digital synthesizer that is most analogy?

[killmaster]

I'm not 100% sure how to ask this since its a little technical, but I'm look to find a synth that creates sounds with the most samples. I don't really care about CPU load, just want to get the most high quality analog resolution. My intention is to record and massively stretch the sound so wanting as many sample/subsamples as possible. When I stretch, I want the sound to interpolate as little as possible.

I do have Repro and was thinking maybe that would do the trick. But I really know nothing about how digital synths actually work in this regard.

thanks for any recommendations,

Greg K.

[Yorrrrrr]

I don't even know where to begin...but most digital VSTi synthesizers can output sound to pretty much whatever sample rate you desire, from 44.1 up to 192 kHz (some may sound different at higher rates). That said, I guess you already know that even 44.1 kHz (the standard) is sufficient when it comes to sample rate, to digitally represent any analog sound from 20 Hz to 20 kHz (the range that our human hears can perceive) with absolute accuracy (see sampling theory and nyquist frequency).

[yung-forever]

Just about any good synth should render at the host sample rate. Bump up your DAW to the highest sample rate you can, render the audio, and then stretch.

[killmaster]

I don't even know where to begin...but most digital VSTi synthesizers can output sound to pretty much whatever sample rate you desire, from 44.1 up to 192 kHz (some may sound different at higher rates). That said, I guess you already know that even 44.1 kHz (the standard) is sufficient when it comes to sample rate, to digitally represent any analog sound from 20 Hz to 20 kHz (the range that our human hears can perceive) with absolute accuracy (see sampling theory and nyquist frequency).

Thanks and no, I did not know this. But, wanting to stretch it out and preserve the quality without smearing I thought might be another consideration...

[killmaster]

Just about any good synth should render at the host sample rate. Bump up your DAW to the highest sample rate you can, render the audio, and then stretch.

thanks! makes sense.

[Chris-S]

I do have Repro and was thinking maybe that would do the trick.

Yes, Repro-1/5 is very good in the emulation of analog sound.
They use oversampling internally when needed.

[Guenon]

But, wanting to stretch it out and preserve the quality without smearing I thought might be another consideration...

If you want to stretch it in the sense of timestretching, i.e. preserving the original pitch, the process will need to make up the needed extra material in any case. Think about it, using a simple waveform as an example. Using a higher sample rate, you just have a higher amount of samples taken along that wave, together representing the oscillations in time -- but there are still just as many oscillations in any given time period. If you want to timestretch it (preserving pitch), you can't just pull it longer like a rubber band or a spring , thinking that the higher number of samples somehow makes that process finer, indeed when "elongating" the wave like a spring, for example. If you "stretch" it like that, you are actually just... slowing it down. You have the exact same waveform, with just as many oscillations, happening during a longer time period --> lower pitch, slower sound.

If you want to stretch it in the sense that the oscillations still happen just as fast as in the original audio (in other words, you hear the same pitch, while also the resulting stretched audio is slower as a whole, as its overall duration increases), you will need an algorithm that creates that extra material -- the extra oscillations you need in order to cover the longer time period. It's not inherent in the waveform itself, no matter how high you boost the sampling rate.

[killmaster]

But, wanting to stretch it out and preserve the quality without smearing I thought might be another consideration...

If you want to stretch it in the sense of timestretching, i.e. preserving the original pitch, the process will need to make up the needed extra material in any case. Think about it, using a simple waveform as an example. Using a higher sample rate, you just have a higher amount of samples taken along that wave, together representing the oscillations in time -- but there are still just as many oscillations in any given time period. If you want to timestretch it (preserving pitch), you can't just pull it longer like a rubber band or a spring , thinking that the higher number of samples somehow makes that process finer, indeed when "elongating" the wave like a spring, for example. If you "stretch" it like that, you are actually just... slowing it down. You have the exact same waveform, with just as many oscillations, happening during a longer time period --> lower pitch, slower sound.

If you want to stretch it in the sense that the oscillations still happen just as fast as in the original audio (in other words, you hear the same pitch, while also the resulting stretched audio is slower as a whole, as its overall duration increases), you will need an algorithm that creates that extra material -- the extra oscillations you need in order to cover the longer time period. It's not inherent in the waveform itself, no matter how high you boost the sampling rate.

o wow, had no idea that it worked this way! Thank you! I still have to read what you wrote several time to grasp this.

So, in my case, I'm stretching stuff in Bitwig, keeping the same pitch, and getting steppy results. But next month, they are adding Elastique Pro and I'm wondering from what you are saying if Elastique will do a better job filling in the gaps smoothly.

[egbert101]

<delete>

[VariKusBrainZ]

Elastique Pro will do a much better job but theres no algo that will perform a stretch without changing the sound of the sample in some form.

Why do you want to stretch a synth sound AND keep its original quality and timbre when you can just keep the note feeding the synth patch sustained for as long as you need and suffer none of the side effects of time stretching?

