VSTI world hate MIDI GUITAR?
-
- KVRAF
- 13446 posts since 14 Nov, 2000 from Hannover / Germany
Glorious double post...
There are 3 kinds of people:
Those who can do maths and those who can't.
Those who can do maths and those who can't.
-
- KVRist
- 165 posts since 17 Jul, 2005
I hate to even touch this thread again, but most of the glitches I get when recording midi with my GR-20 are 1) false notes of miniscule duration in unpredictable locations and 2) a sort of double triggering where you get a little false note on top of the note you played, right at the initiation. I think most of this is from sloppy playing, fret noise, etc. The second case might be caused by too much information, I suppose.
It is easy to edit this stuff out, so it doesn't bother me too much.
John
It is easy to edit this stuff out, so it doesn't bother me too much.
John
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
2ndStringer. For case 2), are there two note-ons or one note-on followed by a rapid pitch bend to the correct note?
Also, you using Mono mode?
Case 1) is just handling noise, I'd expect.
Also, you using Mono mode?
Case 1) is just handling noise, I'd expect.
-
- KVRist
- 380 posts since 30 Aug, 2001 from Pyrlandia
AQ, you seem to take a lot of assumptions...
I'd rather assume that thinning works by passing PB message that differs more than some set value from last passed message than by passing every nth message.
and that there is some integration going on...
as far as using guitar for controlling analogue/VA synth, filtered audio fed into sync input of an oscillator can be very nice even without any tracking
I'd rather assume that thinning works by passing PB message that differs more than some set value from last passed message than by passing every nth message.
and that there is some integration going on...
as far as using guitar for controlling analogue/VA synth, filtered audio fed into sync input of an oscillator can be very nice even without any tracking
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
michu, you admit you're assuming. I was just quoting the manual (which says it's a feature to reduce the amount of data) and indeed the NAME of the feature, "Thinning", is indicative of its function as well. What you're talking about would be better called "Smoothing".
Things happen for a reason. But sure, I make assumptions too, but based on a lot of experience and research.
Lets see what the GI-20 manual has to say on the issue:
Of course, any modern sequencer, particularly a computer based host, will have no problem with any amount of controller info.
The GI-20 also has a MIDI IN port which will merge incoming CC data into its MIDI OUT port (allowing you to use other devices to add controllers) and will duplicate those CCs across all 6 channels when in MIDI Mono mode. Of course, this has the potential for increasing MIDI traffic sixfold and completely flooding the MIDI interface. So they offer a thin function for that too:
Things happen for a reason. But sure, I make assumptions too, but based on a lot of experience and research.
Lets see what the GI-20 manual has to say on the issue:
So there you go. It THINS the data, no mention of smoothing. And their rationale is the same as the GI-10 manual stated, to reduce load on your (old fashioned) sequencer.GI-20 Manual wrote:The Bend Data Thin Function
With the guitar, the finger vibrato technique can be used to
produce organic vibratos with amplitude and speed that can
be freely varied-something that's very hard to do with a
keyboard.
When creating sequencer data, the GI-20 records such finger
vibratos, sliding, and tremolo bar usage as MIDI Pitch Bend
messages. As a result of this, the MIDI performance message
that is output contains a large number of Pitch Bend
messages. In some cases, however, these Pitch Bend
messages can greatly inflate the amount of data and exceed
the sequencer's memory capacity.
The Bend Data Thin function is available to keep this to a
minimum.
The amount of Pitch Bend messages that are sent can be
reduced by using the Bend Data Thin function. Although the
smoothness of pitch changes deteriorates, this can reduce the
amount of MIDI data.
Of course, any modern sequencer, particularly a computer based host, will have no problem with any amount of controller info.
The GI-20 also has a MIDI IN port which will merge incoming CC data into its MIDI OUT port (allowing you to use other devices to add controllers) and will duplicate those CCs across all 6 channels when in MIDI Mono mode. Of course, this has the potential for increasing MIDI traffic sixfold and completely flooding the MIDI interface. So they offer a thin function for that too:
As you can see, the only logic applied to which CC messages get through, is that newer ones override any others that are in the queue of the same CC#.GI-20 Manual wrote: Receive Control Change Data Thin Function
When the GI-20's transmission mode is set to "MONO" (p.
