AL-1 vs Virus question...?

shawnhar · Post by **shawnhar** » Thu Nov 03, 2011 8:55 pm

Hedegaard wrote:But alas, this is again, distorting the original purely mathematical saw wave that preexists in all synths. Its the application of e.g. filters, component quality, envelopes e.t.c. that alters the saw wave shape to "something-else-but-sounds-like-a-saw-wave."

Not really.

No true analog synth produces a pure mathematical waveform, even leaving out such issues as filters and oscillator drift. A perfectly tuned oscillator with a fully open filter will come close to that mathematical ideal, but never exactly match it, and it is in these small differences that much of the interest and musicality of the instrument is to be found.

In the digital world, it is certainly possible to produce a mathematically perfect waveform (and many romplers will include exactly such a 'perfect' single cycle wave as one of their source samples) but it turns out that such waveforms don't make for a good sounding synth.

This is one of the things that distinguishes a quality VA from a crappy cheap digital synth: the better implementations go to great pains to make waveforms that sound good, as opposed to 'perfect' but sterile.

Bertotti · Post by **Bertotti** » Thu Nov 03, 2011 8:59 pm

I thought Bond killed Bloefled!

Seriously though would some one define aliasing in this context and what is blit and blep?

shawnhar · Post by **shawnhar** » Thu Nov 03, 2011 9:11 pm

Bertotti wrote:Seriously though would some one define aliasing in this context

Aliasing is what happens when you digitally manipulate a signal that contains frequencies higher than your sample rate is able to accurately represent.

The highest frequency that can be represented by digital audio is half of your sample rate. So at 44k, you cannot manipulate sounds with frequencies higher than 22k. At 96k, the highest possible frequency is 48k.

But, when you take a signal with higher frequencies and transfer it to digital representation, these impossible to represent higher frequencies are not just cleanly discarded. Instead they will cause all sorts of artifacts and unwanted ringing/distortion, mostly around the higher frequencies but potentially also in lower (and thus more audible) spectrum depending on the details of how the conversion is implemented.

This is why a good AD converter is so hard to make: to avoid unpleasant aliasing, a good implementation makes smart use of filters to remove these problematic frequencies entirely before the signal is converted into digital form.

Even purely within the land of digital, aliasing can occur whenever you manipulate a sound, for instance resampling to a different frequency (ie. changing the pitch of a sample), or applying any kind of DSP effect, or evaluating a digital oscillator. The math of these operations produces potentially infinitely high frequencies (even the humble square or saw wave includes an infinite number of infinitely high frequency bands!), and this will audible as aliasing noise unless the implementation does something clever to avoid that.

Much (I would even say most?) of the skill of implementing digital audio is in the various techniques used to avoid aliasing.

Or of course there are some instruments like the PPG Wave that just let the aliasing happen, and embrace it as a unique part of their sound. This can sometimes be quite musical especially if the digital aliased sound is fed through an analog filter, but it's a very distinctive character that is by no means always desirable!

Bertotti wrote:and what is blit and blep?

https://ccrma.stanford.edu/~juhan/vas.html

Bertotti · Post by **Bertotti** » Thu Nov 03, 2011 9:37 pm

Thanks for the link! And explanation!

Hedegaard · Post by **Hedegaard** » Fri Nov 04, 2011 6:59 am

Yeah, thanks for all the feedback.

I had thought the ideal was to create a perfect saw wave and then use that to manipulate into usable sound.

Its nice to be wrong - you learn something new

EvilDragon · Post by **EvilDragon** » Fri Nov 04, 2011 9:27 am

That indeed WAS the idea, however since nothing in this world can be made absolutely perfect, you end up with a lot of variations on the main idea.

Let's take the filter as an example. There is no ideal filter! The ideal filter wouldn't have a transition slope at all! What's a transition slope? It's the segment of the filter which connects the pass-band and the stop-band. That transition simply cannot be instant in analog world (and in digital world it would require a tremendous amount of processing power), so it's made gradual, and hence you get the designations such as 12 dB/oct, 24 dB/oct, which tell you by how much does the signal get quieter after cutoff frequency of the filter.

So, to get closer to the characteristics of the ideal filter, several approximations exist. Perhaps you've heard about Butterworth, Chebyshev, Cauer (elliptic) approximations. They all have their positive and negative sides.

For example, Butterworth's is the best approximation, but it has a really slow transition slope, so it needs a really high order to approach an acceptably short transition slope.

Chebyshev has a faster transition slope, but since it's formula is cosine based, it has a ripple in the pass-band (which means that the signal which is let through the filter does NOT have a constant amplitude throughout the pass-band, which is a problem). There are two types of Chebyshev approximation - type 1 has the mentioned amplitude ripples in the pass-band, and type 2 has those ripples transfered to stop-band, which is a lot more acceptable. Also, depending on filter order, the frequency of the ripple oscillations gets faster, which is also not nice.

Cauer (or elliptic) approximation has the fastest transition slope of all mentioned approximations, but it has ripples BOTH in the pass-band and stop-band, which is not nice for audio applications.

Here's a comparison image, so you all know what I'm talking about:

jimknopf · Post by **jimknopf** » Fri Nov 04, 2011 1:41 pm

Korg still has three outstanding synths to dream of for possible future modelling:

- the PS 3300, probably hard to model
- the Mono-Poly. If this would be modelled as well as the Polysix and the MS-20, we would get a very nice synth for leads and basses.
- the Trident

I would certainly be ready to pay well for any of these three possible extensions. A Mono-Poly would be a great start!

aron · Post by **aron** » Fri Nov 04, 2011 5:04 pm

>A Mono-Poly would be a great start!

