GR-55 Guitar to MIDI Performance Analysis

Started by germanicus, February 12, 2011, 11:31:04 AM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

germanicus

Ok, here are some results from a quick and dirty test of the gr-55's conversion speed. Wayne Jones midi speed tests were the inspiration for this (http://www.joness.com/gr300/MIDI_SPEED.htm).

I used a Fender strat in which I installed a GK internal kit. I used the "Vibraphone" patch (Lead 18-2) in the gr55.

I recorded the output of the gr55's synth into sonar 8.5 producer edition through a tascam FW-1082 interface.
I recorded the regular magnetic pickup output (lace sensors) into the FW-1082 by directly plugging the regular guitar output into the interface.

Each string was picked 3 times, and the results averaged.

I determine the number of samples between the initiation of each note. Then Divided by 44100 (44.1k sample rate) and multipled by 1000 (to get millisecond delay).

High E string:
22.426 MSec
19.977 MSec
19.886 MSec
AVG 20.763 MSec

B String:
27.347 MSec
27.868 Msec
26.077 MSec
AVG 27.097 Msec

G String:
25.714 MSec
26.689 MSec
26.258 MSec
AVG 26.221 MSec

D String:
30.408 MSec
30.113 Msec
29.274 MSec
AVG 29.931 MSec

A String:
35.170 MSec
35.804 MSec
34.127 MSec
AVG 35.034 MSec

E String:
41.224 MSec
42.244 MSec
41.315 MSec
AVG 41.594 MSec
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

Germanicus,

Do you own other Guitar Pitch to MIDI gear? (Roland GR-20, GR-33, GI-20, Axon, etc.)

Curious to see how those stack up to the GR-55 using your same test method.


ddlooping

Thanks for taking the time to make that test, germanicus. :)

Quote from: Elantric on February 12, 2011, 11:45:52 AM
Do you own other Guitar Pitch to MIDI gear? (Roland GR-20, GR-33, GI-20, Axon, etc.)

Curious to see how those stack up to the GR-55 using your same test method.

+1

Diaz Guitars (work in progress)

germanicus

Yes, VG99, a GR33 and a G50.
Ill attempt to do the same test with the gr33 once I get it back from the shop (needed a new Backlight for the screen).
Can do the VG99 and G50 in the meantime.

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

paults

Do some of the other synth sounds have a faster attack than the Vibraphone? (PCM guitar, snare, rim, side stick, something plucky..)  With the extremely small times you are measuring, any tiny "bloom" in that particular sound will skew your results. 

szilard

#5
I did a quick and dirty analysis as well.

Equipment: Godin lgxt, mag PUs output from guitar directly into left stereo channel on an echo audiofire 4 IF and mono output of gr55 into the right stereo channel. Recorded into Sonar X1 and analyzed in WavLab Essential. I used the "Vibraphone" patch (Lead 18-2) in the gr55.

Each string was played on the fifth fret and I also played the lo open e since it should have the worst latency. I've done a quick setup of the Godin, but haven't tweaked it. Strings: Elixer .010 - .046.

hi e fifth fret: average 22 msec
22, 21, 22

b fifth fret: average 25 msec
25, 24, 25

g fifth fret: average 24 msec
25, 22, 25

d fifth fret: average 28 msec
32, 26, 27

a fifth fret: average 33 msec
31, 31, 37

lo e fifth fret: average 35 msec
44, 30, 32

open e: average 55 msec
54, 56, 56



A2theT

Its very important to also post the string gauges you guys use on your guitars when testing.  Personally I use DR Black Beauties 10-46 and have been thinking about stepping it up a notch to 11-50.  Wonder what difference that will make on tracking...
HEAVY on the METAL
Axe-Fx II, Roland VG-99 + FC-300, Roland GR-55, Digitech Jamman Stereo, Ibanez/ESP/Jackson Guitars

embryo

Hi, thanks for starting this it is definitely a very useful reference.

I'd like to also suggest that a tracking performance chart include mention the open string, 5th fret and 12 fret postions and also include GK models and the String on Piezo types like the RMC and Ghost.

Tanx

R

drjoness2001

Wow...everyone is getting testing data clearly slower than what I had. I never had the low "E" string slower than 30 ms. 55 ms is GR-700 territory.

On the other hand, I used a custom patch with a simple square wave. I wonder if there is something in the vibe sample that is slowing the system down, maybe some envelope controls, or perhaps something in the waveform sample format. I'm pretty sure the samples are saved as compressed waveforms, maybe this is adding some extra time.

paults

I'm with you, Wayne.  I think a faster attack sound, with PCM going through the same signal path would have a waveform with an easier to define starting point.  I promise anyone reading this thread: The 55 tracks *much* better than my trusty old GR-700 ever did :)

   

fizog

I was wondering - do playing dynamics alter the latency at all? For example, if you were to mute then play a string would it have a higher latency than a string that was simply strummed?

Slotracer

Using only my seat-of-the-pants analysis, I find that the GR 55 tracks almost as fast as the Axon MK II. The only real advantage the Axon has over the 55 is the comprehensive editor program supplied with the unit. It allows for some really nice fine tuning to help facilitate the fast tracking. Setup is something I would consider crucial to getting optimum performance. Where the 55 rules is in every other category. Mainly, the on board sounds are, to me, outstanding. And having the ability to stack up to four sounds simultaneously is another big plus.
But to return to the topic, I found the speeds vary greatly from patch to patch. I'm not sure how Roland can substantiate their claim on having the fastest midi tracking ever. In some pcm sounds it's fast, in others it isn't. I certainly don't regret trading my Axon for the 55, though.

gumtown

#12
If Axon are now out of business and defunct, then Roland would be the fastest (based on what is available now today).

There is a comprehensive software Editor for the GR-55 here
http://sourceforge.net/projects/grfloorboard/

Quote from: Slotracer on June 02, 2011, 03:51:15 PM
Using only my seat-of-the-pants analysis, I find that the GR 55 tracks almost as fast as the Axon MK II. The only real advantage the Axon has over the 55 is the comprehensive editor program supplied with the unit. It allows for some really nice fine tuning to help facilitate the fast tracking. Setup is something I would consider crucial to getting optimum performance. Where the 55 rules is in every other category. Mainly, the on board sounds are, to me, outstanding. And having the ability to stack up to four sounds simultaneously is another big plus.
But to return to the topic, I found the speeds vary greatly from patch to patch. I'm not sure how Roland can substantiate their claim on having the fastest midi tracking ever. In some pcm sounds it's fast, in others it isn't. I certainly don't regret trading my Axon for the 55, though.

Free "GR-55 FloorBoard" editor software from https://sourceforge.net/projects/grfloorboard/

Machh_2

in my opinion, the best way to got low and same latency in all strings, is use same gauge string in all guitar, some like E high (0.10) string in all 6 positions and then change the pitch to right tune in GR-55.
at the beginning seems a little weird playing a stringed instrument with all the same, but you get used to the track and the result and latency is great.
already played a guitar so I saw no difference in playability ...