Hey all. One of the things that has intrigued me most about the V-Synth is it's ability to manipulate sampled vocal lines in a realistic way - like on the video demos ("Feel the shift in time".)
What I want to do is record a friend of mine (who is a fabulous singer) singing a variety of single line phrases (like the demo).
Other than the basic sampling procedures, what should I know in order to make these lines move freely between keys and make them sound good when adding harmonies in? I know about making them legato so they don't retrigger but what else? I have read some on encoding, but am not sure what the best route it.
I am going to try this two ways - using the V-synth to do the samples and using ProTools to record wave files of her and importing those into the V-Synth.
If there are any existing posts about this, please let me know.
Thank you all in advance for your advice!!
Mitch
Sampling Vocals
Re: Sampling Vocals
Basically, you just need to encode your samples with "Solo" encoding. After that, you will be able to play realistic harmonies, stretch in time and tweak the formant. "Solo" is meant for vocal samples.
A.
A.
-
The Sherriff
- Posts: 57
- Joined: 13:21, 12 November 2004
- Location: Tenerife,canary islands
Re: Sampling Vocals
Hi Mitch
I've been trying some similar things with my own V-synth with resonable success. I would recommend recoeding in pro tools first ( I use cubase SX) so you can get the best possible sample to import into the V-synth. Sort out any compression EQing etc. to really get the sound you want then import it and encode it as solo as ART suggets then treat it the same as any other sound and spend weeks tweaking it till you can play it how you want to and get the sound to respond nicely to the way you play.
Best of luck
the Sherriff
I've been trying some similar things with my own V-synth with resonable success. I would recommend recoeding in pro tools first ( I use cubase SX) so you can get the best possible sample to import into the V-synth. Sort out any compression EQing etc. to really get the sound you want then import it and encode it as solo as ART suggets then treat it the same as any other sound and spend weeks tweaking it till you can play it how you want to and get the sound to respond nicely to the way you play.
Best of luck
the Sherriff
sampling vocals
if you have an apple g-5 you can sample using the optical i/o
for the best possible sample .the v-synth has optical i/o s.And this will give you the best out possible input and output.hands down
for the best possible sample .the v-synth has optical i/o s.And this will give you the best out possible input and output.hands down
-
The Sherriff
- Posts: 57
- Joined: 13:21, 12 November 2004
- Location: Tenerife,canary islands
Re: Sampling Vocals
Just to clarify my earlier message I don't actually record the wave in the V synth I import what I create from elsewhere as a WAV file, then edit that in the Vsynth to create any events in the sample needed for the different methods of using the sample. This completely bypasses any need for quality inputs on the Vsynth. If your using a quality sound card or other suitable equipment to record the samples you don't need to re-record the wavs.
cheers the sherriff
cheers the sherriff
Encoding..???'s
before encoding a vocal, should the vocal be sung in a single tone line? or can it be melodic? reason why is cuz I recorded a melodic vocal, straight to the XT, encoded it as 'SOLO' and it sounded funny(not smooth).....then I sang a mono tone of the vocal, encoded it and still, it was not smooth....it was kinda quirky....are there other parameters that need to be taken into account other than encoding, to get a vocal to sound smooth like the one on "DA "V" CODE"??? formant maybe...?
my main questions are,,
What does encoding actually do to the waveform? and what do those 'event' points do? and how do I know when or where within the waveform to 'Add' or 'delete' one of these points? and why does 'backing' or 'ensemble' not let me use the Formant? Is it neccesary to encode all samples? How can I tell if the encoding did what I hoped ?(i know,,it would sound the way I want it to.. but....)
Sorry,, I know thats alot of questions but I'm new to encoding and I need alot of answers.
..(if one of u can)
Thank u........
Septimo ,'^)
XT
X6
my main questions are,,
What does encoding actually do to the waveform? and what do those 'event' points do? and how do I know when or where within the waveform to 'Add' or 'delete' one of these points? and why does 'backing' or 'ensemble' not let me use the Formant? Is it neccesary to encode all samples? How can I tell if the encoding did what I hoped ?(i know,,it would sound the way I want it to.. but....)