14), the Control Change messages received at MIDI IN are
increased due to the output from all of the six strings. If a
large number of Control Change messages are being received
at the same time, it may cause data processing to stop, with
the error message (p. 37) being displayed as a result. If this
occurs, you can use the Receive Control Change Data Thin
function to thin out some of the Control Change messages
received at MIDI IN, while giving higher priority to the most
recently received messages.
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
Yet more good stuff from the GI-20 manual's Troubleshooting section:
Also note how they tell you to use Mono mode, and NOT use thinning, unless you need it. "Do not use the Bend Data Thin function when it is not needed." And, as quoted in my previous post... the only reason mentioned that you would need it is if your sequencer is low on memory. They only put this feature in so if you happen to be using an older hardware sequencer, you won't swamp it with too much data.
Sascha suggested above that they wouldn't need such a feature these days, because modern sequencers can handle it (and I agree)... but this is what us developers and manufacturers call "backward compatibility". Making new stuff that breaks old stuff is a great way to piss your customers off... so they offer this feature just to satisfy people who are still using OLDER HARDWARE SEQUENCERS. That's the only reason.
All that other stuff I mentioned about possible MIDI cable congestion I'm sure is mitigated just by prioritizing note events over CC data, and throwing out all but the latest outgoing CC of a particular number from the MIDI queue. This isn't magic or rare stuff, these are WELL KNOWN, COMMON techniques that manufacturers have been using almost universally since the late 80's. It's just the smart way to go about it. And frankly, I didn't even need to read the manual to know that this is how it worked. It's my JOB to know about this stuff. I've been immersed in MIDI since 1983. Not just as a user, but from a programming perspective as well. To me, there aren't any mysteries going on in there, and I can't believe all the flak I'm getting just trying to straighten out some common misconceptions. You're all so SURE of your mythology... but there's a lot of this in the guitar world actually. I shouldn't be surprised. (Let's start talking about amps and effects and THEN watch the fur fly!
)
But whatever you think works for you... do it. But I'm simply trying to explain HOW it works.
So, one thing that can cause (at least on this device) retriggering, is if your bend range is set TOO LOW! If you bend (or due to a glitch it thinks you bent) past the set range, a new note will have to be triggered. This is a lot nastier than just a quick pitch bend that's instantly corrected by the next PB message. (Also you could use a MIDI CC smoothing function in your host to remove sudden jumps caused by a spurious PB message. If there isn't a product that does this already, I'll make you one.)Pitch doesn't change smoothly
❏ Is "BEND RANGE" (p. 23) set to "Chromatic Type 1-
3"?
→ Set BEND RANGE to anything other than Chromatic
Type for patches that require continuous changes in
pitch. (p. 23)
❏ Is data being sent in the Poly mode (p. 14) ?
→ When transmitting in the Poly mode, pitch changes
during chord performances are in semitone steps. Play
with single notes, or use the Mono mode to send data.
❏ Is the Bend Data Thin function (p. 31) in use?
→ Using this function may result in a slight loss of
smoothness when pitch changes, even during normal
play. Do not use the Bend Data Thin function when it is
not needed.
❏ Is BEND RANGE MAX (p. 32) or BEND RANGE (p. 23)
set to anything other than 24?
→ With BEND RANGE MAX set to anything other than 24,
if Glide Up/Down 5-8 or Pitch Bend 3 or 6 is used in
Assign (p. 19), then changing the pitch by an octave
causes retriggering.
Set BEND RANGE MAX and BEND RANGE to 24.
❏ Is "BEND RANGE" (p. 23) set to "1" or "2"?
→ When matching with sending and receiving, use as large
a value as possible.
Also note how they tell you to use Mono mode, and NOT use thinning, unless you need it. "Do not use the Bend Data Thin function when it is not needed." And, as quoted in my previous post... the only reason mentioned that you would need it is if your sequencer is low on memory. They only put this feature in so if you happen to be using an older hardware sequencer, you won't swamp it with too much data.
Sascha suggested above that they wouldn't need such a feature these days, because modern sequencers can handle it (and I agree)... but this is what us developers and manufacturers call "backward compatibility". Making new stuff that breaks old stuff is a great way to piss your customers off... so they offer this feature just to satisfy people who are still using OLDER HARDWARE SEQUENCERS. That's the only reason.
