For those in the know: Is this morphing?

[tulamide]

Guys, please listen to this video: https://www.youtube.com/watch?v=Pd-uSis8xSc

It is about a synth called Nostromo. Ignore that it is a Rack Extension for Reason, I just want to know if what they present in the video is really morphing, or more like blending?

If it is morphing, what does Flowstone lack to not be able to create something like that?

[Spogg]

Guys, please listen to this video: https://www.youtube.com/watch?v=Pd-uSis8xSc

It is about a synth called Nostromo. Ignore that it is a Rack Extension for Reason, I just want to know if what they present in the video is really morphing, or more like blending?

If it is morphing, what does Flowstone lack to not be able to create something like that?

I listened very carefully to the initial part of the demo, where stuff was simple, and this is cross-fading or blending. I made use of the principle in my Quilcom Blender project.

The nomenclature is a tricky one but in my view morphing is different. The comparison with face morphing in Photoshop may be useful. We've all seen one face morph into another. In a good morph all parts of the face change position, intensity and colour progressively but individually. This is not the same as having a start face and an end face and fading from one to the other. In this case the different intermediate faces are never seen and the effect is not the same.

With sound it's a bit more difficult. If you imagine a trumpet morphing into a flute you should hear all the stages in between. This would mean addressing each partial or harmonic number individually and sliding from the start to the end on a per-partial basis both in amplitude and phase. The middle-stage trump-flute should have some characteristics of both but be different from either.

For simple static sounds however, blending may be "sufficient" and it's certainly a powerful technique, as I found by experimenting with the Blender.

The clip below compares the two approaches and the difference between blending and morphing in this is so huge:

https://www.youtube.com/watch?v=iuRxmLRacWQ

I was very impressed with the Nostromo though, I must say. If I made that in Flowstone it would need a super-computer to run it The coding must be remarkably efficient. However I did notice a glitch when a wave was inserted but I guess that's not intended to be done live.

Thanks for sharing this tulamide; fascinating!

Cheers

Spogg

[tulamide]

I was very impressed with the Nostromo though, I must say. If I made that in Flowstone it would need a super-computer to run it The coding must be remarkably efficient. However I did notice a glitch when a wave was inserted but I guess that's not intended to be done live.

That's the really unbelievable good work of Propellerheads audio engine. Contrary to VSTs, a Rack Extension makes full use of all the inner workings of Reason - no need to re-invent the wheel. And how efficient it is? Well, I will blow your mind away: Having done that in Flowstone would have needed a super-computer?

Well, I found this song. The song is lame and boring, I just chose it because of the description. These are 12 Nortromos playing, with full effect chain on each of them - and that wasn't even pushing the cpu load to a third.
https://soundcloud.com/wongothesane/lysurgic
As I mentioned before, that's what I love about Reason. I can literally use several hundred synths and effects at the same time, and the (almost ten year old) processor in my pc doesn't even start to sweat. In fact, in the last 5 years I wasn't able to ever reach the limit of Reason. As if Propellerhead would be able to get 10 times more out of each processor cycle than any other audio programmer in the world.

Of course, it's still a lot of work to build something like Nostromo (Propellerhead gives you the library and a framework, the developing is still your part). But you can build it while being assured you can realize it in a way that let's the user have fun with the instrument.

And thank you very much for your description. According to the way I asked one can see that I also assumed blending. I was confused because the video uses the word "morph".

[adamszabo]

To me it sounds more like simple crossfading. And I already had an idea like that to make, but Im very busy with other stuff, but it shouldn't use that much Cpu. You can make something similar with assembly and use dynamic bypassing. There are 9 wavetables in that video, so you would need 9 oscillators, but when you crossfade between two oscillators you only need 2 active. The rest can be bypassed. When the lfo moves the position it enables the next active wavetable, and disables the last one. Or thats how I would do it anyway

[tulamide]

I thought of something like that as well. But, I read in a Reason-Forum where the creator of the synth anounced it, that the display is not a waveform (or nine waveforms), but that the images show the harmonics. So one of them in the video only has even harmonics (I can see it), while others have a full spectrum, etc. Now this would lead me to think that the harmonics are blended, or aren't they?

