1. Load music file with a piano recording (arbitrary).
2. Pianoteq analyse the sound (Fourier, etc.) and create FXP that is the best approximation for the given sound on the selected piano model (or even suggest the most suitable model too, e.g. "D4" + generated FXP).
Like this, but automatically.
+1
I remember requesting this function some months ago. If you mean arbitrary recording in terms of a piece it will be a tough job. I don't think it's possible.
My approach was a "sampling" function.
I remember requesting this function some months ago. If you mean arbitrary recording in terms of a piece it will be a tough job. I don't think it's possible.
"Arbitrary" is any piano piece, without additional instruments and garbage. I'm sure it's not so hard to identify notes/chords, extract it, make fourier and other analysis. The whole piece provide dynamic range, EQ... I'm not a sound engineer, but I guess this is possible. No need to manually extract note samples from the piece.
"Arbitrary" is any piano piece, without additional instruments and garbage. I'm sure it's not so hard to identify notes/chords, extract it, make fourier and other analysis. The whole piece provide dynamic range, EQ... I'm not a sound engineer, but I guess this is possible. No need to manually extract note samples from the piece.
I don't think you'll get proper results this way. I see the major problem in matching timbre and dynamics. I think you need at least some "velocity layers" per note.
This request require something ver complex, perhaps beyond the reach of actual technology.
If you would give a few velocities for pianoteq, for each key pedal down, instead of jus any music played on the piano, things could be easier.
But even so the pianoteq technology, for a precise timbre recreation, require a detailed analyze that take a relative long time.
In the best, with actual technology, you could get a pianoteq piano model, similar to a given piano you want to try to emulate, and wait to they develope a key by key analyze to pianoteq automatically adjust the parameters (the ones available in Pro version interface and the ones hiden in the progran plataform) to try to match the best as possible.
1. Load music file with a piano recording (arbitrary).
2. Pianoteq analyse the sound (Fourier, etc.) and create FXP that is the best approximation for the given sound on the selected piano model (or even suggest the most suitable model too, e.g. "D4" + generated FXP).Like this, but automatically.
Interesting idea. EQ and tone matching with external apps or devices might get you close, eg. iZotope Ozone (EQ match) or Fractal Audio Axe-FX (Tone match).
Although I have not tried this particular piece of software, please see this Dynamic Spectrum Mapper video. It gets interesting starting about 2:05, as the software may approach the very effect you desire. At 4:08, the video suggests you can apply the spectrum of one thing [i.e., the piano in a particular recording you wish to emulate] to anything else [i.e., Pianoteq].
Agreed, this does not produce an .fxp for Pianoteq, but playing back through an instantiation of DSM might work.
Cheers,
Joe
Hello again,
Following vestapol's advice from two articles ahead of this one, please have a look at the way Izotope Ozone 5 Advanced may be used to achieve similar results as DMS (using EQ instead of multiband compression) above:
https://www.youtube.com/watch?v=NRThYSuqcQ4
The Izotope video gets interesting at 1:50, where it discusses EQ mapping from a reference source.
Cheers,
Joe
Very interesting, Joe. Thanks a lot!
Something like the Izotope with multiple layers could be the starting point. Such a function would be great even if it wasn't fully automatic.