Trying to find software that can "blend" two samples together

so id like to find some software that can, blend, or fuse, two samples into a new sample. ive had conversations with people about this type of thing, and people suggest recording the two samples playing at the same time. this is not at all what i want

 

id like to blend together, say, a bird chirp, and a piece of glass shattering. the resulting noise should be all high pitched kind of like a bird chirp, but would also sound all mangled up and whatnot like glass shattering. thats even an obscure idea, it would be awesome to even mix two drum machines, like fusing two different kicks together or what have you

 

i dont really understand much about software synthesis, but i imagine this type of thing is easily done

 

if anyone could point me in the right direction that would be great

 

thanks

The closest thing I can think of is manually cross fading small elements. It would take a lot of time.

I don't understand. If you don't want them simply layered, then what do want? Can you describe conceptually what the difference would be?

what you're wanting is some kind of morphing between the two samples.

there are different techniques which give different results.

 

just blending two sounds together definitely wont do it.

 

there are a lot of things you could try to get these types of results 'manually', like using pitch shifting on each sound separately, and pitching each sound either up or down while listening to the result. you could also EQ/filter each sound. those two things together could be used to blend two similar sounds into one.

 

i like the free Braindoc pitch shifter vst, which has feedback too so that could help you kind of mutate the sounds into each other.

 

but really for the better morphs you would something more like a vocoder. it gets a bit more of a pain in the ass to set up because you'll have to route the sounds so that both are going into the vocoder, so that one is being vocoded by the other. youre going to have to tinker with parameters a lot too. mixing some of either signal's dry into the wet results, or you could even combine this with the pitch shifting/filtering (for either the dry component or the wet or both) to really fine tune the final sound you get. all vocoders are not created equal so some will give you crappy results while others might be better.

 

there are also more advanced plugins out there just for 'morphing' which, i believe are using something like vocoding but just kind of more advanced. there's this synthesizer vsti for example which i believe you can load 2 different samples into and have them morph into each other in various ways. i think there is even some pretty advanced hardware out there for doing it but i don't really know much about this stuff beyond what i've said here.

convolution reverb, but instead of a space, another sample :p

  On 9/16/2013 at 3:12 AM, logakght said:

convolution reverb, but instead of a space, another sample :p

ha hey yeah thats a good one!

a super expensive but amazing option is the program kyma, which was how they made the waaallllyyyy voice from wall.e. Some incredible morphing features!

 

 

near the end she goes over the morphing a bit.

convolution is probably the easiest way, but even with large FFT window sizes you will still get a lot of 'samey' results and they will mostly be textural not exactly a morph between two sounds (not like the sound equivalent of say morphing a picture of 2 different people's faces together)

Kyma is indeed the best thing I've used to do this, but it takes a lot of manual preparation (including drawing with the mouse the curve for the fundamental frequency of the sound). It also takes a lot of trial and error, sounds with more noise content that can't be converted into sine waves (this is how kyma works by breaking down a sound up to 128 sine wave oscillators) won't sound good or smooth when morphed into other sounds. it works best when using something with a distinct tone or note value. Trying to morph something like drums to piano for example will sound weird and glitchy.

I spent a lot of time trying to get good sounds out of it, and by the end I had about a 20% or lower success rate, and most of the successes weren't sounds I ended up using for anything they were just realistic sounding morphs. It's too bad there isnt anything out that works better than Kyma that I know of. They also haven't bothered to update the process in at least 7 years, if it was automatic and you didn't have pre-process the sound i would probably use it a lot more often, but its such a specific type of sound I can't imagine dedicating myself to spending the amount of time required to get useable results. To this day the best sound morphs I've heard are by the guy who did most of the demos you hear on the Kyma website itself, he has his own page somewhere with a bunch of other morphs that sound very very smooth.

there are other methods to get the sound of a bird chirping to sound like a breaking piece of glass. There is a max/msp patch called 'sound spotter' that's supposed to 'match' the content of two different sounds together. IT only does a basic 50/50 crossfade mix but it uses one sound's spectral content to 'trigger' another sound's spectral content so they sound as if the two sounds are interacting. I've never been able to get good results out of the patch myself but maybe someone else would have better luck.
Another similar piece of software to Soundspotter is 'loopmash' that only comes with Cubase 5 and above, It's very easy to use in theory but i haven't spent much time with it.

Edited September 16, 2013 by John Ehrlichman

Some cool ideas in here.

 

I wonder if anything like this could be done in praat. So far I can't find a praat script that does any more than just layer two sounds.

a consideration: this gets much simpler when you do it with synth patches.

I have an old Emu ESI-32 sampler that has a function that does this.

 

It's called Transform Multiplication.

 

From the manual:

 

  Quote

6. Transform Multiplication

This unique function merges two sounds together in a way that

frequencies common to both sounds are accentuated while uncommon

frequencies are discarded. Because of this characteristic, Transform

Multiplication tends to work best with sounds that are harmonically

rich. Using this function is easy, just pick two sounds and go. The length

of the resulting sample will be equal to that of the current sample.