[adamszabo]

Well the harmonics are just a different representation of the sound. Internally it could simply blend the oscillators.

[Spogg]

To me it sounds more like simple crossfading. And I already had an idea like that to make, but Im very busy with other stuff, but it shouldn't use that much Cpu. You can make something similar with assembly and use dynamic bypassing. There are 9 wavetables in that video, so you would need 9 oscillators, but when you crossfade between two oscillators you only need 2 active. The rest can be bypassed. When the lfo moves the position it enables the next active wavetable, and disables the last one. Or thats how I would do it anyway

In my thought experiment I see three oscillators being active at any time; the current one, the previous one being faded out and the next one being faded in. Of course it would depend on the cross-fade law applied to the blending process. Since these sound like small static wave tables my guess is that cross-fading is "sufficient" and true morphing would be too taxing and not add much.
When a new wave was added there was a small glitch and this suggests to me that a wave array or similar is being updated, so this may be a clue as to the method used.

@tulamide: you've mentioned before that Reason is so very CPU efficient. Would it, in theory, be even possible for a VST to achieve this or is there an in-built limitation cuased by the VST SDK?

One thing I would like for Flowstone 4.0 is for the many Prims to be more optimised, like so many made by Martin Vicanek. He has proved admirably well that it's possible to create very CPU efficient alternatives. It would be really nice if they were available out of the box. When I came to FS I assumed the Prims were all you got and knew nothing of optimisation so I think this would be great for newbies.

Cheers

Spogg

[adamszabo]

In my thought experiment I see three oscillators being active at any time; the current one, the previous one being faded out and the next one being faded in. Of course it would depend on the cross-fade law applied to the blending process.

Hmm, 3 doesnt make any sense to me. Look at this simple example, you can see that when you turn the morph knob there are always a maximum of two waves that are being calculated, and it morphs nicely. The ones with '0' multiplication can be bypassed and they wont use any cpu.

[tulamide]

@tulamide: you've mentioned before that Reason is so very CPU efficient. Would it, in theory, be even possible for a VST to achieve this or is there an in-built limitation cuased by the VST SDK?

There certainly is no in-built limitation by the VST SDK. VST, introduced by Steinberg, is also a good starting point to mention that the founders of Propellerhead Software worked for Steinberg before founding their own company. On of them wrote the manual for Cubase, another programmed the Master Track Editor, and even the presentation of their first production software, Rebirth-338 (1997! Cubase just started that year with VST 1, which only allowed effects and eqs), was done on a Steinberg event. What I mean is, those guys really knew what they were doing. That's why Reason is so remarkable in terms of efficiency.

Comparing RE and VST is not so easy. Imagine being stranded on an island and having to build a house. In the VST version you have to first walk around and look for materials you can use. In the RE version you find wooden planks, nails and a hammer all beaded in the sand.

RE is just way more integrated into the DAW (if we call Reason a DAW for this comparison) than VST, which is an open standard and thus has to work for all DAWs and not just Cubase.

Having said that I can imagine a software synth that wouldn't be doable for VST but for RE. But I don't think it's Nostromo. I can also imagine synths that would be doable in VST, but not in Flowstone

[Spogg]

In my thought experiment I see three oscillators being active at any time; the current one, the previous one being faded out and the next one being faded in. Of course it would depend on the cross-fade law applied to the blending process.

Hmm, 3 doesnt make any sense to me. Look at this simple example, you can see that when you turn the morph knob there are always a maximum of two waves that are being calculated, and it morphs nicely. The ones with '0' multiplication can be bypassed and they wont use any cpu.

Adam, I stand corrected. I only did a thought experiment but you went to the trouble of making a schematic to demonstrate your point. For this I am both amazed and very grateful.

In my imagination I could see a blend rule whereby when the LFO was pointing at the centre of a wave there could be some influence from the waves either side. Of course this would mean you would never hear just the wave pointed to at any time. However, it might be interesting to hear this "blurring effect" so the contrast as the waves cycled would be reduced. I'm very tempted to experiment to try the effect of blurring...

Cheers

Spogg