High quality audio morphing effect plugin, why it doesn't exist yet?

[secretkillerofnames]

A fun thing to do in Metasynth is in the Image Synth, create a spectral analysis from a sound file, set a key and tuning map, apply that analysed values to one (or more) samples. It's not really "Morphing" but it produces similar / more interesting results.

[pdxindy]

yep, exactly this can be done in metasynth as well, all in the same application. you can analyse an audio file and turn it into an image which is resynthesized in an additive way. you can do this with a second audiofile and then just blend the images together the way you like, using all the copy/paste functionality and painting functionality that metasynth offers.

Are you sure Metasynth supports this way of morphing?
You are speaking about "blend", that's different to "morph" and "warp".

You had to select one area (freehand selection) and another target area (2nd picture, different place, size and shape), than you had to define a time frame to morph from the first shape to the second different shape and render this as your output.
Have you tried WinMorph? There are also some basic tuts included that show the process.
btw, I can't test Metasynth, I'm on Windows.

You can do more than just blend in Metasynth... doesn't matter cause in most cases the results will not be satisfying.

Movies do amazing 3d animation these days... and they use lots of custom tools specific to certain tasks. It is relatively trivial to create a specific use algorithm. Creating a general algorithm is far far harder so lots of the software used is too specific to be releasable as a general purpose product.

Same thing with audio... Tv commercials and movies can afford to pay someone the time it takes on a specific example. Someone may write a script that does a good job morphing from one specific type of sound to another specific one. But a general purpose tool that can morph anything to anything? Nobody can do it...

And any examples that are impressive, one has to know how many other examples were not impressive to know what % of successes one should expect... It is very low for this sort of stuff

[pdxindy]

since when does saying 'crossfading pictures of two audio files is not the same as crossfading those two audio files' become a claim that 'spectral proessing doesnt work'.

but it is the exact same set of 1's and 0's so it should be the same no?

[antithesist]

http://music.dartmouth.edu/~larry/SHPap ... ticle.html

[whyterabbyt]

since when does saying 'crossfading pictures of two audio files is not the same as crossfading those two audio files' become a claim that 'spectral proessing doesnt work'.

but it is the exact same set of 1's and 0's so it should be the same no?

Imagine your sine and square waves are represented via black pixels. the 50/50 crossfade between a black pixel and a white pixel is a grey pixel. You're going to wind up with an image containing a grey sinewave superimposed on a grey squarewave. How does that represent the audio mix between a sinewave and a squarewave?

Try it.


(and even if I'm wrong, the statement still has absolutely nothing to do with an opinion on whether spectral processing works. I wasnt talking about spectral processing, or anything to do with it, so the assertion that what I was talking about was somehow implying I 'didnt believe spectral processing works' is the purest bullshit. and yet monas repeated it, even after I told him I believed nothing of the kind.

im not the only person monas tried that on with with either.

just that you don't have much experience with spectral processing doesn't mean you have to say it's all BS. why not be open minded and learn something?

that's to someone who merely said that 'morphing' doesnt automatically mean 'spectral morphing'.

(and it doesnt, otherwise the adjective 'spectral' wouldnt need to be applied to differentiate it. )

that person said nothing about it being 'all BS'. nor did they indicate anything which meant they had no experience with it, quite the contrary.

i have no reason why monas keeps trying to assert that the people he's chosen to disagree with are attacking or 'disbelieving' spectral processing. it seems nothing more than simple trolling to me.)

[monas]

@ whyterabbyt
dude, seriously, quit replying to my posts. you're acting childish, insulting and weird. i have no clue what's wrong with you

there's plenty of concepts already mentioned about morphing soundfiles from good sources. none of them wouldn't involve spectral processing. morphing doesn't really make sense thinking of it in time domain and frequency domain is hard to wrap your head around. it took me a long time to get a vague idea of what happens there and i don't blame anyone for not grasping it. i blame people for making silly, and useless remarks though.

if anyone has a solid idea of how morphing could work outside of frequency domain i am eager to hear it. (by the way, i've already mentioned the filter mask method. it can hardly be done in realtime though)


@ pdxindy
why so pessimistic? of course morphing "anything to anything" doesn't work. as mentioned before, the sources have to be well chosen. and just because 99% of all morphing attempts maybe fail doesn't mean there can't be a general purpose tool. it's like with compression. of course 99% of all compressor settings on a track are "wrong". it's my job to find the 1% that works. that doesn't mean there can't be a general tool for compression.
think of it like picture morphing. morphing two faces can give intriguing results. but morphing two random photos will end up in a blurry mess. but that's not the morphing algorithms fault.

[whyterabbyt]

@ whyterabbyt
dude, seriously, quit replying to my posts. you're acting childish, insulting and weird. i have no clue what's wrong with you

what's 'wrong with me' 'dude', is that I wasnt replying to your posts at all, but you started replying to mine, interjecting while Im conversing with someone else, missing the point and arguing over stuff that was unrelated to an explanation I was trying to give them, and attributing statements to me which had no basis in anything I'd posted.
notably, as per this reply, you've continued to do so.

you're the one who started acting childish, insulting and weird. particularly when you repeated the assertion that I didnt believe spectral processing worked after I'd already told you, politely, that you had no basis for that assertion.

if you want to jump into a conversation im having with someone else, argue with minor trivia that's unrelated to my point, then attribute things to me which are unrelated to snything Ive said, and then act like Im at fault for reacting to that, you tried it on the wrong guy.

and, like i said, i'm not the only person you've tried your straw man crap on.