All that other stuff I mentioned about possible MIDI cable congestion I'm sure is mitigated just by prioritizing note events over CC data, and throwing out all but the latest outgoing CC of a particular number from the MIDI queue. This isn't magic or rare stuff, these are WELL KNOWN, COMMON techniques that manufacturers have been using almost universally since the late 80's. It's just the smart way to go about it. And frankly, I didn't even need to read the manual to know that this is how it worked. It's my JOB to know about this stuff. I've been immersed in MIDI since 1983. Not just as a user, but from a programming perspective as well. To me, there aren't any mysteries going on in there, and I can't believe all the flak I'm getting just trying to straighten out some common misconceptions. You're all so SURE of your mythology... but there's a lot of this in the guitar world actually. I shouldn't be surprised. (Let's start talking about amps and effects and THEN watch the fur fly!
But whatever you think works for you... do it. But I'm simply trying to explain HOW it works.
-
- KVRist
- Topic Starter
- 448 posts since 18 Nov, 2002
Hi that sounds interesting, have you tried it?, it works really that well?guitarzan wrote:Or...f**k all this guitar to MIDI shit altogether and use something that works like FrettedSynth's pitch to "CV" synths for mono lines. They track more expressively than pitch to MIDI ever will.
For polyphony, maybe run a hex pickup into one of those breakout boxes then into a 6 channel audio interface - then run the 6 channnels of audio into something like a multi-mono version pitch to "CV" synth, or 6 FrettedSynths in parallel if the system could handle it.
I really doubt MIDI is even capable of tracking the subtle dynamics and bends that the FrettedSynth pitch to CV type method can and does.
Just trying to think out of the box
I will try for sure, but some opinions are ever welcome.
- KVRAF
- 2342 posts since 3 Sep, 2005 from Outer Bongolia
All the Fretted Synth Audio Triggered Synths work great in mono - and they do track more expressively than any midi guitar stuff, and you don't need any special pickup or equipment. They're all free, go ahead and try them all if you haven't already. They do track best above the 4th fret or so, but I think that could be perfected, in fact FrettedSynth just said "expect an update to the bass synth with some real smooth tracking of the filter cutoff".krraqk wrote:Hi that sounds interesting, have you tried it?, it works really that well?guitarzan wrote:Or...f**k all this guitar to MIDI shit altogether and use something that works like FrettedSynth's pitch to "CV" synths for mono lines. They track more expressively than pitch to MIDI ever will.
For polyphony, maybe run a hex pickup into one of those breakout boxes then into a 6 channel audio interface - then run the 6 channnels of audio into something like a multi-mono version pitch to "CV" synth, or 6 FrettedSynths in parallel if the system could handle it.
I really doubt MIDI is even capable of tracking the subtle dynamics and bends that the FrettedSynth pitch to CV type method can and does.
Just trying to think out of the box
I will try for sure, but some opinions are ever welcome.
I have not tried the deal with 6 in parallel yet (the only Fanout box I knew of before the DIY one refered to in this thread cost > $300) but I see no reason it shouldn't work perfectly IF you build or buy a fanout box for the hex pickup AND you have a 6 channel audio interface AND your system can handle all 6 channels plus 6 instances of FS Audio Triggered Synths. I'm not sure my system could. I'm sure a more CPU friendly 6 channel multi-mono pitch to CV synth could be developed, but it really wouldn't get used by enough people unless someone can come up with an all-in-one interface with a 13-pin Roland hex pickup input and 6 channels out (say, firewire or USB).
OR - there is apparently a way to do polyphonic guitar synth without even using a MIDI pickup - the Electro Harmonix HOG does it, and if anyone could figure out how to do that in VST form then that would be the future right there!!!
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
Guys, those FrettedSynth "products" are made with SynthEdit. I REALLY doubt they've managed to do a better job of pitch to "CV" (there is of course no such thing as a "control voltage" in the digital domain) than the dedicated Roland hardware. For one thing, you are guarranteed probably at least 3 or 4 ms latency on any input into your DAW. So you're already behind there. The dedicateed hardware unit will be able to do much better than that. Latency is one of the few areas where I will admit dedicated hardware devices can beat software... aside from that digital hardware RUNS software, so they're both the same, HOWEVER...krraqk wrote:Hi that sounds interesting, have you tried it?, it works really that well?guitarzan wrote:Or...f**k all this guitar to MIDI shit altogether and use something that works like FrettedSynth's pitch to "CV" synths for mono lines. They track more expressively than pitch to MIDI ever will.
