How to best use audio-to-midi?
-
Newbie Brad
- Member
- Posts: 31
- Location: Nashville
- Contact:
Figure out what range of frequencies your audio input will be inputing.
The human voice is about 300 Hz to 3000 Hz, actualy sung tone being on the lower end and fricatives and plosives on the other end. For voice, the overtone harmonics and plosives/fricatives are useless so band pass to eliminate low boomy background noise and pretty much the entire range above your highest sung note works well.
The human voice is about 300 Hz to 3000 Hz, actualy sung tone being on the lower end and fricatives and plosives on the other end. For voice, the overtone harmonics and plosives/fricatives are useless so band pass to eliminate low boomy background noise and pretty much the entire range above your highest sung note works well.
-
Newbie Brad
- Member
- Posts: 31
- Location: Nashville
- Contact:
Thanks, very nice. Could you suggest a range and approach for my electric guitar?
No, sorry, I'm no guitar player. I only like playing nylon string flamenco ones.
I can, however, recommend you poke around on Fretted Synth's site, as I use his pitch-detecting guitar-to-synth plug-ins and I bet there's some intelligent information there. I'm sure you know the site - version 3 of his ampsim is bumped every day on KVR.
I can, however, recommend you poke around on Fretted Synth's site, as I use his pitch-detecting guitar-to-synth plug-ins and I bet there's some intelligent information there. I'm sure you know the site - version 3 of his ampsim is bumped every day on KVR.
Just for information, the audio to midi conversion works quit well on voice too if you fidle with it to find the good setup.
I've been pretty succesful with the following chain on a (isolated) voice track.
Sampleplayer (mono) -> Compressor/Limiter -> Gate -> AUdio to midi _> Piano roll.
The compressor/limiter is used to get as much as possible dynamics out of the vocal track, this way it's easier to set parameters for the audio to midi component. The gate takes out unwanted noise (if there is any) to avoid unwanted notes.
Then you'll have to play with the tresshold and latency parameter of the audio to midi component.
You can get a quit accurate result. With a vocal track you'll have to delete the ghostnotes you get but the end result is pretty good.
I've been pretty succesful with the following chain on a (isolated) voice track.
Sampleplayer (mono) -> Compressor/Limiter -> Gate -> AUdio to midi _> Piano roll.
The compressor/limiter is used to get as much as possible dynamics out of the vocal track, this way it's easier to set parameters for the audio to midi component. The gate takes out unwanted noise (if there is any) to avoid unwanted notes.
Then you'll have to play with the tresshold and latency parameter of the audio to midi component.
You can get a quit accurate result. With a vocal track you'll have to delete the ghostnotes you get but the end result is pretty good.
Free samples-vsti's-artist hosting at www.kara-moon.com
Music forum at www.kara-moon.com/forum
Did I mention how great Usine is ?
Music forum at www.kara-moon.com/forum
Did I mention how great Usine is ?
What does anyone think the most effective way to capture midi notes from audio-to-midi along with midi CC automation played live (in my casse via p5 glove) and send it off to another program would be?
I've yet to find a way to efficiently capture a performance this way in order to "score" my USB-mic and VR glove data that can then be put in FL for quick composition, bouncing things back and forth.
I've yet to find a way to efficiently capture a performance this way in order to "score" my USB-mic and VR glove data that can then be put in FL for quick composition, bouncing things back and forth.
no ideas?
hmm I may have to post what I've come up with and see if it makes any sense to my colleagues
hmm I may have to post what I've come up with and see if it makes any sense to my colleagues
I have tried, but failed, to create a workspace in which the audio-to-midi will only retrigger a new pitch when you kit a midi controller key or midi button or somesuch, and otherwise will send pitch variation to the pitchbend channel so that the notes don't skitter around all over in semitone steps but haven't gotten it to work.
Has anyone got any insight as to whether this is possible or not?
Has anyone got any insight as to whether this is possible or not?
What's the problem? That you can't get a trigger of the new note when you want, or that the pitchbend isn't working properly? Or both?
Bjørn S
Um, neither.
