Proper SY1000 Speed Test

Started by germanicus, February 07, 2020, 10:13:43 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

germanicus

This weekend I have some time on my hands and thought it would be worthwhile to do a proper speed test.
Test not only the Pitch To Midi performance, but the latency associated with the various synths.

Looking for feedback/recommendations on this process.

I plan to use my Godin LGXT into the SY1000.
I will connect the SY1000 via USB (for Midi) and also take feeds out the Main and Sub Outputs, and the 5 pin output.

Measure the following:
Pitch to Midi (over USB)
Pitch to Midi (from 5 pin)
Dynamic Synth
OSC Synth
Gr300
Regular guitar through the SY1000 (routed through an empty "Normal" guitar chain).
Regular guitar output directly into my Focusrite interface.

I should be able to record 4 audio signal outputs from the SY1000 (each hard panned through one side of the Sub and Main outs), plus the two midi outputs (USB/5 ping), plus the output directly from my guitar simultaneously.

For each synth I will use a tone with a fast attack envelope.

Will record single note hits on each string. Then compare the latency between the regular guitar outputs against the synths and Midi outputs.

Tips?
My albums done with modeling/guitar synth at http://music.steamtheory.com

JTV69/59P/Godin LGXT/Multiac ACS/Variax 700 AC
Helix/FTP/GP10/VG99/SY1000
Traynor k4

admin

#1
Quote from: germanicus on February 07, 2020, 10:13:43 AM
This weekend I have some time on my hands and thought it would be worthwhile to do a proper speed test.
Test not only the Pitch To Midi performance, but the latency associated with the various synths.

Looking for feedback/recommendations on this process.

I plan to use my Godin LGXT into the SY1000.
I will connect the SY1000 via USB (for Midi) and also take feeds out the Main and Sub Outputs, and the 5 pin output.

Measure the following:
Pitch to Midi (over USB)
Pitch to Midi (from 5 pin)
Dynamic Synth
OSC Synth
Gr300
Regular guitar through the SY1000 (routed through an empty "Normal" guitar chain).
Regular guitar output directly into my Focusrite interface.

I should be able to record 4 audio signal outputs from the SY1000 (each hard panned through one side of the Sub and Main outs), plus the two midi outputs (USB/5 ping), plus the output directly from my guitar simultaneously.

For each synth I will use a tone with a fast attack envelope.

Will record single note hits on each string. Then compare the latency between the regular guitar outputs against the synths and Midi outputs.

Tips?

1 ) I use a separate stereo audio recorder. I like the Zoom H4N, or an old Iphone4 with Fostex AR4i Audio interface and an IOS audio recording app like Refire or White recorder.


2 ) Left channel is Audio output from a headstock mounted piezo clip on pickup typically sold as an accessory for Peterson Strobe tuners. This provides immediate response from your guitar.

Left Channel audio will be used as the Master timing reference.

Peterson TP-3 Tuning Pickup
http://www.amazon.com/Peterson-TP-3-Tuning-Pickup/dp/B000NKKHVE/ref=sr_1_2?ie=UTF8&qid=1418052095&sr=8-2&keywords=peterson+tuner+clip






3 ) Right channel is final output audio from your "unit under test" ( DAW sound card output , Guitar FX output , etc.

4 ) After making the 1 minute recording of your playing, load resulting Stereo Wave file in WaveLab or Soundforge and zoom in and look at the Left Channel. Use this Left Channel audio (from the headstock piezo pickup) as a reference to measure the Right Channel's  "unit under test" millisecond(s) offset delay. 
This is your latency.
The above strategy can be employed to measure any latency of any DSP guitar gadget or full rig.


** Note if you want higher accuracy  - replace the stereo audio recorder from step #1 with a dual trace storage oscilloscope

--

Measurement Tool


https://www.usb-audio.com/latenc-o-meter/
https://web.archive.org/web/20160531001447/https://www.usb-audio.com/latenc-o-meter/

Ploytec Latenc-o-meter

Latency Tester / Delay Meter


Pricing: Latenc-o-meter EUR 199.00 including shipment.


Here you can order the Ploytec Latenc-o-meter, a unique tool for finding the truth about latency.

The unit generates a "click" signal on its analog output (the red cinch connector) and waits until it returns on the input (the black cinch connector).

In the meantime it samples that input on its A/D converter running at 100kHz samplerate in order to get this amazing accuracy: The measurement range goes from 0.01ms to 999.99ms!