For polyphony, maybe run a hex pickup into one of those breakout boxes then into a 6 channel audio interface - then run the 6 channnels of audio into something like a multi-mono version pitch to "CV" synth, or 6 FrettedSynths in parallel if the system could handle it.
I really doubt MIDI is even capable of tracking the subtle dynamics and bends that the FrettedSynth pitch to CV type method can and does.
Just trying to think out of the box
I will try for sure, but some opinions are ever welcome.
We're comparing apples to oranges here. There's no MIDI Out from their product, they just morph the monophonic guitar performance into a new waveform and then filter it and envelope track it; it's more of an effect than a controller. You're stuck with their sounds, and their sounds only. They specify to use the neck pickup because they need the roundest waveform possible otherwise there's too many harmonics and they can't filter/distort the incoming signal to a square wave (from which they derive their other waveforms)
(This is actually probably not unlike some of the stuff the roland VG's do! Of course, the VGs can do it for each string, polyphonically. And their pickup goes as close to the bridge as possible, as they have smarter fundamental frequency determination logic.)
Thanks for pointing them out though.
OK, I'm actually trying them out now... latency is even worse than it needs to me (I have 3 ms latency on my DAW). Delay is probably because it uses a pitch shifter to get the different intervals. Mostly sounds just like I'm playing a rectifier (Octavia) through an envelope follower like my own SCAMP. But it sounds worse because it's SynthEdit with unspectacular filters, etc.
Not impressed, sorry. And this isn't MIDI-guitar. It's a big fat effect. I can't seem to get a clean undistorted sound out of it either. Even when I feed it sine waves (playing harmonics on the 12th fret.)
It's what it is though, sure it has some use. But you get what you pay for. And it certainly ain't "MIDI-guitar". Not even for monophonic leads.
- KVRAF
- 2342 posts since 3 Sep, 2005 from Outer Bongolia
No, it's not MIDI guitar, that's why it works. They have oscillators- they are synths, not effects. They track FAR smoother than MIDI can and with much less latency. MIDI guitar sucks, and that is why they have never caught on in a big way. It can be useful, but you can't play it like a real instrument - your walking on egg shells all the time!
Last edited by guitarzan on Thu Sep 14, 2006 12:20 am, edited 1 time in total.
-
- KVRAF
- 13446 posts since 14 Nov, 2000 from Hannover / Germany
Something easily reached by a lot of hardware devices as well - of course, 3-4ms of latency are quite optimistic when it comes to overall latency (input to output), realistic figures are more in the 15-20ms range (doubled buffers for I/O, some "safety" buffers usually not reported by whatever DAW, probably even some AD/DA converter latency, probably some plugin latency). So yes, hardware often is more responsive.AdmiralQuality wrote: For one thing, you are guarranteed probably at least 3 or 4 ms latency on any input into your DAW.
Anyways, as all guitar to MIDI converters still need at least two string amplitudes to analyze the pitch properly, MIDI guitarists are quite f**ked as there's even more latency happening (even if, on the positive side of things, you're just used to latency due to that...)
True. But when you listen to some patches of the VG-8 or 88, it's still amazing. The downside obviously being that it's audio data rather than easily tweakable MIDI data. Plus, who really wants to deal with six channels of audio data at once (in case you're doing it inside a computer)?We're comparing apples to oranges here. There's no MIDI Out from their product, they just morph the monophonic guitar performance into a new waveform and then filter it and envelope track it; it's more of an effect than a controller. You're stuck with their sounds, and their sounds only. They specify to use the neck pickup because they need the roundest waveform possible otherwise there's too many harmonics and they can't filter/distort the incoming signal to a square wave (from which they derive their other waveforms)
Yeah well, from what I know, the VGs don't even use pitch to CV but deal with the individual pickup signals straight ahead (so it's indeed just a 6x FX unit).(This is actually probably not unlike some of the stuff the roland VG's do! Of course, the VGs can do it for each string, polyphonically. And their pickup goes as close to the bridge as possible, as they have smarter fundamental frequency determination logic.)
As far as the Fretted Synth thing goes, while his FreeAmp is amazing for the money, I haven't found much use for the guitar "synth" thing either. I seem to be getting more inspirational sounds out of whatever FX chains - which even play better as they don't have to deal with pitch to CV.
There are 3 kinds of people:
Those who can do maths and those who can't.
Those who can do maths and those who can't.
