What kind of Pro Audio Software are you most interested in?

[jens]

I agree with you xamido, when you say there's an overload of 'Virtual Analogue' softwares in the pro audio market, but I think it is due to the HUGE attention by the users for this kind of efforts (analogue electronic emulation). It seems to be something with a continuously growing demand by users...

Yes, obviously that is indeed the case - just look at the huge interest people have in e.g. U-HE's Diva - and there also Oxium, Retrologue, Charlatan, MauSynth,etc. - and when something with a new/unusual type of synthesis comes along it rarely get's a lot of attention... so those who say 'not yet another VA please' are clearly in the minority.

[schnapsglas]

Yes, obviously that is indeed the case - just look at the huge interest people have in e.g. U-HE's Diva - and there also Oxium, Retrologue, Charlatan, MauSynth,etc. - and when something with a new/unusual type of synthesis comes along it rarely get's a lot of attention... so those who say 'not yet another VA please' are clearly in the minority.

Apparently so. Surprised to learn this. Could it be because we are on KVR where we feel compelled to look at every plugin that comes out?

[xhunaudio]

Attention for Analogue Modeling is often deeply linked to the musical genres most of the people like/make/perform, not only for 'pure technology interest'...

By the way - so far - the poll results are very interesting ( specially from my point of view :) )

[jens]

Don't let yourself be deceived though - physcial modelling instruments are usually bound to more or less flop... people might want them in theory, but then they don't know what to do with them... which is just 'natural'. The Ondes Martenot and the Trautonium do not sound so organic because of their type of synthesis but rather because of the way in which they are controlled.

[xhunaudio]

Don't let yourself be deceived though - physcial modelling instruments are usually bound to more or less flop... people might want them in theory, but then they don't know what to do with them... which is just 'natural'. The Ondes Martenot and the Trautonium do not sound so organic because of their type of synthesis but rather because of the way in which they are controlled.

Hi jens,

I don't agree, once informed about its exitence, people know very well what to do with Physical Modeling. This technology has a strategic importance in the Electronic music scenario. I think people just need a few time to start considering this relatively new approach as the *standard* for their every-day-productions. Times are now mature for the technology transition from 'Sampling' to 'Modeling', and my opinion is that it is sufficient to shake our consuetude 'cages' to move on and start with something revolutionary.

[jens]

It's not enough to know about its existence... you need to find a way to play it properly and few people own a breath controller and/or a ribbon-controller...
- hell, many don't even have/use aftertouch - and even with the extensive RT controls it offers, it's nigh impossible for me to play some of my Z1's PM programs without making it sound like a Bontempi. Needless to say, it was a flop - and so were the Alesis Fusion and the Technics WSA-1 ... and as far as I know, the development of Modelonia 2 is currently frozen due to the lack of interest.

I don't want to discourage you and if you have a great idea, then please please go ahead... but don't be disappointed when it doesn't become a huge success.


on the other hand:

I haver a purely theoretical idea for PM that would have a totally different approach than everything that is available up to now... and if you managed to come up with something like this I bet it would sell like hot cakes.

[acousticglue]

Something like Monolith that is for guitar players and is a synthesizer using the 1/4" plug. No delay and has many add-on options.

[xhunaudio]

It's not enough to know about its existence... you need to find a way to play it properly and few people own a breath controller and/or a ribbon-controller...
- hell, many don't even have/use aftertouch - and even with the extensive RT controls it offers, it's nigh impossible for me to play some of my Z1's PM programs without making it sound like a Bontempi. Needless to say, it was a flop - and so were the Alesis Fusion and the Technics WSA-1 ... and as far as I know, the development of Modelonia 2 is currently frozen due to the lack of interest.

The fact is that in the passed years there wasn't enough processing power to make ACCURATE real-time simulations of SPECIFIC instruments. So, what came out in the 90s for example (but also later) were generic Physical Modeling softwares that simulated 'everything' and at the same time 'nothing'. Personally I think a good approach to increase phy.mod. diffusion is to make accurate simulations of real/specific instruments USING phy.mod. technology. And not to sell a generic phy.mod. TOOLS/technology to the end-user asking him to make his own instruments. This slowed phy.mod. diffusion. Because in real life each instrument is really and deeply different from all the others, and expecting a faithful reproductions of an existing gear (wich makes it really usable in productions) using a generic phy.mod. tool/approach is simply something impossible. For example one aspect of IronAxe is the fact it manages materials, non-linear damping and time-variant tension energy dispersion for each single string. Something really too complex to be achived with/using any generic (also if good) phy.mod. 'monolithic' tool.

[David Carpenter Wind Core]

Physical modeling is the most expressive and sampling is the most realistic. True Pianos atlantis module is a good representation of what happens when you combine Physical modeling with sampling. Also sample modeling instruments to some degree uses this more on the sampling side then the physical modeling though.