The red button switches it on while pressed, new clicks are generated every few seconds. It runs from a 9V battery which is easy to replace (but should last for ages depending on your application).

We use it for our development of external USB soundcards and drivers, but there's tons of further applications including checking your studio setup, testing digital audio equipment and last not least VoIP.

Pricing: EUR 199.00 including shipment. If you order from a non-European Union country there might be additional taxes charged by your national customs.

In the EU, you have the legal right to cancel your order (and return the unit) anytime within 14 days after ordering.

Estimated delivery time: Typically we ship on the day you order, max. 3 days. Your credit card will be charged on the day we ship the unit.


===


https://www.thegearpage.net/board/index.php?posts/26499772/

Quote from: Sascha Franck, post: 26499742, member: 187257You don't need any specific measurement tools to find out about your latencies. You can even do it with the same system and interface that you're measuring (you do however need two inputs).
And yet I never get accurate results with that method.

and many folks on forums say

QuoteI have only 1.09ms round trip latency on my Windows DAW!

and I say B.S.

https://web.archive.org/web/20170506174453/https://www.presonus.com/learn/technical-articles/The-Truth-About-Digital-Audio-Latency



Details:

The Truth About Digital Audio Latency
By Wesley Elianna Smith

https://web.archive.org/web/20170506174453/https://www.presonus.com/learn/technical-articles/The-Truth-About-Digital-Audio-Latency

https://www.vguitarforums.com/smf/index.php?topic=12835.msg134782#msg134782


Today I only trust using my independent third party hardware tool Ploy-Tec Latence-o-meter, and I'm gathering the data



===

https://digitech.com/en/products/the-drop






here's a handy reference to learn how to ear training and listening identify milliseconds delays



fokof

Can't wait to try it too !!  (with a bass)

The lack of delay @ the guitar input worries me.
My guess is that it's the reason why they put a phase switch , but it's NEVER a 180° phase thing but always a delay thing.

We'll probably have to use a delay block like I do with the VB99 .......
(with are 1ms increment , sadly)


germanicus

Some Results:
USB Midi, used the sub and main outs into my Focusrite 18i20 interface. Cakewalk by Bandlab DAW.

I found lots of variability on the low E latency. One outlier triggered at 30 ms, rest were higher.
Something consistent however, the OSC synth was consistency only 2-3 ms faster than the Midi Note.

Low E String.
Time between Normal Guitar and Midi triggered note: 48 ms
Time between Normal Guitar and Gr300: 4 ms
Time between Normal Guitar and Dynamic Synth: 3 ms
Time between Normal Guitar and OSC synth: 46 ms

High E string:
Time between Normal Guitar and Midi triggered note: 25 ms
Time between Normal Guitar and Gr300: 4 ms
Time between Normal Guitar and Dynamic Synth: 5 ms
Time between Normal Guitar and OSC synth: 21 ms

My albums done with modeling/guitar synth at http://music.steamtheory.com

JTV69/59P/Godin LGXT/Multiac ACS/Variax 700 AC
Helix/FTP/GP10/VG99/SY1000
Traynor k4

admin

#4
QuoteUSB Midi, used the sub and main outs into my Focusrite 18i20 interface. Cakewalk by Bandlab DAW.

Question
Since the SY-1000 is the unit under test for Guitar to MIDI  - explain the actual SY-1000 complete Guitar to MIDI signal chain that was used for your testing?

Historically Roland/Boss products yield lowest Guitar to MIDI Latency using the Roland/Boss Core Audio  / MIDI OSX/macOS USB Driver and a USB 2.0 MIDI cable between SY-1000 USB port and Mac Pro USB Input, driving AU Instruments in Apple Mainstage  / Garageband / Logic-X

A bit higher latency  would be  the Guitar to MIDI Latency using the SY-1000 with Boss ASIO Audio / MIDI USB Driver for Windows and a USB 2.0 MIDI cable between SY-1000 USB and Windows USB 2.0 Input, driving VST Instruments in NanoHost VST Host 
https://www.tone2.com/nanohost.html


Still higher latency  would be  the Guitar to MIDI Latency using the SY-1000 5 pin MIDI Out into a fast triggering 5 pin MIDI Tone module
- Alesis QS6/QS7/QS8

And Not all MIDI Tone modules respond with the same speed

jassy

Quote from: germanicus on February 07, 2020, 07:51:52 PM
Some Results:
USB Midi, used the sub and main outs into my Focusrite 18i20 interface. Cakewalk by Bandlab DAW.