Θ

Tip:

Transform Multiplication can take

quite a bit of time with longer samples.

Begin your experiments with short samples

or even sample attacks as these will give

good initial results.

Transform Multiplication Ideas:

λ

Try using the same sound for both

samples.

λ

Splicing silence to the beginnings or

ends of short samples can change the

spectral characteristics of the result.

λ

Using speech as one of the sources, it is

possible to “speak from within" violins, bassoons, cymbals, etc.

 

 

 

Drag both samples into their own audio tracks in Ableton. Open a Vocoder on one track, go to Carrier, and select External. Then select the other audio track. From there on is tweaking the unvoiced/voiced, gender, dry/wet, bands, etc. You could also reverse this and put the vocoder on the other track, then export both layer the two.

  On 9/16/2013 at 3:09 AM, MisterE said:

. there's this synthesizer vsti for example which i believe you can load 2 different samples into and have them morph into each other in various ways.

 

damn, that Alchemy synth sounds pretty intense. plus it is pretty much exactly what the op was asking for.

 

I knew Camel Audio was working on stuff like this a while back but i had no idea it had gotten this advanced.

 

listen to that audio demo on their site with the woman speaking. really good!

The grain delay in ableton can do just about everything that Alchemy synth can do.

Alchemy is cool but still sounds inferior to kyma in terms of a realistic blend. It's really surprising to me that a 10+ year old algorithm has never been out-done by any independent software company.

wow, this is all fascinating.. I've sometimes wondered the best way to do this myself. Great ideas/tips guys. I wonder if rdj used one of these morphing methods (maybe something like kyma) in windowlicker to get some of those crazy glassy morphing/stretching sounds?

Edited September 17, 2013 by Lane Visitor

maybe a vocoder with one sample has the carrier

  On 9/17/2013 at 11:51 PM, Lane Visitor said:

wow, this is all fascinating.. I've sometimes wondered the best way to do this myself. Great ideas/tips guys. I wonder if rdj used one of these morphing methods (maybe something like kyma) in windowlicker to get some of those crazy glassy morphing/stretching sounds?

 

I'm pretty sure a lot of that was done with spectrographic vst/dsps.

S-Layer from Twisted Tools is excellent for this, although you would need a full version of Reaktor.

I used to use a Mac software that did this... I think it was a beta version of Sound Hack or something... (in the same folder, fuck that was like 10 years ago, sorry can't remember...)... Or some other freeware thing at the time. Sorry, no help. However, I do recall that the output sounded all garbled, so I used it to make my samples sound more watery.

 

I have not used software that can somehow perfectly morph 2 sounds (besides vocoder, but there's a lot of manual work involved).

 

Alchemy seems impressive.

For a successful transition between samples, you need to put in work, apply different effects depending on the sound characteristics of the sounds you're trying to blend.

 

  Quote

 

  On 4/17/2011 at 4:00 PM, 'Jonas' said:

Our ears are very sensitive to pitch changes and recognizing physicality in a sound. For a morph to be successful it helps if you can emphasize similarities or differences between the sounds and create some kind of intermediate stage first (a lot of pre-sculpting work). It depends on the source material, what settings will be most useful keeping in mind the trade of between time and frequency resolution using different phase-vocoding settings...

 

What I found useful is not choosing two completely different sounding material beforehand, but have the first transformed sound determine the next choice.

 

Electro Acoustic maestro Trevor Wishart is / has been investigating sound morphing for quite some time and has designed software for it working with CDP. Some observations from the CDP site on morphing ( http://people.bath.ac.uk/masjpf/CDP/sndexs/exmph1mo.htm ) :

 

  Quote

 

General observations

• Morphing with sounds generally takes longer than morphing with visual images. The ear is acutely sensitive to the minutest change and is not easily fooled.
• When there are significant amplitude envelope differences between the sounds (e.g., the second one is rhythmic), one technique is to impose the amplitude envelope of the second sound on at least a part of the first before morphing (ENVELOPE REPLACE). Blurring might also be used to minimise various peaks in one or both sounds.
• A smooth morph may not sound any different from a cross-fade. There are times when one wants to feel the first sound being pulled out of shape and reformed as something new. Determining the nature and degree of the mid-point distortion takes considerable thought and, more often than not, experimentation.
• A judicious use of the stagger parameter can help to place the morph section after a portion of unaltered first input and before the end of the second. This helps to give the ear time to hear the before and after, and also results in a longer output soundfile.
• Some sense of disturbance or warping of the first sound as the morph gets underway can add to the drama and effectiveness of the morph: you can hear that the first sound has gone awry and don't yet know where it will end up.
• I have been experimenting with amplitude and formant envelope transfers, blurring, transposition, spectral crossing and some aspects of waveshape distortion to achieve this sense of disturbance, with variable degrees of success.
• It is best not to underestimate the challenge of achieving effective audio morphs.

