Raveshaper wrote:What I mean is, mathematically, operations performed on the floating point values of an audio stream behave identically. It all comes down to math in the end.
If I have to get explicit, I will only invest in plugin suites and bundles that are on a platform I can resell and migrate as I choose.
But my main argument is math works the same way regardless of code, and that REs are constrained to the same range of capabilities as stock devices.
As for portability of code issues and optimization problems, I refer you to this thread:
http://www.kvraudio.com/forum/viewtopic ... 5&start=15
And the final feelings from a developer's perspective:
http://www.kvraudio.com/forum/viewtopic ... 5&start=75
You misunderstand his comments, he's mainly talking about vector operations. As said, you can use vector operations in REs now, even do so in a platform independent way - without using assembler code. Maybe you ask him if he thinks his plugins are the same thing as others that exist just because they're math - I'm pretty sure he'll disagree.
Math works the same way in the whole universe (as far as we know) sure thing but how you apply that math is a whole different thing. Otherwise there would be no reason to do research or develop new algorithms, would there? The last two decades saw massive improvements in audio processing algorithms from analog simulation to completely new ways to work with audio material. Melodyne, new denoising algorithms, dynamic convolution processing just to name a few. Some of these just became possible because of the current processing power of computers and were conceptually known for a longer time, some are completely new developments.
So no, I still don't follow you at all - audio algorithms definitely don't behave "identically" just because they're based on math. As Jiggery indicated, thats the same logic as saying all poetry is the same because its based on words or all pictures are the same because they're made with paint.