I found lots of variability on the low E latency. One outlier triggered at 30 ms, rest were higher.
Something consistent however, the OSC synth was consistency only 2-3 ms faster than the Midi Note.

Low E String.
Time between Normal Guitar and Midi triggered note: 48 ms
Time between Normal Guitar and Gr300: 4 ms
Time between Normal Guitar and Dynamic Synth: 3 ms
Time between Normal Guitar and OSC synth: 46 ms

High E string:
Time between Normal Guitar and Midi triggered note: 25 ms
Time between Normal Guitar and Gr300: 4 ms
Time between Normal Guitar and Dynamic Synth: 5 ms
Time between Normal Guitar and OSC synth: 21 ms

The real latency is an important part of the equation but not the most important I think.
The most important to me is consistency, realiability and predictability.
Now that you have done those tests, can you comment about those aspects in the OSC and midi out?

There is a quote from some similar tests with MG2:
https://www.vguitarforums.com/smf/index.php?topic=7530.msg180828;topicseen#msg180828

QuotePitch to Midi Conversion Times for Midi Guitar 2
(using Midi Guitar 2 as plugin within Reaper DAW)

Polyphonic Setting
E4: 20 ms
B3: 25 ms
G3: 26 ms
D3: 23 ms
A2: 36 ms
E2: 32 ms
Average: 27 ms

Monophonic Setting
E4: 10 ms
B3: 14 ms
G3: 20 ms
D3: 20 ms
A2: 26 ms
E2: 30 ms
Average: 20 ms
MG2 is doing very well compared to the SY1000 and the most important its very realiable while playing.

arkieboy

Quote from: jassy on February 08, 2020, 01:57:11 AM
The real latency is an important part of the equation but not the most important I think.
The most important to me is consistency, realiability and predictability.


The slowest tracking Roland unit was the GM70.  But with a G707 it was the only one I felt comfortable attacking anything like a proper piano part.  I found that combination to be really consistent such that with good monitoring I could play around the tracking delay
Main rig: Barden Hexacaster and Brian Moore i2.13 controllers
Boss SY1000/Boss GKC-AD/Boss GM-800/Laney LFR112

Other relevant gear: Line 6 Helix LT, Roland GR-33, Axon AX100 MkII
Oberheim Matrix 6R, Supernova IIR, EMu E5000, Apple Mainstage, Apple Logic, MOTU M4

fokof

Quote from: germanicus on February 07, 2020, 07:51:52 PM

Low E String.
....
Time between Normal Guitar and Gr300: 4 ms
Time between Normal Guitar and Dynamic Synth: 3 ms
Time between Normal Guitar and OSC synth: 46 ms

Thanx for that  , it confirms what I was thinking : the lack of adjustable delay at the input of "normal gtr/bass " will be problematic.
They should REALLY think of implementing it with future firmware update.

With a bass , blending the normal signal + a synth with a 3ms delay will be Über problematic    :(



germanicus

#8
Here is another screen shot.
In this example I am recording the USB and 5 pin midi simultaneously, as well as the output from my seymour duncan pickups in the Godin LGXT straight from the "Electric" jack It is the top waveform in the picture. The RMC polydrive piezo signal is being routed through the 13 pin and recorded on the second track. Notice there is a VERY small delay between the start of the picking transient. Less than half a millisecond.

The two midi notes have a VERY small delay as well. The 5 pin taking a half millisecond or so longer. HOWEVER. There were other cases where they were identical. I cant account for the variation. In most cases they were identical.

Once again the OSC synth is almost as slow as the Midi notes.
The GR300 and Dynamic synths have roughly 3-4 ms latency compared to the Piezo and Seymour Duncan signals.

This demonstrates the relative latency, not the ultimate one can get. In this example the sy1000 is not being used to route audio, the DAW only sees the USB as a Midi device (my Focusrite is the audio interface).
My albums done with modeling/guitar synth at http://music.steamtheory.com

JTV69/59P/Godin LGXT/Multiac ACS/Variax 700 AC
Helix/FTP/GP10/VG99/SY1000
Traynor k4

germanicus

#9
Quote from:  Bluesbird on February 08, 2020, 06:51:51 AM
Also, the normal guitar sound took some time to get into the daw, and the midi note after triggering a soft synth will take some time in getting to the output. So the actual triggering time of the midi note will have to add in the input latency of the normal guitar signal (around 2 to 3 ms in the least).  Also the output latency once the synth is triggered has to be added in. So the total latency will be longer than the already dismal results you have posted. I found the SY-1000 to have high output latency using it as an interface (7 ms reported output latency at lowest buffer setting at 48000 sample rate.  Also, with the SY-1000 pitch-to-midi, I found that it dropped notes from chords and had delayed firings. The GM70, according to Wayne Joness's midi speed test had a latency of 25 ms on the high e and 40 ms on the low E.  These results from Germanicus suggest that the SY-1000 has the distinction of being the worst performing pitch-to-midi of any prior Roland unit.