Sorry,, I know thats alot of questions but I'm new to encoding and I need alot of answers.
..(if one of u can)Thank u........
Septimo ,'^)
XT
X6
I can't resist to try to answer these questions (I hope I ge
- "before encoding a vocal, should the vocal be sung in a single tone line? or can it be melodic?"
I would say that this depends on the effect you are trying to achieve, if we sample a melodic line it will be transposed to every key we play (but the melodic relation inside the line remains unaltered)
- "encoded it as 'SOLO' and it sounded funny"
in the manual they advise that if the recording has too much reverb or delay it may nor sound as you expected, in this case re-encode in backing or ensemble
- "What does encoding actually do to the waveform?"
what I know is that an encoded wave is allways bigger than the original, so I imagine it creates information that helps V-Synth to work on them
- "what do those 'event' points do?"
event points are like indexes inside a wave, they work mainly if you use STEP or EVENT in the PCM selection, witch splits the wave exactly at those points
- "and how do I know when or where within the waveform to 'Add' or 'delete' one of these points?"
listening to the wave and according to what you intend to do with it you can ad or remove events, when done automaticaly it looks for abrupt volume changes; but one can do perfectly without events as long as we don't intend to use it
- "why does 'backing' or 'ensemble' not let me use the Formant?"
a SOLO encoding uses more memory than the others, I depreend that these last encodings don't create the information necessary for the V-Synth control the formant
- "How can I tell if the encoding did what I hoped ?"
encoding does not alter the original wave, so if you are not satisfied with any of the parameters of the encoding you have made you can allways go back and aletr it without quality loss
I hope I got this right, please correct me if I didn't
I would say that this depends on the effect you are trying to achieve, if we sample a melodic line it will be transposed to every key we play (but the melodic relation inside the line remains unaltered)
- "encoded it as 'SOLO' and it sounded funny"
in the manual they advise that if the recording has too much reverb or delay it may nor sound as you expected, in this case re-encode in backing or ensemble
- "What does encoding actually do to the waveform?"
what I know is that an encoded wave is allways bigger than the original, so I imagine it creates information that helps V-Synth to work on them
- "what do those 'event' points do?"
event points are like indexes inside a wave, they work mainly if you use STEP or EVENT in the PCM selection, witch splits the wave exactly at those points
- "and how do I know when or where within the waveform to 'Add' or 'delete' one of these points?"
listening to the wave and according to what you intend to do with it you can ad or remove events, when done automaticaly it looks for abrupt volume changes; but one can do perfectly without events as long as we don't intend to use it
- "why does 'backing' or 'ensemble' not let me use the Formant?"
a SOLO encoding uses more memory than the others, I depreend that these last encodings don't create the information necessary for the V-Synth control the formant
- "How can I tell if the encoding did what I hoped ?"
encoding does not alter the original wave, so if you are not satisfied with any of the parameters of the encoding you have made you can allways go back and aletr it without quality loss
I hope I got this right, please correct me if I didn't
-
The Sherriff
- Posts: 57
- Joined: 13:21, 12 November 2004
- Location: Tenerife,canary islands
Re: Sampling Vocals
Hi guys nice to be back
With the vocals (and more recently Sax samples) I've been working with some of them do sound very strange but I have found that if you try turning off the Vari SW switch sometimes this improves. You can't always rely on this but mess with all the buttons ( ie tempo sync, time trip sw & beat keep) some odd combinations work some don't.
The other thing to bear in mind is don't expect to record a deep bass voice and turn it into a stunning soprano. It might work with this technology but it can just as easily sound revolting.
Also mess with the formant control as sometimes you actually want less of this effect if you intend to keep the original character of the sound.
I've just been using some Sax phrases that a friend of mine recorded for me. We kept the sound as dry as possible and used a really good mic. these seem to work quite well but you have to learn to play the soundin the same way you would with a clav or synth sound. There is still a tendency for chords to sound like a dodgy brass patch on the D50 if you don't play precisely and voice the chords in the same way a sax section would.
I would suggest that if you can't get the results you need from basic sampling then get hold of the VC2 card or spend a lot more time with your recording hardware to get the best possible sound in the first place so that you have less to do to the sound in the Vsynth to make it useable.