[antithesist]

http://music.dartmouth.edu/~larry/SHPap ... ticle.html

[highkoo]

You guys gave me another piece of a half an idea...

Go to the websites for various Adobe PS/AE plugins,
Dump their sample before/after pics into Metasynth and see what the plugins can 'sound' like.

Working in Metasynth, you won't want to use any video morphing plugins. Metasynth represents a sound of X duration as a single image, so you would want to use a filter that could somehow "interpolate" points between the left and right. I'm not sure what that would look like.

I wonder if there are more of these morphing type tools on OSX, that derive from the work done in the 1990s on Macs. Back when I was young(-ish) and poorer, I used Csound on a Pentium II running Windows 98 to create my sounds. I always had serious envy of those folks that had Macs, as they could run Soundhack, Supercollider, Metasynth, Max/MSP, and so on. All of these tools are still available, and I think that Soundhack and Metasynth are OSX only.

Sean Costello

I didnt really explain myself very well....
In my mind the process I describe would really only be useful to a dev who was looking at image morphing as a way to inform an audio morphing project.
It would just give an idea of what the visual code was doing when run through the Metasynth conversion.
From what I have seen, a lot of the image synth experimentation works with fairly basic imagery and basic image manipulation. Some graphic plugs are capable of very subtle and very intricate stuff, so I wonder if that code has ever been shoved though Metasynth in any way.

I guess I am thinking of this as if there were a universal visual>audio standard that is non-configurable, but just has never been explored.
Which is basically wrong, but it provides a place to start from.
And so there are this pile of hi-end visual plugins that have never been applied to that conversion standard.

So, there is an Adobe plugin that does XYZ. Given a specific Metasynth set up, what does the 'dry' image sound like, and what does the 'wet' image sound like? The plugin itself is not useful, but the two sounds might inform us on what that visual code does sonically, and what a useful audio plugin might have to have inside it.
For PS plugs, you just take the processed images, and for the AE plugs you just take screen shots that show the breadth of the effect it is producing. Just to get an idea. I would guess that even video plugins sometimes use images to show what they do.

It seems to me (likening programming to magic ) that making a morpher from scratch, or making a visual>audio system from scratch, are massive projects. Maybe some fancy Adobe could be informative and have a place to hitchhike from...

[pdxindy]

@ pdxindy
why so pessimistic? of course morphing "anything to anything" doesn't work. as mentioned before, the sources have to be well chosen. and just because 99% of all morphing attempts maybe fail doesn't mean there can't be a general purpose tool. it's like with compression. of course 99% of all compressor settings on a track are "wrong". it's my job to find the 1% that works. that doesn't mean there can't be a general tool for compression.
think of it like picture morphing. morphing two faces can give intriguing results. but morphing two random photos will end up in a blurry mess. but that's not the morphing algorithms fault.

I was answering in the context of why developers have not created more morph tools...

your compressor analogy is wrong... it is up to you to find the 1% settings in the compressor where it works... but those settings are there and findable in most if not all cases and with not that much time needed

with morphing, there is no general purpose tool that can consistently find a useful solution... the task is magnitudes more difficult...

Users, when they think of morphing, imagine all sorts of morphs from A to B that would not give a result that they imagine. That is, they would not be happy with the result, therefore not happy with the tool.

So there is not a good incentive for a developer to invest a huge amount of development time for a tool that is very much a specialty tool, and which will not be useful in the vast majority of cases anyway.

[monas]

hmmm...
there's plenty of articles that were posted on the issue of morphing here.
i don't believe that the "fact" that you can't get a consistent and good result from a morphing tool is the reason why developers haven't invested in it. i think it's simply because it's a kind of processing that's too exotic. why run a risk when there's easier money elsewhere?

[lofty]

...

[vurt]

ive been thinking about treating myself to photosounder for my birthday in a few weeks.
if any wants to treat me to alchemy...

[monas]

@ lofti

yep. 1) is what i am working on right now in reaktor.

2) is problematic, because i think you need to work with the frequency data here, and that's the hard part, not the amplitude morph. interpolating i/r values or amplitide/phase pairs don't really work. (does it? can someone tell me how?). what you need is amplitude/frequency, which can't be resynthesised by iFFT, only in an additive way. if i get this up and running it should make a lot more than just morphing possible in good old reaktor, and i am already most of the way there.
fine tuning isn't going to happen of course, but modulating the morph factor is a nice idea

i think i'll start a thread about it over at reaktor board where discussions stay productive and technical (mostly)
fft geeks are welcome to join in.

[Neon Breath]

@ lofti

yep. 1) is what i am working on right now in reaktor.



i think i'll start a thread about it over at reaktor board where discussions stay productive and technical (mostly)
fft geeks are welcome to join in.

good idea