KVR Audio

timandmonica · Post by **timandmonica** » Mon Apr 04, 2016 7:16 pm

I like how he says in the OP "Oh, and if you happen to have a real Elka Synthex..."

Hahahaha!

PAK · Post by **PAK** » Mon Apr 04, 2016 7:30 pm

moscom_electronics wrote:Please play with OSC B -> Cutoff! Then change OSC B settings and play with the algos. The differences then are obvious!

They are indeed. But then carefully tweak the modulation and frequency rate. Sometimes it's the main thing the "model" appears to be adjusting..

After playing a bit more, it seems option 1 does actually sound fairly bad at certain settings too..

Overall there doesn't appear to be any major level of detail modeling differences between them, beyond some noise and the amusing chirpings from model one sometimes. Non-zdf?

Can I vote for modelling the Retro Sound dude's Pro One instead?

Vesius · Post by **Vesius** » Mon Apr 04, 2016 7:30 pm

Absolutely, I wouldn't expect a pure sine - in this case the give away is that the filter goes up and up, and then goes quiet, and then starts to come back down in pitch - it's definitely 100% absolutely classic aliasing - it isn't part of an accurate emulation, unless we are saying that the Pro1 aliases?

Anyhow, I know this is a filter test, and for all I know the sync emu isn't finished, but I'm also gonna give a little feedback on osc sync. Firstly, it sounds better than pretty much any other osc sync I know, but secondly - and given the state of osc sync by and large out there, this could almost be a compliment - it's a bit demure. I don't know how other people feel, but with hard sync, I want it to be rude ... you know full on sticking it's hand down my trousers and having a rummage rude.

Which is nothing I ever thought I'd write on a KVR forum, so there you go.

akira2 · Post by **akira2** » Mon Apr 04, 2016 7:41 pm

Urs wrote:
Boardwalk wrote:https://m.youtube.com/watch?v=GgSCZTckcqk
Woah... this one sounds completely different from ours!

I wish the repro1 sounds like this !

Huge sound

About the filters... when the cutoff is to the max and then we turn the reso...the volume decrease a lot (too much in my opinion)...is it normal ?

besides... I prefer the 1 which add some character to the sound

EvilDragon · Post by **EvilDragon** » Mon Apr 04, 2016 7:44 pm

akira2 wrote:About the filters... when the cutoff is to the max and then we turn the reso...the volume decrease a lot (too much in my opinion)...is it normal ?

Yes it's normal with some common filter designs.

moscom_electronics · Post by **moscom_electronics** » Mon Apr 04, 2016 7:53 pm

Urs wrote: Pssst, no worries, we do not feel obliged to use the poll result for the final version.

I know, I know! I just overreact

u-he-william · Post by **u-he-william** » Mon Apr 04, 2016 7:55 pm

chk071 wrote:I'm a bit confused though. I thought audio signals alias, not the filter.

To clear things up about aliasing. Since a digital signal has to be sampled and is represented by a limited number of samples per second (44100, 88200 ,...), you can only "display" frequencies up to half the sampling rate. Every "operation" that in some way adds frequencies to a signal, such as oscillators, certain filters,... may create aliasing. So an audio signal does not really alias on it's own.

Urs · Post by **Urs** » Mon Apr 04, 2016 8:00 pm

akira2 wrote:
Urs wrote:
Boardwalk wrote:https://m.youtube.com/watch?v=GgSCZTckcqk
Woah... this one sounds completely different from ours!
I wish the repro1 sounds like this !
Huge sound

About the filters... when the cutoff is to the max and then we turn the reso...the volume decrease a lot (too much in my opinion)...is it normal ?

besides... I prefer the 1 which add some character to the sound

I hope you realize that the Retrosound demo is processed... I love his demos, but adding delay whatsoever does not draw an objective picture! I'm fairly sure RePro will sound as strong with a similar effect chain behind it.

exmatproton · Post by **exmatproton** » Mon Apr 04, 2016 8:04 pm

Urs wrote:
akira2 wrote:
Urs wrote:
Boardwalk wrote:https://m.youtube.com/watch?v=GgSCZTckcqk
Woah... this one sounds completely different from ours!
I wish the repro1 sounds like this !
Huge sound

About the filters... when the cutoff is to the max and then we turn the reso...the volume decrease a lot (too much in my opinion)...is it normal ?

besides... I prefer the 1 which add some character to the sound
I hope you realize that the Retrosound demo is processed... I love his demos, but adding delay whatsoever does not draw an objective picture! I'm fairly sure RePro will sound as strong with a similar effect chain behind it.

Little processed demo..6 repro 1's, bit of flange here and there, some verbs...

https://soundcloud.com/exmatproton/u-he-repro-test-2

.jon · Post by **.jon** » Mon Apr 04, 2016 8:05 pm

.jon wrote:It lacks bottom end authority of an analog, otherwise filters 2-5 all sounds really very nice and even 1 would be just fine.

I retract this argument after comparing these to analog filters, Repro does have that thickness in there right before the filter closes.

Lesson learned, memory isn't a substitute for A/B.

Funkybot's Evil Twin · Mon Apr 04, 2016 8:22 pm

With crossmod and high resonance, I'm really digging model 2. Filter 1 completely turns to mush with that combo and low cutoffs whereas 2 holds together nicely. 5 does well here too. 3 and 4 get noisy.

chk071 · Post by **chk071** » Mon Apr 04, 2016 8:24 pm

u-he-william wrote:
chk071 wrote:I'm a bit confused though. I thought audio signals alias, not the filter.
To clear things up about aliasing. Since a digital signal has to be sampled and is represented by a limited number of samples per second (44100, 88200 ,...), you can only "display" frequencies up to half the sampling rate. Every "operation" that in some way adds frequencies to a signal, such as oscillators, certain filters,... may create aliasing. So an audio signal does not really alias on it's own.

Thanks for explaining.

Numanoid · Post by **Numanoid** » Mon Apr 04, 2016 8:27 pm

u-he-william wrote:To clear things up

To quote Faith No More: Introduce yourself

How many programmers are involved with making this synth?

moscom_electronics · Post by **moscom_electronics** » Mon Apr 04, 2016 8:29 pm

Test Repro.zip

moscom_electronics wrote:I think that 3 is winning for me actually. Here are two sound examples why... The first example is a sound I made played at C3. The sound is played 5 times, each time with different filter from 1 to 5 in chronological order.
https://soundcloud.com/moscom/c3-test

the second one is the same patch played two octaves higher (C5)
https://soundcloud.com/moscom/c5-test

On these examples, the following things are pretty clear :
- 1 is digital and noisy as hell when modulation is extreme
- 2 and 5 just weaken the sound and make almost inaudible
- 4 is good, but has some weird artefacts
- 3 is clearly the winner, musical all along.

This is why my vote goes to 3.

Also, I am starting to think that it'd be fun that we start posting audio examples of sounds that we find compelling and nice, eventually comparing the filters on these sounds to explain what we find nice. What do you think?

and if it is of any interest, here is the patch I used to generate these sounds...

Urs · Post by **Urs** » Mon Apr 04, 2016 8:41 pm

Numanoid wrote:
u-he-william wrote:To clear things up
To quote Faith No More: Introduce yourself

How many programmers are involved with making this synth?

We're 5 developers. William has done most of the oscillator+envelope code while the others have kept my back free for filter design. The project is also used as a blueprint for modeling other such synths.

KVR Audio

Repro-1 (out now)

To your ears, which filter behaves most analogue