Finally don't forget the Vsynth is not designed to just copy sounds verbatem that's what a CD, Tape, DVD, etc is for. Use your hardware/software sequencer for that use the Vsynth to make the sound and texture interesting and original.
cheers
The Sherriff
With the vocals (and more recently Sax samples) I've been working with some of them do sound very strange but I have found that if you try turning off the Vari SW switch sometimes this improves. You can't always rely on this but mess with all the buttons ( ie tempo sync, time trip sw & beat keep) some odd combinations work some don't.
The other thing to bear in mind is don't expect to record a deep bass voice and turn it into a stunning soprano. It might work with this technology but it can just as easily sound revolting.
Also mess with the formant control as sometimes you actually want less of this effect if you intend to keep the original character of the sound.
I've just been using some Sax phrases that a friend of mine recorded for me. We kept the sound as dry as possible and used a really good mic. these seem to work quite well but you have to learn to play the soundin the same way you would with a clav or synth sound. There is still a tendency for chords to sound like a dodgy brass patch on the D50 if you don't play precisely and voice the chords in the same way a sax section would.
I would suggest that if you can't get the results you need from basic sampling then get hold of the VC2 card or spend a lot more time with your recording hardware to get the best possible sound in the first place so that you have less to do to the sound in the Vsynth to make it useable.
Finally don't forget the Vsynth is not designed to just copy sounds verbatem that's what a CD, Tape, DVD, etc is for. Use your hardware/software sequencer for that use the Vsynth to make the sound and texture interesting and original.
cheers
The Sherriff
Re: Sampling Vocals
Cromio,
thanx for the carification, and your time....I understand it a little more now...It's just not stuck in me yet..but with more experience I'll get it....If not I'll come back and read your post over again....
The Sherriff,
Thanx for the advice. I'll try turnin off the Vari SW to see what I get....... but the VC2 is already in there..Its not really what I want it for, and I'm not doin any specific vocal projects on the XT, but I want to learn how the machine does it so I know in the future how to manipulate short vocal phrases, not entire songs. for knowledge sake..I don't want to abuse it, just take advantage of it....and also thank U for your time....
Septimo ,'^)
XT
X6
thanx for the carification, and your time....I understand it a little more now...It's just not stuck in me yet..but with more experience I'll get it....If not I'll come back and read your post over again....
The Sherriff,
Thanx for the advice. I'll try turnin off the Vari SW to see what I get....... but the VC2 is already in there..Its not really what I want it for, and I'm not doin any specific vocal projects on the XT, but I want to learn how the machine does it so I know in the future how to manipulate short vocal phrases, not entire songs. for knowledge sake..I don't want to abuse it, just take advantage of it....and also thank U for your time....
Septimo ,'^)
XT
X6
-
The Sherriff
- Posts: 57
- Joined: 13:21, 12 November 2004
- Location: Tenerife,canary islands
Re: Sampling Vocals
Hi septimo
I've been talking today with a guy in sound control about recording vocals in the Vsynth we were looking at the DA vcode patch and I noticed that if you turn off all the vari phase etc it sounds like the original sample has had a lot of processing done to it. I think it also sounds as though the original sample is as close to being on one note as possible. ie. monotone. You might want to bear that in mind when trying to sample phrases..
hope this helps a little more keep up the good work let us know if you get some that really work. I still miss the sample DANCE in the version 1.51 soundset that was a great sound which worked whatever pitch you played it at but was only a one word hit.
cheers
the Sherriff
I've been talking today with a guy in sound control about recording vocals in the Vsynth we were looking at the DA vcode patch and I noticed that if you turn off all the vari phase etc it sounds like the original sample has had a lot of processing done to it. I think it also sounds as though the original sample is as close to being on one note as possible. ie. monotone. You might want to bear that in mind when trying to sample phrases..
hope this helps a little more keep up the good work let us know if you get some that really work. I still miss the sample DANCE in the version 1.51 soundset that was a great sound which worked whatever pitch you played it at but was only a one word hit.
cheers
the Sherriff
Re: Sampling Vocals
Hi septimo & The Sheriff !