Or you could stick in a sampler and set up loop points

Unless you want time stretching side effects

[fluffy_little_something]

I don't get what exactly he wants to stretch. And at which stage of the workflow.
Synths often use different oversampling factors in different sections.
In another thread someone mentioned a separate oversampling plugin that oversamples the global output of a synth.

[killmaster]

But, wanting to stretch it out and preserve the quality without smearing I thought might be another consideration...

If you want to stretch it in the sense of timestretching, i.e. preserving the original pitch, the process will need to make up the needed extra material in any case. Think about it, using a simple waveform as an example. Using a higher sample rate, you just have a higher amount of samples taken along that wave, together representing the oscillations in time -- but there are still just as many oscillations in any given time period. If you want to timestretch it (preserving pitch), you can't just pull it longer like a rubber band or a spring , thinking that the higher number of samples somehow makes that process finer, indeed when "elongating" the wave like a spring, for example. If you "stretch" it like that, you are actually just... slowing it down. You have the exact same waveform, with just as many oscillations, happening during a longer time period --> lower pitch, slower sound.

If you want to stretch it in the sense that the oscillations still happen just as fast as in the original audio (in other words, you hear the same pitch, while also the resulting stretched audio is slower as a whole, as its overall duration increases), you will need an algorithm that creates that extra material -- the extra oscillations you need in order to cover the longer time period. It's not inherent in the waveform itself, no matter how high you boost the sampling rate.

o wow, had no idea that it worked this way! Thank you! I still have to read what you wrote several time to grasp this.

So, in my case, I'm stretching stuff in Bitwig, keeping the same pitch, and getting steppy results. But next month, they are adding Elastique Pro and I'm wondering from what you are saying if Elastique will do a better job filling in the gaps smoothly.

Pauls Extreme Stretch is your friend.

http://hypermammut.sourceforge.net/paulstretch/

thanks! though appears to be PC only and I'm on mac.

there is this though which I've used with great results before. Just hoping to process within DAW.
Soundhack (in lower right of window). You can use the phase vocoder to ultra stretch. very cool...

[killmaster]

But, wanting to stretch it out and preserve the quality without smearing I thought might be another consideration...

If you want to stretch it in the sense of timestretching, i.e. preserving the original pitch, the process will need to make up the needed extra material in any case. Think about it, using a simple waveform as an example. Using a higher sample rate, you just have a higher amount of samples taken along that wave, together representing the oscillations in time -- but there are still just as many oscillations in any given time period. If you want to timestretch it (preserving pitch), you can't just pull it longer like a rubber band or a spring , thinking that the higher number of samples somehow makes that process finer, indeed when "elongating" the wave like a spring, for example. If you "stretch" it like that, you are actually just... slowing it down. You have the exact same waveform, with just as many oscillations, happening during a longer time period --> lower pitch, slower sound.

If you want to stretch it in the sense that the oscillations still happen just as fast as in the original audio (in other words, you hear the same pitch, while also the resulting stretched audio is slower as a whole, as its overall duration increases), you will need an algorithm that creates that extra material -- the extra oscillations you need in order to cover the longer time period. It's not inherent in the waveform itself, no matter how high you boost the sampling rate.

o wow, had no idea that it worked this way! Thank you! I still have to read what you wrote several time to grasp this.

So, in my case, I'm stretching stuff in Bitwig, keeping the same pitch, and getting steppy results. But next month, they are adding Elastique Pro and I'm wondering from what you are saying if Elastique will do a better job filling in the gaps smoothly.

Pauls Extreme Stretch is your friend.

http://hypermammut.sourceforge.net/paulstretch/

thanks! though appears to be PC only and I'm on mac.

there is this though which I've used with great results before. Just hoping to process within DAW.
Soundhack (in lower right of window). You can use the phase vocoder to ultra stretch. very cool...

[killmaster]

Elastique Pro will do a much better job but theres no algo that will perform a stretch without changing the sound of the sample in some form.

Why do you want to stretch a synth sound AND keep its original quality and timbre when you can just keep the note feeding the synth patch sustained for as long as you need and suffer none of the side effects of time stretching?

Or you could stick in a sampler and set up loop points

Unless you want time stretching side effects

I want to mangle it and take advantage of whatever artifacts some about as part of the process. Only thing I was thinking is that there are so many harmonics going on that you can't hear til you severely slow the sound down and to the degree that I can preserve these "hidden" harmonics, it could be worth the effort. I tried a demo of ircam lab thing awhile back and it was phenomenal. It seemed to do such a great job preserving the original quality. That is if I can even tell or not! Again, though, wanting to know what I can similar "quality" doing it on a track in a DAW.

[killmaster]

I don't get what exactly he wants to stretch. And at which stage of the workflow.
Synths often use different oversampling factors in different sections.
In another thread someone mentioned a separate oversampling plugin that oversamples the global output of a synth.

I want to take either recorded material, like real nature sound OR synth sounds and stretch them. I did it here with a track. The opening part that gradually pitches down. The whole track is a short segment of little blips that I stretched 20 times or so.

https://soundcloud.com/ret-sam-llik/sehnstucht

it sounds pretty good but I had to add some processing to reduce the "steppiness"