Is crowdfunding an option for my sound design tool?

[MackTuesday]

Interesting tool definitely, do you have sound examples of the algorithms?

I've finally been able to throw together a sound clip for public consumption. It's not much, but I'll try to keep putting up more.

This is a demonstration of my "demod-remod" effect. I multiply one of the analysis components by 2 and resynthesize.

[ghettosynth]

This sounds very interesting. Heavy FScape and CDP user here. Been waiting for a binary version of Fscape Next for what seems like forever!

Thanks for the reference, I didn't know about FScape. Mostly posting so that this shows up yellow in my feed.

[cron]

Interesting tool definitely, do you have sound examples of the algorithms?

I've finally been able to throw together a sound clip for public consumption. It's not much, but I'll try to keep putting up more.

Interesting stuff. Sounds like an non-linear frequency shift, but not quite like any other I've heard. Keen to hear more.

Thanks for the reference, I didn't know about FScape. Mostly posting so that this shows up yellow in my feed.

It's an interesting toolbox but I'm not sure anyone but the creator knows how to use it properly. I mostly use the wavelet transform as it spits out a wav. You can then process the result elsewhere and bring it back to FScape for the inverse transform. You can get all sorts of subtle effects provided you don't disrupt the 'time structure' of the analysis, and all sort of frequency mirrored, time smeared madness if you do.

The Serial Killa (energy output) and Seek And Enjoy modules can be used together to do my favourite audio party trick: cut sound into loads of tiny chunks then rearrange them from loudest to quietest. If you use the noise analysis mode instead of the energy mode on voice, you end up with all the sibiliants and mouth noises rapid-fire jammed together at the end of the file - a gloriously horrible noise.

[MackTuesday]

These processes sound similar to dictionary-based/matching pursuit granular processes. Kinda like the granular equivalent of resynthesis where resynthesis is achieved by selecting and scaling elements from a dictionary of 'atoms'. Any similarities there?

I looked into that approach, even took a swing at regression analysis, but I didn't see how to make it computationally cheap.

[nerd]

I have no idea about crowdfunding and sound mangling, but I like the idea to have a "wave editor" (?) as sketch and design tool instead of late/post production tool or DAW.

(I can't really tell why. Recording takes into the DAWSs is somehow not fast enough, trimming takes and quanzing their positions, there's not really the perfect tool for it. Most applications seem to offer a lot more and nothing that would be quicker than usual for the recording. Though, I could be building workflow castles in the sky.)

[kamalmanzukie]

I've made a leap of faith and moved to part time work in order to have more time to develop a sound design idea that doesn't seem to exist*. I think it's definitely marketable.

Making money part time isn't working as well as I'd hoped. I'm living like a monk, and that's fine, but I've had to take on more work to make ends meet. I'm working about 12 hours a day and I simply don't have time to work on my project.

If I could get some support as I'm working on it, it would help a lot. Is crowdfunding worth an attempt? Where might be the best place(s) to introduce it?

*So here's the idea. Applications like Sound Forge are very goal oriented. Their features are chosen for getting certain tasks done. There's nothing in there for experimentation and getting unusual results. The good ones are also quite expensive. We have Fscape, whose experimental disposition is refreshing, but the interface is rudimentary and the documentation lacks enough information about the algorithms to empower the user to make good use of them.

I have effects that give weird but sonically viable results. For example, I have a spectral decomposition algorithm that separates sounds into sinusoidal partials, noise, and transients, but also components with intermediate time/frequency localization. Furthermore, it's designed so resynthesis gives you sample-accurate reconstruction if you don't alter the analysis components. The same can't be said of McAulay-Quatieri or traditional bandwidth-enhanced methods, which throw away the original sound and rely on the fidelity of the resynthesis engine to get good output.

I have another decomposition algorithm based loosely on Hilbert-Huang that separates a sound into sonically useful but unusual pieces that can also be sample-accurately reconstructed. I have never seen its like in other sound design tools.

I'm also working on an idea inspired by attempts at circuit emulation that allows users to explore systems of differential equations. The trick with this idea is in the interface, which makes the process intuitive. Because who wants to tweak individual functions and hope the system is stable?

I would pack this tool with other such strange but useful oddities as well as the usual editing functions.

i've built some analysis rensynthesis tools in reaktor. the hardest part is trying to think what it should be used for. it actually stopped me uploading it to their user library. it works, but for what??

[MackTuesday]

i've built some analysis rensynthesis tools in reaktor. the hardest part is trying to think what it should be used for. it actually stopped me uploading it to their user library. it works, but for what??

My preferred technique is to analyze, molest the analysis data, then resynthesize and see what you get. The process is ripe for abuse. Subvert math for your devilish scheming!

[kamalmanzukie]

absolutely. i've developed quite a few different ways of separating sinusoids from noise. sometimes using them in series so successively refine the results. have you ever messed with logarithmic bin spacing at all?

[sonicpowa]

Keen to hear more.

Yes me too, maybe something odd that takes a lot of offline processing.