"phase corrected" DEF and PRF
- Thread starter JohnnyMarks
- Start date
synthetic
Senior Member
I'm not sure it really means anything, if you're talking about the Garritan Strad phase-corrected samples. Their diagram shows two waveforms in perfect sync. Well, actual acoustic instruments don't produce perfect sine waves, let alone super-stable pitches. I could see if they had samples of synthesizer waves, then they might be in perfect sync. But even synths have pitch drift, especially the good ones. Perhaps they have somehow pitch-corrected the waveforms to be perfectly pitched and in-phase, using Melodyne or something. In which case... Autotune Violin?
More to your question, PRF is a way to scale pitches using formant-controlled filters along with pitch. It's a way to get legato or portamento from a non-legato sample. You can control the portamento speed from a MIDI controller. (The one specified in the original MIDI spec for portamento speed! Who knew.)
DEF is a way to add filtering to a sample so that you don't need to use as many crossfades to get realistic dynamics. In that way, the end result is less phase problems between the crossfaded samples. You could replace a three-way crossfade with a DEF filter and it would eliminate the phase problems associated with crossfades.
Did any of that answer your question?
More to your question, PRF is a way to scale pitches using formant-controlled filters along with pitch. It's a way to get legato or portamento from a non-legato sample. You can control the portamento speed from a MIDI controller. (The one specified in the original MIDI spec for portamento speed! Who knew.)
DEF is a way to add filtering to a sample so that you don't need to use as many crossfades to get realistic dynamics. In that way, the end result is less phase problems between the crossfaded samples. You could replace a three-way crossfade with a DEF filter and it would eliminate the phase problems associated with crossfades.
Did any of that answer your question?
JohnnyMarks
Active Member
So "phase corrected" just means "corrects the problem of phasy-sounding velocity layer cross-fading, since you don't have to do this if you're using DEF"? I though it had something to do with correcting phase anomalies introduced by the EQ/spectral processing DEF performs.I'm not sure it really means anything, if you're talking about the Garritan Strad phase-corrected samples. Their diagram shows two waveforms in perfect sync. Well, actual acoustic instruments don't produce perfect sine waves, let alone super-stable pitches. I could see if they had samples of synthesizer waves, then they might be in perfect sync. But even synths have pitch drift, especially the good ones. Perhaps they have somehow pitch-corrected the waveforms to be perfectly pitched and in-phase, using Melodyne or something. In which case... Autotune Violin?
More to your question, PRF is a way to scale pitches using formant-controlled filters along with pitch. It's a way to get legato or portamento from a non-legato sample. You can control the portamento speed from a MIDI controller. (The one specified in the original MIDI spec for portamento speed! Who knew.)
DEF is a way to add filtering to a sample so that you don't need to use as many crossfades to get realistic dynamics. In that way, the end result is less phase problems between the crossfaded samples. You could replace a three-way crossfade with a DEF filter and it would eliminate the phase problems associated with crossfades.
Did any of that answer your question?
The first paragraph is a somewhat different - but interesting! - topic from my intended question. There has been work around phase alignment of audio samples, originally for concatenating syllables in speech synthesis I believe. Anyway, it's complex stuff, papers have been written (e.g. Miller Puckette...google "phase bashing"), and there are a bunch of Italians who are pretty active in this kind of thing, so...the Garritan effort could well be fairly sophisticated.
Thanks.
Dynamitec
Senior Member
Hmm, interesting topic. But only one thought: did nobody notice he only says that the samples are "phase aligend". He didn't say that he really recorded different velocity samples and "phase aligned" them, if i recall this correctly. If anyone would really have a software that can phase align different velocity samples, tihs one would make more money in seeling this software to sample library producers than selling a library with phase aligned samples! :wink:
Best,
Benjamin
Best,
Benjamin