-
- KVRAF
- 13446 posts since 14 Nov, 2000 from Hannover / Germany
Hm, I actually *have* to agree with AQ on the Fretted Synth part (as much as I'd prefer not having to...guitarzan wrote:No, it's not MIDI guitar, that's why it works. It has oscillators- they are synths, not effectst. It tracks FAR smoother than MIDI can and with much less latency
There are 3 kinds of people:
Those who can do maths and those who can't.
Those who can do maths and those who can't.
-
- KVRist
- Topic Starter
- 448 posts since 18 Nov, 2002
Just tried it and I like it. Translates surprisingly well the playing nuances. It resembles the VG synth sounds, now Im sure Roland can do a much better thing than the VG one. Not perfect, some sounds work better than others, but there is a lot of less glitches than with guitar midi converters. You are stuck to the internal sounds, of course, that's why I would like to see more development with this technique it has a great potential.
INTERESTING
I will keep trying it.
pd: Thanks to point it guitarzan.
pd2: this technique is not similar to the one used by Roland with the GR300? The sound horn that Metheny uses with this synth is amazing and tracks almost perfect many guitarist techniques.
INTERESTING
I will keep trying it.
pd: Thanks to point it guitarzan.
pd2: this technique is not similar to the one used by Roland with the GR300? The sound horn that Metheny uses with this synth is amazing and tracks almost perfect many guitarist techniques.
- KVRAF
- 2342 posts since 3 Sep, 2005 from Outer Bongolia
No problem! Glad you like themkrraqk wrote: pd: Thanks to point it guitarzan.
Yes...I think the GR300 was pre MIDI and used analog pitch to CV, that's why the GR300 is still considered the best tracking guitar synth ever by many. The FrettedSynth ones use virtual pitch to CV.krraqk wrote:pd2: this technique is not similar to the one used by Roland with the GR300? The sound horn that Metheny uses with this synth is amazing and tracks almost perfect many guitarist techniques.
-
AdmiralQuality AdmiralQuality https://www.kvraudio.com/forum/memberlist.php?mode=viewprofile&u=83902
- Banned
- 6657 posts since 10 Oct, 2005 from Toronto, Canada
Heh, I actually *have* to agree with Sascha too. The VGs are probably great and I'll bet (though I haven't played them) MUCH more responsive than the FrettedSynth stuff.
OH and please note, the FrettedSynth things AREN'T tracking. They're filtering, distorting, and pitch shifting the ORIGINAL WAVEFORM. Then applying envelopes (basically triggered envelope "followers". All stuff you could do with regular guitar effects! I don't have it, but I bet you could do most of what they do with G.R. They have octavers, pitch shifters, envelope followers, don't they? Sure, it might not be as integrated and "to the point", but again, don't be fooled, they're NOT pitch tracking. There is no "CV" stage in them, the output wave is just a distortion of the input wave, and then pitch shifted to get whatever interval you select (which again is the process that I think is introducing the bad latency I hear. Without the pitch shift this technique SHOULD be very responsive!
)
Anyway, just a guess after 5 minutes of playing with it, but... maybe we can get the FrettedSynth author in here to tell us how it works?
Still, its hard to complain about something that's free. So more power to em! And sorry for putting it down. Like I said, there's a use for EVERYTHING. I just tend to be REALLY FUSSY when it comes to sound, so don't take the fact that I'm not impressed too badly. I'm RARELY impressed.
Cheers everyone.
OH and please note, the FrettedSynth things AREN'T tracking. They're filtering, distorting, and pitch shifting the ORIGINAL WAVEFORM. Then applying envelopes (basically triggered envelope "followers". All stuff you could do with regular guitar effects! I don't have it, but I bet you could do most of what they do with G.R. They have octavers, pitch shifters, envelope followers, don't they? Sure, it might not be as integrated and "to the point", but again, don't be fooled, they're NOT pitch tracking. There is no "CV" stage in them, the output wave is just a distortion of the input wave, and then pitch shifted to get whatever interval you select (which again is the process that I think is introducing the bad latency I hear. Without the pitch shift this technique SHOULD be very responsive!
Anyway, just a guess after 5 minutes of playing with it, but... maybe we can get the FrettedSynth author in here to tell us how it works?
Still, its hard to complain about something that's free. So more power to em! And sorry for putting it down. Like I said, there's a use for EVERYTHING. I just tend to be REALLY FUSSY when it comes to sound, so don't take the fact that I'm not impressed too badly. I'm RARELY impressed.
Cheers everyone.