I also did some vocal-recording and imported the samples into the v-synth. As far as i can see the v-synth detects the pitch over time during encoding. With that information and the given user defined note-pitch of the sample (sample-view, button "i") the v-synth can calculates a pitch-difference.
If you have a mono-line at e.g. exactly A3 and you define the note pitch in sample-view with A3 the v-synth do not need to process the sample and therefore the sample sounds clean when you play an A3. (Other notes are also transformed great, cause the v-synth is now able to show its full-power)
Otherwise, if you want, that it sounds funny, you can take a vocal line with big intervals and encode it with "solo" and no Events (those vertical bars). Then some parts of the sample are heavily processed (pitch/formant) and other parts (those with the same pitch as defined for the sample) not.
c.u. johnny.
I also did some vocal-recording and imported the samples into the v-synth. As far as i can see the v-synth detects the pitch over time during encoding. With that information and the given user defined note-pitch of the sample (sample-view, button "i") the v-synth can calculates a pitch-difference.
If you have a mono-line at e.g. exactly A3 and you define the note pitch in sample-view with A3 the v-synth do not need to process the sample and therefore the sample sounds clean when you play an A3. (Other notes are also transformed great, cause the v-synth is now able to show its full-power)
Otherwise, if you want, that it sounds funny, you can take a vocal line with big intervals and encode it with "solo" and no Events (those vertical bars). Then some parts of the sample are heavily processed (pitch/formant) and other parts (those with the same pitch as defined for the sample) not.
c.u. johnny.
-
The Sherriff
- Posts: 57
- Joined: 13:21, 12 November 2004
- Location: Tenerife,canary islands
Re: Sampling Vocals
Thanks for the tip johnny I'll try that with my next lot of samples.
cheers the Sherriff
cheers the Sherriff
Cool!.....
Thank U guys for the much needed information....I'll keep all this in mind...
Septimo ,'^)
XT
X6
Septimo ,'^)
XT
X6
Re: Sampling Vocals
This should work Awsome with real instruments, (gtrs, violins, brass,etc.) by playin only one note and sampling various rythms of the note.....This'll be my next experiment..Thanx again guys....
Septimo ,'^)
XT
X6
Septimo ,'^)
XT
X6
-
Watermelon
- Posts: 138
- Joined: 10:35, 16 August 2004
"Da V Code" secrets revealed ;-)
I was looking for reasons why the "Da V Code" vocal patch sounds that good:
1. As previously stated, the original sample has been sung on one single note. This definitely eases Variphrase's job and ends up in a higher sound quality than using the "Robot Voice" function on a "melodic" sample.
2. The FORMANT KF parameter has been assigned to a positive slope (+10): this automatically assigns lower formants on lower keys and higher formants on higher keys (in other words, this matches formant and pitch across the keyboard and heps producing a more realistic voice across a wider range of keys).
3. PITCH LFO=+2, this adds some faint vibrato the original sample lacks
4. TIME LFO=+1, TIME ENV DEPTH=+1 controlled by a +32 VELOCITY SENS. This subtly changes the speed of the sample, and breaks the perfect synchronization when playing chords. When these parameters are set to zero, I noticed how the perfect key sync ends up in producing kind of unpleasant perfection (and faint "flange" artefact).
5. Anything else?
1. As previously stated, the original sample has been sung on one single note. This definitely eases Variphrase's job and ends up in a higher sound quality than using the "Robot Voice" function on a "melodic" sample.
2. The FORMANT KF parameter has been assigned to a positive slope (+10): this automatically assigns lower formants on lower keys and higher formants on higher keys (in other words, this matches formant and pitch across the keyboard and heps producing a more realistic voice across a wider range of keys).
3. PITCH LFO=+2, this adds some faint vibrato the original sample lacks
4. TIME LFO=+1, TIME ENV DEPTH=+1 controlled by a +32 VELOCITY SENS. This subtly changes the speed of the sample, and breaks the perfect synchronization when playing chords. When these parameters are set to zero, I noticed how the perfect key sync ends up in producing kind of unpleasant perfection (and faint "flange" artefact).
5. Anything else?