I think using the same modeling technology that yamaha has been using and combining that with sampling can yield realistic emulations with a more expressive playability. A benefit is less memory and cpu footprint.

ON the other hand I find that Diva is a great representation of how analog emulation is evolving. It sounds so organic and like a real analog synth it blows my mind. It also blows away my cpu and sample buffers.

I think refinements are happening in every aspect of synthesis sampling, physical modeling, additive, and analog.

I think also that the hardware controllers are a bit behind the software. New innovative instruments that are quality are rare.

To me it takes all aspects the software the hardware and the user to complete the circle of performance. Having proper and intuitive ways of controlling and playing these instruments is just as important as the synth it self. For emulation based performances the controller must suit the instrument. Anything that is a wind instrument should be controlled by the users breath. Anything that is a percussion instrument should be played on a electronic drum kit etc.

When the strengths of the synthesizers are complemented by the control method then a more realistic emulation can be achieved. This allows the musician more realism more control and a broader musical palette to work with.

Looking at all aspects of what is involved to perform an accurate emulation is a good start. By identifying the main key prerequisites and incorporating the appropriate technology a better emulation or even an all new instrument sound can be realized.

No emulation can beat a real instrument but no real instrument can be anything but itself. It is important to remember what we use emulations for. To emulate the original instrument and be able to play said instrument and any other with a midi controller of choice. This allows the musician to incorporate instruments that he or she doesn't know how to play quickly and as seamlessly as playing the instrument they do know. This can overcome the drawbacks of realism as long as the performance is musical. The better the emulation and the more performance friendly it is the easier it is to trick the listener into thinking that is the original instrument.

I don't know how many songs used the pan flute patch over the years a lot but how many people stopped to say hey that's not a real pan flute. It is because people became so familiar with the synth pan flute that the real thing didn't matter because of the newly found sound association.

I think the future of synths are interconnected it takes all the different aspects to be further developed in order to evolve.

[jens]

gh processing power to make ACCURATE real-time simulations of SPECIFIC instruments. So, what came out in the 90s for example (but also later) were generic Physical Modeling softwares that simulated 'everything' and at the same time 'nothing'.

No,actually that is not the case. Z1 has several very specific algorithms. And with the other PM sythesizers it's very similar, I think (only the Z1 has several different algorithms though). Just the WSA-1 had a more generic approach,
I think.

what I have in mind would be very generic, and at the same time could be very realistic, I assume.

Additive synthesis is in theory very capable (at least the advanced one as in the K5K - but with more partials than the Kawais). The problem however is that it is almost impossible to adjust all the parameters by hand, so an algorithm would instead need to do it. And to be realistic, it would need to get fed and to analyze a good sampleset - i.e. different pitches, different velocities, different playing techniques. Then it could predict how the instrument behaves under the different conditions and set the parameters accordingly. What you would then get is an actual recreation of the instrument. And with this it would also be possible to create all kinds of new instruments by purposely feeding it with the wrong samples.

[stillshaded]

I think the future of synths are interconnected it takes all the different aspects to be further developed in order to evolve.

Great post, and this last line sums it up nicely. I'll expand on this by saying, that my preference is when lots of different synthesis methods are cleverly integrated together.

People ITT are saying that people just want VA's. Yes there's a lot of that, but what's the most popular synth you can think of? maybe Mm m m m m MASSIVE? Definitely not a pure VA. But it integrates some novel(ish) features, with old as dirt ones, and with a great interface. IMO this is what devs should be aiming for.

Another synth that I think does this well is the Virus TI, I'm definitely not one of these folks who thinks software can't match it, but I do think it's a great synth. They spent years perfecting a solid VA, and then they added some wicked new oscillators! Seriously the wavetables, granular, and formant oscillators in that thing are SO fantastic. I get sick of people only talking about it's supersaw.. SHEESH gimme a break. That'd be like if you were like MAN I NEED A MOOG VOYAGER FOR IT"S WHITE NOIZE!!11!!

so yea.. stand on the shoulders of giants, as it were.

[Jace-BeOS]

I would love to see a new "king" of physical modeling. The problem is that it's more the realm of expert physics and maths not just for the software design but for the programming of sounds. It has not yet been offered to users in both a rich and accessible format. Existing tools are archaic and existing common controllers don't allow for accessibility to expressive playing. Imagine if it were standard operating procedure to have a full length, pressure sensitive ribbon controller on all master keyboards instead of a few Kurzweils. Or a bowing device. Imagine if Yamaha still sold breath controllers and they weren't retardedly expensive.