Well there are many variables. I would not take this as a bellwether for one's best possible Pitch to Midi performance with this unit.
This was mainly done to see the variation in responsiveness between the various synths, particularly the OSC synth. I think this a relatively decent test for THAT comparison, but not best possible PTM.

My Daw is set up for stability for mixing large projects, not lowest possible latency.
Its not super cutting edge, something I built myself. Core i7-6700k running @4.00GHZ, 32 gigs ram, Windows 10.

128 samples, with a reported:
Input: 5.5msec, 242 samples

However, I am using the recorded raw regular guitar pickup as the starting point. How the buffer setting impacts the MIDI input is unknown to me. The buffer is for audio, not midi data.

I recorded the Midi over USB and 5 pin into the Focusrite.

For comparison sake, I just recorded the Tripleplay. It reports 10 ms for the High E string.
I recorded magnetic pickup from my JTV69 direct to the focusrite, and used the wireless tripleplay receiver.

The SY1000 is 9 milliseconds slower on the High E string.
In the attached pic the top tracks are from the SY1000, the bottom two the Tripleplay test. I drag adjusted the tripleplay test tracks to show how they compare.

The Tripleplay take 18-20 msecs on the Low E. That is an astounding difference compared to Roland/Boss.

My albums done with modeling/guitar synth at http://music.steamtheory.com

JTV69/59P/Godin LGXT/Multiac ACS/Variax 700 AC
Helix/FTP/GP10/VG99/SY1000
Traynor k4

szilard

Thanks for the measurements, I have a few more days to wait...

susbemol

This is very much in line with the measurements I performed a while back. In essence, it always seemed to me that there has been no guitar to MIDI improvement of any note in Roland/BOSS products over recent years. In fact, my tests showed that the GI-20 tracked considerably better (less glitches and more predictably) than the GR-55 and VG-99 that I owned at the time and the GP-10 I still own.

My TriplePlay setup (using the exact same GK3 pickup and my home made GK to TriplePlay adapter) performs better than those in all aspects.

germanicus

Quote from:  Bluesbird on February 08, 2020, 09:19:08 AM
Forgive my ignorance, but does the osc synth incorporate pitch recognition? If so, is the same pitch recognition algorithm tied to the pitch-to-midi conversion process? Could it be that in improving the performance of the OSC synth (perhaps by using more recognition cycles as is done in the fingerstyle mode of the Fishman Tripleplay) that this has the side effect of degrading the pitch-to-midi performance?

The OSC synth incorporates Pitch Detection. From the Reference Manual:
This is a synthesizer that detects the pitch and attack
information in the input sound and outputs signals
produced by the built-in oscillator.

In my tests, it was consistently only a few ms faster than the pitch to Midi output.

Its a triggered oscillator.

I don't think Roland has done anything to improve their PTM performance.

The Tripleplay is in another league.

I still very much like the sy1000. Great synth sounds, guitar modeling, alternate tuning, and amp models.
But if you want PTM, get the Fishman.


My albums done with modeling/guitar synth at http://music.steamtheory.com

JTV69/59P/Godin LGXT/Multiac ACS/Variax 700 AC
Helix/FTP/GP10/VG99/SY1000
Traynor k4

GuitarBuilder

Quote from: germanicus on February 08, 2020, 09:55:51 AM
The Tripleplay is in another league.

I still very much like the sy1000. Great synth sounds, guitar modeling, alternate tuning, and amp models.
But if you want PTM, get the Fishman.

Amen!
"There's no-one left alive, it must be a draw"  Peter Gabriel 1973

jassy

Quote from: germanicus on February 08, 2020, 09:55:51 AM
The OSC synth incorporates Pitch Detection. From the Reference Manual:
This is a synthesizer that detects the pitch and attack
information in the input sound and outputs signals
produced by the built-in oscillator.

In my tests, it was consistently only a few ms faster than the pitch to Midi output.

Its a triggered oscillator.

