Synth effects for kpa

  • Could you implement some synth effects like piano, trumpet, violin or so on?

    Sounds such as these aren't ever likely 'cause they'd require samples (not going to happen 'cause of RAM demands) or sophisticated modelling (quite CPU-hungry), Frank.


    I wouldn't rule out conventional-synth sounds, 'though, 'cause they'd generally only require no more than a few oscillators of sine, square and sawtooth nature's be generated (much-lower CPU load than the former instruments' raw sources) and then filtered. These would generally be used for pads, string sounds and mono-synth leads.

  • That's exactly what I was saying should be doable in the Kemper, deadpan.


    A simple sine / square / saw-oscillator-driven affair shouldn't be too much of a headache to implement, nor should it tax the RAM or CPU overhead much at all IMHO.


    Modelling involving complex waveform generation, such as would be required for semi-authentic-sounding acoustic instruments such as the ones Frank mentioned, would be an entirely-different affair 'though, obviously.


    Cheers mate.

  • I thought you said above that would never happen due to needing samples and high cpu usage, hence my comment.

    Ha! I see mate. Please forgive me if you know all this stuff, DP; my "answer's" in there, but I figured the basic and relevant grounding was not only relevant but could be helpful to those without a synthesis background in understanding why we're not likely to ever see a "serious" synth in our beloved beasties.


    In the first instance, I was talking about authentic-sounding sounds (I think that's your answer, DP):


    Sounds such as these aren't ever likely 'cause they'd require samples (not going to happen 'cause of RAM demands) or sophisticated modelling (quite CPU-hungry), Frank.

    I then made the case for a traditional subtractive synth:


    I wouldn't rule out conventional-synth sounds, 'though, 'cause they'd generally only require no more than a few oscillators of sine, square and sawtooth nature's be generated (much-lower CPU load than the former instruments' raw sources) and then filtered.

    Subtractive synthesis that isn't sample-based (where the samples are of real-world instruments; I'm not talking about SQR / SINE / SAW oscillator samples used to create pseudo-analogue results either), such as that we could expect of a Kemper synth, couldn't possibly hope to mimic guitar, piano, flute, violin (to use the examples Frank provided) or any other harmonically-complex instrument accurately.


    Emulations in the analogue world are obviously what we all had to make do with until the Synclavier, Fairlight and then FM synthesis from Yamaha came along, followed by the ROMpler revolution and finally uber-accurate samplers and their attendant libraries, and boy were said approximations lousy. Thankfully, playing the various instruments idiomatically went a very-long way towards "authenticating" the "emulations". This is what we'd have to do with a potential Kemper synth, assuming we ever saw one and that it was traditionally-oscillator-based. For anything meaningful to emanate from our KPAs, emulation-wise, we'd have to master idiomatic playing for each instrument we intend to mimic. Dedicated synths controlled via hex pickups or whatever, such as Roland has offered, have this area covered, especially in terms of realism.


    Seeing as this idiomatic implementation, which takes a lot of effort and practice, is doomed to failure authenticity-wise, especially in this age of multi-gigabyte sample libraries, I saw it fitting to suggest that:


    These would generally be used for pads, string sounds and mono-synth leads.

    Classic guitar-synth stalwarts. The strings won't sound authentic, but used as support, background sounds such as pads, they'd work quite well. Roland's VG system and its ilk don't qualify for this category as the DSP is dedicated for the task of carrying out sophisticated synth as well as real-world-instrument modelling, and this isn't the Kemper's brief, nor has it been designed with this in mind... I assume.


    Hope this helps someone and that I have it more-or-less "right".

  • I don't think that we need a full fledged synth in the KPA as a guitar driven synth will always be a compromise no matter how good it is. I have a Boss SY-300 and I really like it but it will never replace a real (virtual) analog synth or a sampler.


    However, it would be great to have some basic synth building blocks in the KPA like assignable LFOs, envelopes and things like that. The KPA already has a lot of pitch and filter capabilities so this would allow us to create a lot of the usual synth like guitar sounds directly in the device (strings pads, wobbles, filter sweeps, arpeggios and so on).


    A slow gear effect like the ones found in even the cheapest effects units out there would be a good start. It should be very easy to implement as this is just a compressor setting after all. And if you're feeling really sophisticated then you'll add a gate parameter as well for reversed cymbal like effects.


    Add some more simple waveforms to the tremolo, filter and vibrato blocks (saw, square, S&H). This shouldn't be too difficult for a synth manufacturer to create and will add great VCA, VCF and VCO modulation capabilities to the KPA (especially if the parameters are morphable).


    Add a voice delay parameter to the pitch shift block and you'll have simple arpeggios. Very basic but covers a lot of ground.


    Along with the upcoming delay and reverb upgrades (reverse, shimmer etc) this should give us a lot of stuff to play with and it shouldn't be too difficult to implement.