I'm trying to make something which will *not* allow a new note to retrigger just because the audio note has changes, and instead send out one-octave pitchbend information that tracks the audio's pitch *until I hit a key to say it's ok to retrigger* so that controlling VST synths when I glide up or down notes. I don't want semistone steps, so I've been trying to find something to do this forever. I figured I *might* be able to get Usine to do this.
I guess I don't ask for help here all that often because I can finally get Usine to do most anything I'd like it to hehe.
Essentially I want the audio-to-midi to track in cents instead of full semitones, but VSTs don't respond to note on messages that way.
I figure that I've gotten Fractal Tunesmithy to work via MidiOx to re-tune things to weird scala scales, so there must be some way to trigger a midi note based on whatever it thinks is the closest semitone, and then send pitch bend information for fine pitch control until a new note is allowed to trigger, in my suggestion above sent manually by me via a keystroke or midi controller button or something like that.
Alas, I've not figured out a way to make this happen yet. I have managed to create some pretty interesting modular set-ups that do all sorts of weird things via the analysis modules, just not this thing.
For that matter, somehow forcing the audio to modo to send a new note only when I hit a key, and at that time determining what the incoming audio's pitch is in order to send that in the note on message would be pretty useful.
That way incidental noise coming into the mic wouldn't trigger a note, only when I was holding down a key. Now that I've thought this out in posting it I think I can get *that* to work at home. Hmm...
I'm trying to make something which will *not* allow a new note to retrigger just because the audio note has changes, and instead send out one-octave pitchbend information that tracks the audio's pitch *until I hit a key to say it's ok to retrigger* so that controlling VST synths when I glide up or down notes. I don't want semistone steps, so I've been trying to find something to do this forever. I figured I *might* be able to get Usine to do this.
I guess I don't ask for help here all that often because I can finally get Usine to do most anything I'd like it to hehe.
Essentially I want the audio-to-midi to track in cents instead of full semitones, but VSTs don't respond to note on messages that way.
I figure that I've gotten Fractal Tunesmithy to work via MidiOx to re-tune things to weird scala scales, so there must be some way to trigger a midi note based on whatever it thinks is the closest semitone, and then send pitch bend information for fine pitch control until a new note is allowed to trigger, in my suggestion above sent manually by me via a keystroke or midi controller button or something like that.
Alas, I've not figured out a way to make this happen yet. I have managed to create some pretty interesting modular set-ups that do all sorts of weird things via the analysis modules, just not this thing.
For that matter, somehow forcing the audio to modo to send a new note only when I hit a key, and at that time determining what the incoming audio's pitch is in order to send that in the note on message would be pretty useful.
That way incidental noise coming into the mic wouldn't trigger a note, only when I was holding down a key. Now that I've thought this out in posting it I think I can get *that* to work at home. Hmm...
-
grego mondo
- Member
- Posts: 466
- Location: 22110 Rostrenen
- Contact:
audio to midi seems hard to ajust for (big) tuba (lots of harmonics) . . .
I tried to convert an audio loop to midi so I can record a loop of tuba and then it's played by a midi instrument but I couldn't get a clean result, may be I should try to compress the incoming sound (as it's written up there)...
I tried to convert an audio loop to midi so I can record a loop of tuba and then it's played by a midi instrument but I couldn't get a clean result, may be I should try to compress the incoming sound (as it's written up there)...
En attendant cordialement l'apocalypse ... 100% Usine !!
https://lesformesdufond.kaz.bzh/grego-mondo/
PC i7 2.6Ghz - Win11 / RME Fireface UFX
Tubas / Serpent / Music Box
Interfaces Midi / OSC / DMX
https://lesformesdufond.kaz.bzh/grego-mondo/
PC i7 2.6Ghz - Win11 / RME Fireface UFX
Tubas / Serpent / Music Box
Interfaces Midi / OSC / DMX
grego mondo , maybe you can also add a filter or /and an equalizer to remove the harmonics...?
just an idea...
just an idea...