I don't think Roland has done anything to improve their PTM performance.

The Tripleplay is in another league.

I still very much like the sy1000. Great synth sounds, guitar modeling, alternate tuning, and amp models.
But if you want PTM, get the Fishman.
Sorry to ask again, but would like to know your opinion, the latency question is clear, but what about consistency and realibility in the OSC synth tracking while playing? same as the PTM? worse like the GP10 OSC was?

germanicus

Quote from: jassy on February 08, 2020, 10:06:19 AM
Sorry to ask again, but would like to know your opinion, the latency question is clear, but what about consistency and realibility in the OSC synth tracking while playing? same as the PTM? worse like the GP10 OSC was?
I never used the GP10 OSC, so cant compare.

The OSC Synth in the SY1000 seems decent, but you have to play to its latency. Its not perfect. But you can certainly get musical results out of it. If you double the sound with a clean guitar sound you notice the latency more.

Its not a jumbled mess when you play it. You can tweak its "low velocity cut" setting, and this goes a long way to adjusting it for your needs. Getting it so that small very light touches do not trigger the oscillator and keeps it more "clean".

Im glad its in there, and will probably use it.
The question is would it be better using a tripleplay to trigger similar soft synths? Yeah probably.

But you also have the GR300, Poly FX, Vio, and Dynamic synths which are different beasts.
My albums done with modeling/guitar synth at http://music.steamtheory.com

JTV69/59P/Godin LGXT/Multiac ACS/Variax 700 AC
Helix/FTP/GP10/VG99/SY1000
Traynor k4

AlakaLazlo

I really appreciate the work you did to provide this info. 
I'm curious if it would be possible to compare the time delay between the raw audio (directly off the guitar) and the midi data (off either 5 pin or usb) - as opposed to the sound generated from the midi applied to an audio generator?
Hexstainocaster, Fender Strat and Electric XII, Godin ACS, Axon AX50 - Moog One, IIP and Mini, SEM, Dot.Com/Moon/STG/FSFX 110, Cubase Pro, 2xMR816, HR824, NS10M, Komplete, Omnisphere/RMX/Trilian, Z3ta+, Analog Lab, Slate MTi2, ML1 and Everything Bundle, Social Entropy Engine, ESQ1, DX7, Lavalamp.

germanicus

Quote from: AlakaLazlo on February 08, 2020, 11:00:28 AM
I really appreciate the work you did to provide this info. 
I'm curious if it would be possible to compare the time delay between the raw audio (directly off the guitar) and the midi data (off either 5 pin or usb) - as opposed to the sound generated from the midi applied to an audio generator?

I did that.
I compared the recorded raw signal from the guitar quarter inch out jack and compared that against the midi note recorded from both the USB and 5 pin. The midi note was not triggering a soft synth.
My albums done with modeling/guitar synth at http://music.steamtheory.com

JTV69/59P/Godin LGXT/Multiac ACS/Variax 700 AC
Helix/FTP/GP10/VG99/SY1000
Traynor k4

jassy

Thanks booth for your opinion.
It is been obvious that the PTM tech from Roland is old and has not been improved over the years, sadly. The OSC is dependant of this tech, so same PTM limitations, I said that previously, it was at least to me, logical.
Anyway I agree It's good to have it inside, can be a good compliment to have more colors and when combined with the other synth the problems can be minimized, same thing that in the GR55 with the PCM tones.

alexmcginness

Quote from: susbemol on February 08, 2020, 09:20:37 AM
My TriplePlay setup (using the exact same GK3 pickup and my home made GK to TriplePlay adapter) ....

Do you have plans for that? All of the other info here on the form has it the other way around where the triple play hex is driving the roland 13 pin out. I want to do what youve done and have the GK pickup drive Triple play. Thanks.
VG-88V2, GR-50, GR-55, 4 X VG-99s,2 X FC-300,  2 X GP-10 AXON AX 100 MKII, FISHMAN TRIPLE PLAY,MIDX-10, MIDX-20, AVID 11 RACK, BEHRINGER FCB 1010, LIVID GUITAR WING, ROLAND US-20, 3 X GUYATONE TO-2. MARSHALL BLUESBREAKER, SERBIAN ELIMINATOR AMP. GR-33.

admin

#20
Quote from: alexmcginness on February 08, 2020, 12:05:04 PM
Do you have plans for that? All of the other info here on the form has it the other way around where the triple play hex is driving the roland 13 pin out. I want to do what youve done and have the GK pickup drive Triple play. Thanks.


