aliasing. Let's say we go lower, to 100 Hz. 60*100, so highest partial at 6k Hz. This is audibly not a wide-open analog saw, it is dulled in comparison. Now, let's say we want a saw that's "analog" sounding at 100 Hz. It has 240 partials in our system. But this saw taken to 400 Hz will have 180 aliasing frequencies!

So, in response to the post to which I was responding: it's not so simple. You either need dull waveforms, big oversampling, or an approach which varies generated partial content (or some combination of these).

And the approaches you can use are the same whatever the form of the digital processor you are using. "Hardware" offers no magic alternative.

[VariKusBrainZ]

Anyone can quote theory all day long. Without audio proof it's just hearsay.
I'm not saying aliasing doesn't exist, just that you are baboons who repeat other baboons nonsense.

Whats your problem kid?

Do we now have to experience everything personally in order for us to make acceptable statements.

I heard if I jump from the 5th storey window its likely to kill me, Im going to have to test that theory as it cant be true.

[hakey]

Let's say we have a saw at 400 Hz, which is what, a about a soprano F#. Our sampling rate is 48k, so we generate waveforms which have no content above 24k to avoid aliasing. A 400 Hz saw has partials at 400, 800, 1200, 1600, and so on. Our waveform therefore can have 60 partials. At 400 Hz it's going to sound rich and "analog" when wide open. But lets say we go higher. Whoops, aliasing. Let's say we go lower, to 100 Hz. 60*100, so highest partial at 6k Hz. This is audibly not a wide-open analog saw, it is dulled in comparison. Now, let's say we want a saw that's "analog" sounding at 100 Hz. It has 240 partials in our system. But this saw taken to 400 Hz will have 180 aliasing frequencies!

VA oscillators aren't sampled at a given frequency and then repitched. Rather, they're generated for each pitch, with the result that the cut off point relative to the fundamental (and thus the number of harmonics) will be different dependant upon the frequency of the fundamental.

[hakey]

"Hardware" offers no magic alternative.

Of course.

[Aroused by JarJar]

Let's say we have a saw at 400 Hz, which is what, a about a soprano F#. Our sampling rate is 48k, so we generate waveforms which have no content above 24k to avoid aliasing. A 400 Hz saw has partials at 400, 800, 1200, 1600, and so on. Our waveform therefore can have 60 partials. At 400 Hz it's going to sound rich and "analog" when wide open. But lets say we go higher. Whoops, aliasing. Let's say we go lower, to 100 Hz. 60*100, so highest partial at 6k Hz. This is audibly not a wide-open analog saw, it is dulled in comparison. Now, let's say we want a saw that's "analog" sounding at 100 Hz. It has 240 partials in our system. But this saw taken to 400 Hz will have 180 aliasing frequencies!

VA oscillators aren't sampled at a given frequency and then repitched. Rather, they're generated for each pitch, with the result that the cut off point relative to the fundamental (and thus the number of harmonics) will be different dependant upon the frequency of the fundamental.

Yes- but this is not "filtering out frequencies which will alias".

And VA's reading raw waveforms from tables do do "pre-dulling", they're not storing "naive" (aka theoretically perfect) saws and such.

[hakey]

"naive" (aka theoretically perfect) saws and such.

theoretically perfect = harmonically perfect?

[Lotuzia]

Try Oxium, pure crispy sound in the higher octaves

Even with audio rate FM? I doubt it.

First, Xils-Synths use Ring Modulators, not FM FM produces much more sidebands than RM, provided that the modulator signal is strong enough. Maybe you could demo Xils synths and understand what they do before speaking of them ?

Then, well, until now each time you had doubts, or should I say each time you attacked Xils-Lab with intentionally false information, like in the odf, then Oxium affairs, you were proved totally wrong, had to make excuses, to the point to declare yourself a -blind- 'fanatic' to justify your actions ).

So, very few have changed since this time uh ? you were only "recovering"

Some guys just never learn their lessons I guess. ( Though eventually having only doubts, wich is a delicate way to declare that we could be lying, or just be "as stupid as usual", can be seen as a WIP Vs spreading false information )


" Our oscillators are aliasing free, which gives an incredible pure and warm "analog" sound. The DSP algorithm behind the oscillators is an advanced improved version of the well-known BLIT (Band Limited Impulsion Train). We would call it ABLIT (Advanced Band Limited Impulsion Train), but we think that the result is more important than the words : see the XILS 3 comparative screen shots for exemple."

(Quote From XIls site)

So well, believe it, or not, and have doubts, or not (especially about things that do not exist ). Who cares. People who really use our synths already know the answer, while those who might be concerned by this particular subject can easily demo them. End Of/

-------____

Btw synthesizers with aliasing can still be good. Some famous FM synths and their software counterparts have aliasing. Still, wonderfull synthesizers/musical instruments.

[hakey]

AJJ, I think you're argument is either wrong, or perhaps just confusing.

In stating that 48kHz is insufficient to capture all of the harmonics of a real analog waveform and then arguing that repitching a sampled waveform will result in either aliasing or a duller waveform you appear to imply a connection between sampling a waveform and generating one.

Sampling and generating a waveform are not the same process.

[Aroused by JarJar]

"naive" (aka theoretically perfect) saws and such.

theoretically perfect = harmonically perfect?

Don't know what you mean by "harmonically perfect". I just meant for example that a theoretically perfect saw would have zero rise time (instantaneous change from -1 to 1) and an absolutely straight direct ramp between 1 and -1. This doesn't happen in reality, as the finest analog doesn't operate at infinite speed and bandwidth. In digital this waveform is represented by a "naive" line from 1 to -1. This aliases like a hell.

[Aroused by JarJar]

AJJ, I think you're argument is either wrong, or perhaps just confusing.

In stating that 48kHz is insufficient to capture all of the harmonics of a real analog waveform and then arguing that repitching a sampled waveform will result in either aliasing or a duller waveform you appear to imply a connection between sampling a waveform and generating one.

Sampling and generating a waveform are not the same process.

I think it's rather a matter of you're being confused by your own intent to pick at whatever I say rather than actually try to understand it. All I was doing was pointing out that it's not so simple as "filtering out frequencies which will alias".

And you still seem to be suffering from the delusion that all VA variably generate waveforms. No, some VA's simply read from wavetables. In all likelihood, judging from my memory of the Virus A, it probably did just that and doesn't even approach the sophisticated methods common in plugin synths these days.

[hakey]

First, Xils-Synths use Ring Modulators, not FM FM produces much more sidebands than RM, provided that the modulator signal is strong enough. Maybe you could demo Xils synths and understand what they do before speaking of them ? Maybe you could demo Xils synths and understand what they do before speaking of them ?

So audio rate OSc>Osc and Osc>Filter Cutoff modulation isn't possible in Oxium?

Funny how, having tried it before writing that post, it looked to me like the mod matrix allows such routings and how the results do indeed sound like audio rate FM with plenty of aliasing.

Must have imagined that.