I agree. I bet it's coming.

shawnhar · Post by **shawnhar** » Fri Nov 04, 2011 5:08 pm

Hedegaard wrote:I had thought the ideal was to create a perfect saw wave and then use that to manipulate into usable sound.

Attitudes about this have changed a lot over the years.

Back in the day, that was exactly the idea. Laurens Hammond set out to create an additive instrument based on pure sine waves, and viewed the fact that his tone wheel design could not quite produce perfect sines as an unfortunate flaw. Similarly, early synth designers would most likely tell you that their goal was to create mathematically perfect waveforms. Of course that was impossible, so they had to choose between various compromises. Some chose wisely and produced instruments of great musicality, while others understood less about what was going on, yet serendipitously landed on designs that provided interesting characters. And others chose badly: these we hear of less today as they went bankrupt long ago

But technology improved. Analog was replaced with digital. It became possible to make actual pure waveforms without any of these compromises.

Shock horror! Turns out that if you build a Hammond clone using pure sine waves, it sounds terrible. Flat, boring, just not an interesting sound in any way. Laurens Hammond was wrong: the 'flaws' in his implementation were actually the very thing that made it such a classic instrument! Hence modern clonewheels go to great pains to carefully emulate each of these 'imperfections'.

Same goes for synthesis. A perfect digital waveform, played through a perfect digital filter with perfectly accurate digital LFOs and envelopes, just doesn't sound very interesting. So modern digital synths go to great effort to make things imperfect. Sometimes they emulate the details of a classic analog synth, other times they strike out on their own doing new things in a different way, but we've learned that a mathematically perfect waveform is not the end goal, but actually just a starting point for the process of messing things up in interesting and musically fulfilling ways.

hermanmusic · Post by **hermanmusic** » Sat Nov 05, 2011 12:10 am

I own an Oasys and a Virus TI Desktop. I find that the Virus complements the Oasys beautifully. For me, in one word, it's the "sound". For the most part the Virus seems like an extension of the Oasys.

I use my Oasys for sequencing rather than a DAW and find it adequate.
But, you may not if you're used to the DAW environment. Read the manual for the Kronos (Sequencer section especially) and see if you can live with it.
I sequence exclusively on my Oasys and with the exception of limited MIDI track count (16), and the timing resolution (192 PPQN), it is fine for me. I believe the Kronos sequencer sports 480PPQN. It would also be nice if it could send/receive Midi Time Code (instead of just Clock).

If you have an opportunity to play the Kronos and the Virus side by side it wouldn't surprise me if you started pillaging around the house for stuff to throw up on eBay!
My opinion is that one is not better than the other.
It's like chocolate fudge in the chocolate ice cream: one is not better - they just compliment each other so tastefully. Or, you might say, that in this case the sum of the parts IS greater than the whole!

Good Luck!
David -

RonF · Post by **RonF** » Sat Nov 05, 2011 12:49 am

Owner of a Virus TI Polar and Kronos.....

Virus sound is great.....its architecture is deep.....its effects are epic....its programming interface is outstanding.

Kronos AL-1 is deep, its effects are more diverse than Virus, but its reverb is not as good as Virus. Its programming interface is quite simply not as good. Biggest complaint about AL-1 is its lack of a dedicated mod matrix screen where you can see an overview of your connections. It makes programming Kronos synths a chore at times.

Neither sounds better than the other.....they both have a unique sonic signature. I wouldn't trade one for the other (strictly focusing on AL-1....because if I could only have one unit, I would pick Kronos for the REST of its diversity).

Morshu · Post by **Morshu** » Sat Nov 05, 2011 7:47 am

the virus ti 2 is the god of synths, you KNOW you want it. the kronos sounds spectacular, sometimes, but the build quality is bad and its very similar to the oasys, whic used ancient technology. choose the virus.

robinkle · Post by **robinkle** » Sat Nov 05, 2011 9:29 am

I agree with Ron F there.

jimknopf · Post by **jimknopf** » Sat Nov 05, 2011 10:11 am

All boasting is futile, and something like the "god of synths" does not exist.
And by the way, the Access TI2 sounds exactly like the TI, with additional processing power for higher polyphony.

My impression is that real longtime Access Virus owners know very well what they have, and neither magnifiy this music making tool into myth making, nor are ignorant about what it can do and really is good for - like some reviewers who obviously can't handle it and didn't even manage to use the power of total integration (sorry, EvilDragon, couldn't resist after reading your linked "review"

).

Ron, concerning the reverb of the TI: it sounds well, but is pretty basic. Don't you think the Kronos effect section offers much better reverbs? For my ears, the Lyle Mays Piano on board of the Kronos for example showcases the best reverb available on synths and workstations so far.

robinkle · Post by **robinkle** » Sat Nov 05, 2011 10:19 am

jimknopf wrote: Ron, concerning the reverb of the TI: it sounds well, but is pretty basic. Don't you think the Kronos effect section offers much better reverbs? For my ears, the Lyle Mays Piano on board of the Kronos for example showcases the best reverb available on synths and workstations so far.

It all comes down to taste. I think the reverb on the virus would be better on a virus. It is how I see it.