Search on Susbemol's  old posts

https://www.vguitarforums.com/smf/index.php?action=profile;area=showposts;u=7390

===

FTP - DIY adapter to use GK 13 Pin Guitars with Tripleplay
(https://www.vguitarforums.com/smf/index.php?topic=8413.msg122194#msg122194)


QuoteIn essence, I built a box very similar to the one I used to connect a GK equipped guitar to the FTP

https://guitarsandsynths.wordpress.com/2016/01/19/roland-gk-to-fishman-tripleplay-adaptor/


AlakaLazlo

Quote from: germanicus on February 08, 2020, 11:14:01 AM
I did that.
I compared the recorded raw signal from the guitar quarter inch out jack and compared that against the midi note recorded from both the USB and 5 pin. The midi note was not triggering a soft synth.

Thank you! 
Hexstainocaster, Fender Strat and Electric XII, Godin ACS, Axon AX50 - Moog One, IIP and Mini, SEM, Dot.Com/Moon/STG/FSFX 110, Cubase Pro, 2xMR816, HR824, NS10M, Komplete, Omnisphere/RMX/Trilian, Z3ta+, Analog Lab, Slate MTi2, ML1 and Everything Bundle, Social Entropy Engine, ESQ1, DX7, Lavalamp.

jassy

Quote from:  Bluesbird on February 08, 2020, 09:11:59 PM
Today I tried the SY-1000 some more using the six dry hex inputs into six instances of Midi Guitar 2 each in monophonic mode.  The performance is better than any Roland unit I've tried and approaches very closely the performance of the Fishman Tripleplay--and in some respects performs a lot better as far as the sustain of midi notes and legato playing. Using 6 instances of MG2 also gives you the benefit of individual gain settings for each string adjusted within the MG2 vsts. Keep in mind the hex pickup I am using is a little hotter than the standard Roland hex, but I think using the Roland hex would yield the same results with the gain adjustment set higher. So if you want to use the SY-1000 with synth vsts, I would highly recommend this approach.
Well thats interesting MG2 in mono mode is the best (i own a TP also).
How is the cpu hit using 6 MG2 instances?  MG2 is not easy on the cpu, what cpu do you have?
The vsti used is also important, I see you are driving Omnisphere (what host is it?), it has a multi mode, so you can run only 1 instance, but with other vsti will need to run 6 instances...running all that  can be a cpu hog, thats why I not tried  with the GP10.


admin

#23
Quote from: jassy on February 09, 2020, 03:06:41 AM
Well thats interesting MG2 in mono mode is the best (i own a TP also).
How is the cpu hit using 6 MG2 instances?  MG2 is not easy on the cpu, what cpu do you have?
.



Quote from: admin on February 07, 2020, 11:27:31 AM
What type computer / CPU/RAM/Drive/ Mobo Model if known ?

https://www.vguitarforums.com/smf/index.php?topic=27372.msg201101#msg201101
Bluesbird wrote>
It is a Dell Inspiron 5676
Dell Inspiron 5676 Gaming PC 16GB RAM, 128GB SSD+1TB HDD, AMD Ryzen 7 2700 8-Core up to 4.10 GHz, Radeon RX580 4GB, VR Ready Desktop, RJ-45 Ethernet, Windows 10
https://www.amazon.com/Dell-Inspiron-Desktop-Ethernet-Windows/dp/B07XJLJ8CT


jassy

Quote from:  Bluesbird on February 09, 2020, 09:22:29 AM
I tried something similar with the GP-10, but it was not stable doing this. The audio driver for the SY-1000 is more stable. It is not the lowest in terms of latency (7.2 ms reported roundtrip latency) at the smallest buffer setting and 48000 sample rate, but it is stable. I can throw anything at it and it doesn't choke. With an octacore processor and Reaper's multiprocessor support, my computer handles it without breaking a sweat. Just for a test I stressed it out even more by not only loading six instances of MG2, but also 7 instances of Helix Native (one on the SY-1000 track and one for each separate string). Again, no problems at all (only 17% cpu usage).
Amazing setup, probably the best guitar to midi in existence today and add to that 6 re-guitar and 6 instances of Helix... endless...
Maybe would be ideal to use some sub host like bluecat patchwork, so you can create a preset with all the above and be able to recall it just like a single (super) preset... that would be much more efficient to manage sounds created like simple presets... the downside maybe is that in this way the process cant be distributed to the different cpu cores or maybe patchwork can do the core distribution itself.