 

A good book to read on sound transformations and composition in general is Trevor's book 'Audible Design' :

 

  Quote

 

Audible Design -- ISBN 978-0-9510313-1-5
A plain and easy introduction to practical sound composition
from Electronic Music Foundation or Digital Music Archive or Integrated Circuit Records

 

  On 9/18/2013 at 10:48 AM, Djeroek said:

For a successful transition between samples, you need to put in work, apply different effects depending on the sound characteristics of the sounds you're trying to blend.

 

  Quote

 

 

  Quote

Our ears are very sensitive to pitch changes and recognizing physicality in a sound. For a morph to be successful it helps if you can emphasize similarities or differences between the sounds and create some kind of intermediate stage first (a lot of pre-sculpting work). It depends on the source material, what settings will be most useful keeping in mind the trade of between time and frequency resolution using different phase-vocoding settings...

 What I found useful is not choosing two completely different sounding material beforehand, but have the first transformed sound determine the next choice.

 

Electro Acoustic maestro Trevor Wishart is / has been investigating sound morphing for quite some time and has designed software for it working with CDP. Some observations from the CDP site on morphing ( http://people.bath.ac.uk/masjpf/CDP/sndexs/exmph1mo.htm ) :

 

  Quote

  General observations

• Morphing with sounds generally takes longer than morphing with visual images. The ear is acutely sensitive to the minutest change and is not easily fooled.
• When there are significant amplitude envelope differences between the sounds (e.g., the second one is rhythmic), one technique is to impose the amplitude envelope of the second sound on at least a part of the first before morphing (ENVELOPE REPLACE). Blurring might also be used to minimise various peaks in one or both sounds.
• A smooth morph may not sound any different from a cross-fade. There are times when one wants to feel the first sound being pulled out of shape and reformed as something new. Determining the nature and degree of the mid-point distortion takes considerable thought and, more often than not, experimentation.
• A judicious use of the stagger parameter can help to place the morph section after a portion of unaltered first input and before the end of the second. This helps to give the ear time to hear the before and after, and also results in a longer output soundfile.
• Some sense of disturbance or warping of the first sound as the morph gets underway can add to the drama and effectiveness of the morph: you can hear that the first sound has gone awry and don't yet know where it will end up.
• I have been experimenting with amplitude and formant envelope transfers, blurring, transposition, spectral crossing and some aspects of waveshape distortion to achieve this sense of disturbance, with variable degrees of success.
• It is best not to underestimate the challenge of achieving effective audio morphs.

 

A good book to read on sound transformations and composition in general is Trevor's book 'Audible Design' :

 

  Quote

 

Audible Design -- ISBN 978-0-9510313-1-5

A plain and easy introduction to practical sound composition

from Electronic Music Foundation or Digital Music Archive or Integrated Circuit Records

 

I was going to mention Trevor Wishart. We all got put onto Sound Loom at university - mainly cause I think our lecturer was good mates with Wishart - for one of our synthesis modules and despite it being one of the most un user friendly pieces of software I have ever used, the results were fucking fantastic.

  On 9/19/2013 at 2:05 AM, b born droid said:

....despite it being one of the most un user friendly pieces of software I have ever used, the results were fucking fantastic.

 

 

yep, program use is very archaic, windows 95 feel. Version 6.0 should have streamlined some functions in a more friendly way compared to earlier versions, removing unnecessary steps between transformations. Still I wouldn't recommend it for purchase, better off with a freeware vst like dtblkfx ( http://rekkerd.org/dtblkfx/ ).

 

If you don't mind going hardware I'd highly recommend the Electrix Warp Factory for this sort of thing. They're pretty hard to come by these days though.

nah. forget it. probably not your thing.

Edited September 19, 2013 by Gocab

  On 9/19/2013 at 2:50 AM, Djeroek said:

 

  On 9/19/2013 at 2:05 AM, b born droid said:

....despite it being one of the most un user friendly pieces of software I have ever used, the results were fucking fantastic.

 

 

yep, program use is very archaic, windows 95 feel. Version 6.0 should have streamlined some functions in a more friendly way compared to earlier versions, removing unnecessary steps between transformations. Still I wouldn't recommend it for purchase, better off with a freeware vst like dtblkfx ( http://rekkerd.org/dtblkfx/ ).

 

 

so have you guys had a chance to use Kyma? I'm curious how much the CDP spectral morphing compares, because it's equally as much of a pain in the ass to use that program as it is to 'prepare' a sound file in Kyma for a spectral morph (drawing the fundamental frequency with dots on a graph with a GUI that feels like windows 3.1)

I bought a copy of CDP about 5 years ago but didn't get any results out of it I liked, but that was probably because I had no idea what the hell I was doing.

Trevor Wishart is awesome though, his Voiceprint release has some very good examples of spectral morphing sounds being used in a composition. Mostly voices turning into inanimate objects, things like that, but they sound amazing.