Standardized Guitar to Internal Synth Performance Test - Help out!

Started by germanicus, February 12, 2011, 05:33:13 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

germanicus

Using a quick short test, I think it may be beneficial to see the results folks are getting from their GR55's, as well as comparing it to other units (ie stringport, gr33, gr20, Axons etc.).

Obviously there are numerous factors at play which impact the performance of a given unit, ranging from player technique, string type, string 'freshness', guitar set up (ie buzzing frets), unit settings (sensitivity, low velocity cut), etc. This being the case, I still believe there may be value in seeing 'real world' practical numbers folks are getting with their setups.

I propose a quick test that people can perform on their gr55's, using the Vibraphone patch 18-2 in the "Lead" preset section. Other guitar synths can use a similar patch, preferably one with a sharp quick attack, that allows one to differentiate doubled or missed notes.

If folks have a few minutes, they can post some information on their gear and settings, and the results of this quick test.

Id like to hear how quick people are able to play (what tempo) sequential chromatic 8th notes up the low E string (6th string).
What tempo are you able to play "E F F# G"  consistently over and over without getting doubled or missed notes.

The Vibraphone patch appears good for this test as it has a percussive attack, and you can easily hear doubled triggers/glitches.

If folks have multiple units or 13 pin guitars, they can post multiple entries.

I propose people post the following info (if you have ideas and/or additions to this informal survey shout them out):

1.  Guitar Synth model: (Gr55)
2.  Guitar Type: (Fender RR Strat, Godin xtsa, etc)
3.  Pickup Type: (Gk3, RMC Piezo, Ghost etc.)
4.  String dampening above nut: (Y/N)
5.  String type used: (D'Addario xl 10's etc)
6.  String 'Freshness': (Brand new, 5 hours use, 10 hours use)
7.  Palm muting other strings: (Y/N)
8.  E String sensitivity setting (6th String):
9.  Velocity Dynamics:
10. Play Feel setting:
11. Low Velocity Cut setting:
12. 6th String PU <-> Bridge distance: (20mm)
13. Patch used: Vibraphone <Lead 18-2>
14. Tempo in BPM clean 8th notes attained: (160)
15. GTR-MIDI MODE (Mono, Poly)
16. GTR-MIDI CHROMATIC (On, Off)
17. GTR-MIDI Data Thin (On, Off)
18. GTR-MIDI EXP PDL BEND RANGE (-24-+24)
19. GTR-MIDI MODE (Mono, Poly)
20. GK SETTING PU TYPE (list GK type)
21. GK SETTING PU SCALE (list mm, ST, LP)
22. GK SETTING PU POSITION (Normal / Reverse)
23. GK SETTING PU PHASE (Normal / Reverse)
24. GK SETTING DISTANCE PU-BRIDGE (mm)
25. GK SETTING SENSITIVITY (0-100)
26. GK SETTING VELOCITY DYNAMICS (1-10)
27. GK SETTING PLAY FEEL (1-5)
28. GK SETTING LOW VELOCITY CUT (0-10)
29. GK SETTING NUANCE DYNAMICS (0-10)
30. GK SETTING NUANCE TRIM (0-10)
31. GK SETTING DOWNTUNE SHIFT (0-5)
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

1.  Guitar Synth model: Gr55
2.  Guitar Type: Variax 300
3.  Pickup Type: GK2a
4.  String dampening above nut: No
5.  String type used: D'Addario EXL110's
6.  String 'Freshness': 5 hours use
7.  Palm muting other strings: No
8.  E String sensitivity setting (6th String): 15
9.  Velocity Dynamics: 5
10. Play Feel setting: 2
11. Low Velocity Cut setting: 6
12. 6th String PU <-> Bridge distance: 14mm
13. Patch used: Virbaphone <Lead 18-2>
14. Tempo in BPM clean 8th notes attained: 170
15. Any other info people feel is relevant: Faster tempos begin hearing doubled notes.
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

Elantric

If possible, during the test, record the audio, the Dry guitar on the Left, PCM Vibraphone on the right, and post as an MP3.
And post ALL your GK PU settings, including the new GR-55 controls, and your pick gauge.


Elantric

Germanicus, lets re-sort data we want collected in the 1st post, perhaps make it s competition on who has the fastest GR-55 GTR- to internal Synth?
We can keep the GR-55 Vibraphone patch, but lets gather more data on all the possible settings that "may" impact these numbers.   

Also we still need a true "Standardized GTR to MIDI Test" using external MIDI sound generator.

Realize not every Sound Generator triggers with the same response timing speed to external MIDI Note events.

germanicus

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

Interesting I will try with my Axon.

Only to clarify (not native english may use different nomenclature), by play "8th notes" you mean 2 notes per beat = 8 notes every 4/4 bar?

germanicus

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

vanceg

Hmmmmm.  I applaud the test...and at the same time suggest that it's not necessarily very scientific or accurate in it's current form.