-
grego mondo
- Member
- Posts: 466
- Location: 22110 Rostrenen
- Contact:
good idea, I'll try ... thanks
right know I'm trying to run 2 FP10(chainned) with Usine
right know I'm trying to run 2 FP10(chainned) with Usine
En attendant cordialement l'apocalypse ... 100% Usine !!
https://lesformesdufond.kaz.bzh/grego-mondo/
PC i7 2.6Ghz - Win11 / RME Fireface UFX
Tubas / Serpent / Music Box
Interfaces Midi / OSC / DMX
https://lesformesdufond.kaz.bzh/grego-mondo/
PC i7 2.6Ghz - Win11 / RME Fireface UFX
Tubas / Serpent / Music Box
Interfaces Midi / OSC / DMX
-
RunBeerRun
- Member
- Posts: 18
- Contact:
I started using the audio to midi with my guitar today, I run a maximizer before it so my volume is pretty loud.
-
Newbie Brad
- Member
- Posts: 31
- Location: Nashville
- Contact:
Hi RBR, I like your videos and things. NB
-
nix
Hey folk,
I am curious to find how it performs(another guitar player). I guess I'll get a license soon. For another pitch to midi option(which needs an update) you could try 2SYN, available here, along with some other pluggable goodies(hopefully)->
http://rekkerd.org/nix-plugs/
Wishing contents in creation, nix
I am curious to find how it performs(another guitar player). I guess I'll get a license soon. For another pitch to midi option(which needs an update) you could try 2SYN, available here, along with some other pluggable goodies(hopefully)->
http://rekkerd.org/nix-plugs/
Wishing contents in creation, nix
Hello everyone.
I wanted to start a new topic but seems like I can not because I am a newbie.
So this is going to be a little off topic. Sorry about that.
I use an old Intel Pentium4 with onboard soundcard (I guess the maindboard says something about "PEARL 845"... its an original Intel one). 512Mb ram. And I got that ASIO4ALL driver.
The question is that I want process in realtime so I can play along with the drummer tapping the tempo and aplying FX to the incoming audio... with real (decent at least) latency... I dont care much to quality audio as long as it sounds decent... (I use some tape recordings too).
Since I am using Usine I tried an awkward attempt to do something realtime with an 8 minute .WAV file. I feel like I am missing something because I cant get the right timing with me tapping to making a couple of drum loops to fit the performance in the .WAV file.
I really apreciate so much any ideas, advice and stuff.
I dont have an internet connection and I usually do the net stuff from a friend pc. So its possible I will read any reply in like week or two. Dont hurry.
Thanks.
and goodluck in that usine user meeting thing... cheers with that.
and sorry for the poor english.
I wanted to start a new topic but seems like I can not because I am a newbie.
So this is going to be a little off topic. Sorry about that.
I use an old Intel Pentium4 with onboard soundcard (I guess the maindboard says something about "PEARL 845"... its an original Intel one). 512Mb ram. And I got that ASIO4ALL driver.
The question is that I want process in realtime so I can play along with the drummer tapping the tempo and aplying FX to the incoming audio... with real (decent at least) latency... I dont care much to quality audio as long as it sounds decent... (I use some tape recordings too).
Since I am using Usine I tried an awkward attempt to do something realtime with an 8 minute .WAV file. I feel like I am missing something because I cant get the right timing with me tapping to making a couple of drum loops to fit the performance in the .WAV file.
I really apreciate so much any ideas, advice and stuff.
I dont have an internet connection and I usually do the net stuff from a friend pc. So its possible I will read any reply in like week or two. Dont hurry.
Thanks.
and goodluck in that usine user meeting thing... cheers with that.
and sorry for the poor english.
hello welcome to the forum!
Generally the tap tempo is not precise enough to synchronize a drum loop for several minutes. Even if you make only a small error (like 0.5 bpm) you will, slowly loose the synchro.
You'll probably have to make a smaller loop and resynchronize it periodically with the 'restart' buton (in the synchro).
Generally the tap tempo is not precise enough to synchronize a drum loop for several minutes. Even if you make only a small error (like 0.5 bpm) you will, slowly loose the synchro.
You'll probably have to make a smaller loop and resynchronize it periodically with the 'restart' buton (in the synchro).
Olivier Sens
www.brainmodular.com
www.brainmodular.com
Who is online
Users browsing this forum: Google [Bot] and 24 guests