The day that it requires mere imagination rather than a degree in math and physics to creatively program a physically modeled synth is the day it becomes as ubiquitous as analog emulations. You can get fun sounds and even usable sounds out of analog synth emulations by twiddling dials and hitting randomizer buttons. If you try that method with physical modeling, you get horrible noises or, more likely, nothing at all (& sometimes program crashes).

[xhunaudio]

gh processing power to make ACCURATE real-time simulations of SPECIFIC instruments. So, what came out in the 90s for example (but also later) were generic Physical Modeling softwares that simulated 'everything' and at the same time 'nothing'.

No,actually that is not the case. Z1 has several very specific algorithms. And with the other PM sythesizers it's very similar, I think (only the Z1 has several different algorithms though). Just the WSA-1 had a more generic approach,
I think.

what I have in mind would be very generic, and at the same time could be very realistic, I assume.

Additive synthesis is in theory very capable (at least the advanced one as in the K5K - but with more partials than the Kawais). The problem however is that it is almost impossible to adjust all the parameters by hand, so an algorithm would instead need to do it. And to be realistic, it would need to get fed and to analyze a good sampleset - i.e. different pitches, different velocities, different playing techniques. Then it could predict how the instrument behaves under the different conditions and set the parameters accordingly. What you would then get is an actual recreation of the instrument. And with this it would also be possible to create all kinds of new instruments by purposely feeding it with the wrong samples.

Z1 uses the MOSS system, 13 different algorithms including physical modeling 'architectures', covering different real world situations (plucked strings, bowed strings, brass models, woodwind models,...). MOSS is a very powerful system (I studied/owned it for years long, long time ago :)), but its phy.mod. algorithms are the result of preliminary (=not yet complete) researches (Stanford/Yamaha) of 20 (!) years ago re-adapted to run on real time on 6 (if I remember well) low computing power DSP chips. Those algorythms let you reach good results, but defenitly not PARITY with real instruments. For example they have no coupling, excitators are very primitive... Until you are deeply involved in developments of that , you can't imagine how much complex phy.mod. is eheh... :)

About the additive approach, it has nothing to do with phy.mod., they are Spectral Modeling synthesis systems. With this approach you cannot have DIRECT control on the physics of the sound genenration process, so at this point it is better to use directly Samples. Even if you reach to make an algorythm that predicts the behaviour of about 600 partials (maybe we need a new Einstein :) for that) for having a single piano note for example, that algorythm is 'freezed' and cannot interact with the environment also because its upper harmonics are 'fake' and not direct depending and 'real-time generated' from the excitation (like in phy.mod.).

By the way thank you for reminding me the old good MOSS system. A lot of touching memories are floating around me... good old days...:)

[xhunaudio]

I would love to see a new "king" of physical modeling. The problem is that it's more the realm of expert physics and maths not just for the software design but for the programming of sounds. It has not yet been offered to users in both a rich and accessible format. Existing tools are archaic and existing common controllers don't allow for accessibility to expressive playing. Imagine if it were standard operating procedure to have a full length, pressure sensitive ribbon controller on all master keyboards instead of a few Kurzweils. Or a bowing device. Imagine if Yamaha still sold breath controllers and they weren't retardedly expensive.

The day that it requires mere imagination rather than a degree in math and physics to creatively program a physically modeled synth is the day it becomes as ubiquitous as analog emulations. You can get fun sounds and even usable sounds out of analog synth emulations by twiddling dials and hitting randomizer buttons. If you try that method with physical modeling, you get horrible noises or, more likely, nothing at all (& sometimes program crashes).

Hi Jace-BeOS, I agree with you, we need a greater diffusion for hardware (and also unconventional) controllers, but I have to say that in the last years something has changed and now availability for this kind of things growed up. In addition to Yamaha and Akai wind controllers, and Terratec/Axon series, now we can have at an affordable price things like YouRockGuitar MIDI controller that is very effective IMO. With a few adjustment (in its software via control panel) and a good electric/guitar simulation (:)) really *great* result con be reached. By the way I would see other controllers (eg. bowed string). And don't forget that another kind of controller technology is just appeared (touch-surfaces and Kinect-style controllers) opening new scenarios for the future...

I agree also about the fact physical-modeled-build products have to be ease-of-use to the END User. It is a lot more effective to interact with real counterparts than physical/ideals counterparts. It's something I always believed in. For example in IronAxe there's a parameter called 'String Material'. Only 1 knob. The equivalent of changing a string in your guitar. By the way, turning that knob, a lot of paramaters in the physical modeling simulations are affected/changed. It changes string vibration, string tension, tension->velocity, non-linear damping,... And all is linked to that SINGLE knob, and this let the end User to remein in the REAL world, not in the Physician's. It is why I agree with you. The idea basically is I'm selling a *real* product using a complex technology, not the complex technology itself. I'm selling a Guitar, not Comb Filters, Envelopes, FIRs...