If we are going to record the Synth output on one side, and the dry output on the other side; where are we tapping the dry signal from? Is the dry signal also running through the GR-55 (or stringport or whatever other device we are using?). The GR-55 kind of has the ability to do this, but some converters (such as the Axon) don't.

Also - If we compare a GR-55 to an Axon or StringPort, we are comparing one product that does NOT require conversion to MIDI to trigger an internal synth (the GR-55) to other products that DO require conversion to MIDI and triggering external synths (such as StringPort and Axon).  Furthermore, the StringPort can send MIDI directly from it's software to SoftSynths whereas the Axon cannot.    So, this surely isn't an 'Apples to Apples' comparison.

Again - I applaud the comparisons of one product to another and I applaud the test!  Let's do it.  But let's also understand that we aren't going to end up with any sort of objective comparison about which converter is actually "fastest" or Best.

Now, to counter my OWN point I would also say: Who cares if we are creating an objective test?  We ARE sort of determining which of these devices is able to accomplish one particular task most accurately.

In addition to testing the internal sounds of the GR-55, I'd suggest that we compare how the GR-55 triggers external synths and compare that to the StringPort and Axon.






musicman65

Do we know for sure that internally the GR-55 isn't using MIDI? Is this info from Roland? What other method would they use to trigger the internal synth? I've read several places that the latency of MIDI alone is 1ms. Seems fast enough for internal use. Just curious.

musicman65

germanicus

I took the dry signal right from the guitars regular guitar pickups. I agree theres tons of variables involved. I didnt wish to use the USB because it could potentially introduce another variable (driver performance). My thinking was to record the analog outputs from both units, 1 the guitar itself, and 2 the gr55's outputs. Several areas are ripe for investigation, including:

-Analog synth output of internal sounds (the test I initially proposed)
-USB midi performance
-USB audio performance (both of the internal synth sounds and modelled sounds)
-Midi out performance (is there a speed difference going into a DAW through a seperate interface, ie how good are the Roland ASIO drivers)?

My interest in testing was bourne out of curiosity to see what kind of results others are getting from their GR-55's, to potentially see if my setup could be improved upon. I also think its just interesting to measure rolands claims of improvements and as a general matter of consumer evaluation for those considering purchasing a guitar synth.

Im also a bit confused about the notion that Roland doesnt convert to midi internally for its internal sounds. I remember reading something that hinted at this, but it still has to be determining pitch correct? The process of converting that pitch information to MIDI cant be all that involved right? Or am I mistaken? I really just dont know.




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

Macciza

Quote from: germanicus on February 13, 2011, 09:38:54 PM
Im also a bit confused about the notion that Roland doesnt convert to midi internally for its internal sounds. I remember reading something that hinted at this, but it still has to be determining pitch correct? The process of converting that pitch information to MIDI cant be all that involved right? Or am I mistaken? I really just dont know.
Yes that was certainly the case with the GR1 - not convert to midi to control internal sounds.

Basically MIDI sent to a synth module needs to be decoded to actually play the synth engine so instead Roland control their synth engine directly rather convert to midi then decode the midi to play the voice - or thats the best way I can explain it . . .

MM
'70s Strat, Brian Moore iM, VG-8, VG-99, FC-300, VL-70m, StringPort, SoftStep, Sentient6, iMac QC i7 27".

Spider

1. I would only suggest that Costas founder of Even Harmonics (Gtak 4) is working on VST plugin called TransM (it's beta know) which will do similar task like Gtak and has one unique feature I hope very useful for "Standardized Guitar to Internal Synth". It's "training" function which show gap (in ms) between audio and midi and percentage of ghost/correct notes. Maybe we can use this for test?

2. One problem with TransM: When we route dry guitar to computer we create additional latency which isn't a problem when the guitar is connected to amp/PA.6

Sorry for my English (the/a thing etc...)

Quote from: vanceg on February 13, 2011, 08:07:01 PM
Hmmmmm.  I applaud the test...and at the same time suggest that it's not necessarily very scientific or accurate in it's current form.

If we are going to record the Synth output on one side, and the dry output on the other side; where are we tapping the dry signal from? Is the dry signal also running through the GR-55 (or stringport or whatever other device we are using?). The GR-55 kind of has the ability to do this, but some converters (such as the Axon) don't.

Also - If we compare a GR-55 to an Axon or StringPort, we are comparing one product that does NOT require conversion to MIDI to trigger an internal synth (the GR-55) to other products that DO require conversion to MIDI and triggering external synths (such as StringPort and Axon).  Furthermore, the StringPort can send MIDI directly from it's software to SoftSynths whereas the Axon cannot.    So, this surely isn't an 'Apples to Apples' comparison.

Again - I applaud the comparisons of one product to another and I applaud the test!  Let's do it.  But let's also understand that we aren't going to end up with any sort of objective comparison about which converter is actually "fastest" or Best.

