Audio reverse engineering questions & ideas

Official support for: meldaproduction.com
Post Reply New Topic
RELATED
PRODUCTS

Post

While exploring sound design I sometimes go backwards and reverse engineer a sound that is near to what I am looking to recreate/ use for inspiration. The process has me wondering if reverse engineering plugins exist and if not, who better to distract and derail from all of his current projects than Vojtek!

Things that I am looking for/ wondering about:
- Can a sample's waveform be imported/ imprinted onto the background of MPowerSynth's ADSR graph in order to sync the sample and then mirror its waveshape with the ADSR?
- Would the above be an accurate representation of the sample's movement in time?
- Can MAnalyzer or any of the EQ analyzers be connected to the EQ bands in order to create a 'sonic/ tonal impulse response' of the sample? Thereby displaying the tonal tendencies and harmonic structure of the analyzed sample. These results could be filtered according to a comparison with basic building block tones - sine, triangle, square, saw, noise.

Obviously, these would need a dry sample to start with because effects would have affected things further from the starting point. Though suggested effect chains might be interesting too!

If any of these are plausible ideas...could they be considered for inclusion in MYYY?
(No, I have no clues about where MYYY is in development or what it includes, I'm just imagining adding creative approaches to a synth engine.)

I'd love to hear alternative reverse engineering approaches/ ideas from you guys, maybe all I need is inspiration or instruction!

Thanks,
werzel

Post

So your looking to reverse engineer a synth sound?
Well the answer to your first question, yes you can copy the amp envelope of any sound exactly using the sidechain of Phatik.
Also it is possible to analyze sound files and use them as oscillators.
And Freeform EQ/or Convolution can match a sounds frequency response.
Good if you are trying to clone a sound, but not really reverse engineering. As it doesn't show how/what was used to get the sound.
Jason @ Melda Production

Post

Those ideas might have some use, but I don't think they will work out well if your trying to reproduce the sounds you hear in songs. That method is similar to measuring the size and density of a cake to figure out the recipe.

What has worked for me is watching tutorials and messing around with FX until I understand what things sound like. Once I have the sound of certain FX or techniques in my head it's easy to identify when others use them and copy what they did. For example if you play around with low pass filters long enough you'll understand what they sound like and all the different variations in ways they're used. Before I had no idea why on when to use them and a lot of synth sounds were a mystery to me. Now I can easily hear them in pads, plucks, bass sounds etc.

Another good method is look at other people's presets. Look at how they route everything and then try turning things off and see what difference it makes in the sound.

Once you know what technique is used to create a sound, reverse engineering it is somewhat easy. Of course getting the nuances perfect is difficult.

Post

Well, it's not that simple really. If it would, everyone would do that and no samplers would be needed :D. The problem is that the envelope and spectrum are only a few of many parameters of the sound. Sampling is currently the only reliable way to do that, not perfect, memory and storage consuming etc., but at least it works. I was thing about some neural networks approach, but no time for that yet, we'll see in the future. Anyways what you suggested could get you something, but won't work well at all in general.
Vojtech
MeldaProduction MSoundFactory MDrummer MCompleteBundle The best plugins in the world :D

Post Reply

Return to “MeldaProduction”