Now, to counter my OWN point I would also say: Who cares if we are creating an objective test?  We ARE sort of determining which of these devices is able to accomplish one particular task most accurately.

In addition to testing the internal sounds of the GR-55, I'd suggest that we compare how the GR-55 triggers external synths and compare that to the StringPort and Axon.

vanceg

Quote from: musicman65 on February 13, 2011, 09:10:13 PM
Do we know for sure that internally the GR-55 isn't using MIDI? Is this info from Roland? What other method would they use to trigger the internal synth? I've read several places that the latency of MIDI alone is 1ms. Seems fast enough for internal use. Just curious.

musicman65

Yes.  Confirmed. Direct from Roland. Same with the GR-33. 
What happens is that internally the device detects the pitch and amplitude of the signal and then uses that information to trigger the internal synth.  It is VERY similar to what happens in a Pitch to MIDI converter, but the signal is never converted to the MIDI data format and does not go through the normal process of being transmitted (even internally) as MIDI nor received as MIDI... This does reduce the time between picking a note and hearing a sound from the synth significantly.


musicman65

Thanks for the info on internal triggering. I was hoping that was not the case though. This means external midi won't be as fast. That's a shame because with today's processing speeds, the time it takes to create a midi note message should be negligable in comparison to the time it takes to sample a half-wave of the fundemental frequency of a string and calculate it's pitch....dang.

I am still waiting to hear reports by external hardware synth users on what the latency will be. The USB midi folks are saying good things. Anyone tried it yet?

musicman65

jassy

Tried it briefly with my axon ax50.

1.  Guitar Synth model: Axon AX50 connected via midi to my pc using a soft synth with a vibraphone  sound (connecting with usb could be better?).
2.  Guitar Type: Godin Multiac Jazz
3.  Pickup Type: piezo internal
4.  String dampening above nut: No
5.  String type used: D'Addario EXL110's
6.  String 'Freshness': 3 or 4 weeks
7.  Palm muting other strings: No
8.  E String sensitivity setting (6th String): -
9.  Velocity Dynamics: -
10. Play Feel setting: -
11. Low Velocity Cut setting: -
12. 6th String PU <-> Bridge distance: piezo
13. Patch used: Virbaphone
14. Tempo in BPM clean 8th notes attained: 225

That phrase you suggested maybe is a bit cumbersome because the armonics generated by the pulloff needed (at faster velocities) between the last and the first note, also its an area where is somewhat difficult to play and  probably all that introduces aditional problems not exactly related to the latency question.

Playing only the G note in that #6 string I can play faster: triplets around 250 BPM, but I think we are getting near the limits of physical playing (at least to me).

Because the velocity also its difficult to judge if all the notes are ok, but my impression is that they are "musically speaking" ok.

As i said its not a very scientific test but a "musical" one, so take it as its. Sorry its not a GR55 test, seems that unitl april I cant have one.

macjones

Quote from: musicman65 on February 14, 2011, 05:01:33 PM
Thanks for the info on internal triggering. I was hoping that was not the case though. This means external midi won't be as fast. That's a shame because with today's processing speeds, the time it takes to create a midi note message should be negligable in comparison to the time it takes to sample a half-wave of the fundemental frequency of a string and calculate it's pitch....dang.

I am still waiting to hear reports by external hardware synth users on what the latency will be. The USB midi folks are saying good things. Anyone tried it yet?

musicman65

Perhaps not.
Ideally, you would want the best possible performance, and it's logical that internally it would be superfluous to have a midi stage if it can be done more efficiently without it. But nothing would prevent good midi conversion in a parallel for use with external synths.

musicman65

Agreed...I do want the best performance. I already know what the GR-55's internal performance is. I was hoping that was with true midi conversion so external would be just as fast. Alas...it is no to be.

Thanks,

musicman65

vanceg

Quote from: musicman65 on February 15, 2011, 05:40:53 PM
Agreed...I do want the best performance. I already know what the GR-55's internal performance is. I was hoping that was with true midi conversion so external would be just as fast. Alas...it is no to be.

Thanks,

musicman65

Well, kind of "by definition" it CAN'T be as fast simply because you have more stages to go through triggering external synths than internal ones. 

I mean, you have to convert the Pitch and Amplitude (and bend) information to MIDI. Takes a little time.
Then you have to get that information to the transmission systems. Takes a little time.
Transmitting it over a cable takes a (very) little time.
Receiving at the other end and making sense of the data takes some time. (see note below).
Then the synth engine in the external synth has to trigger.
There is an attack time associated with the sound you are making...that needs time to happen.
Then the audio needs to be converted to analog (takes some time).

SURE - all of that can be "very fast" and may not be the major part of Guitar Synth latency (in fact I'm fairly sure it's not the cause of most GuitarSynth glitches or general dissatisfaction with the "feel" of guitar synths), but it DOES take some time.

germanicus

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