Any feedback on midi tracking on gm-800 ?

Started by strudl, August 02, 2023, 05:01:38 PM

Previous topic - Next topic

0 Members and 1 Guest are viewing this topic.

arkieboy

Quote from: admin_shawnb on August 07, 2023, 01:51:35 PMI was just thinking pretty much the same thing:
You cannot view this attachment.
The 'audio out of HW synth' part of this is somewhat meaningless as it depends completely on the hardware synthesiser you've plugged in.  Early Motifs were notoriously slow whereas external drum modules you would expect to be really fast.

Just midi out of the GM-800 is all that's needed
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

arkieboy

What are the consensus numbers for the GR55 and Fishman TriplePlay?
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

adamqlw

Quote from: arkieboy on August 08, 2023, 01:28:38 AMWhat are the consensus numbers for the GR55 and Fishman TriplePlay?

Germanicus cited this on another forum

Sy1000 High E string - 25 msec
Sy1000 Low E string - 48 msec

Tripleplay High E string- 10 msec
Tripleplay Low E string- 18-20 msec

arkieboy

Quote from: adamqlw on August 08, 2023, 01:48:08 AMGermanicus cited this on another forum

Sy1000 High E string - 25 msec
Sy1000 Low E string - 48 msec

Tripleplay High E string- 10 msec
Tripleplay Low E string- 18-20 msec

(that is actually of more interest to me, I have a SY-1000!!)

So it's looking like it's much better than previous Roland, still not quite in the TriplePlay range.

Thanks
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

adamqlw

#29
Quote from: Bluesbird on August 08, 2023, 06:20:29 AMWhat was the input latency of the sound card?  Let's assume it took 4 ms. for the guitar audio to reach the daw. That would add 4 ms to your figures (the audio arrived later to the daw and made the difference between the audio and the midi shorter than it actually is). So we would be looking at about 14 ms for high e and 28 to 30 for the Low E. Assuming these numbers are accurate and comparing to Wayne Jones's speed tests, this performs very close to but faster than the Roland GI-20 (which was the last dedicated pitch to midi DSP device that Roland manufactured). It is faster than any Roland device ever made (again assuming these numbers are accurate). It does not however knock Fishman Tripleplay Wireless or Connect (usb cabled version) out of first place as the fastest pitch to midi converters on the market (assuming these numbers are accurate).

Reaper reported input latency was 3 ms so 4 ms is in the ballpark

I would add that I also recorded the audio output of the GM800, which was within 1 ms of the MIDI note that was recorded. Reaper adjusts for the reported audio driver latency, which seems consistent with the MIDI note and the synth audio being aligned. As such I would not be inclined to add an addition 3-4 ms to measured figure

Elantric

#30
Quote from: Bluesbird on August 08, 2023, 08:04:49 AMWith latency numbers like this, and the endless options and sounds to be derived from the GM-800, I see no reason why this will not be a very successful product. Add in the new serial digital cable technology and I think you could score this as a home run for Boss/Roland. I think they got this one right—-too bad they left the SY-1000 to wither on the vine.

The battle will be the rampant
 "I dont read Owners Manuals" culture - getting the GM-800 to behave and not be a random note generator required weeks of reading and applying problem solving strategies, and in my case awaiting release of new firmware 1.03 and GM-800 documentation release and Boss Tone Studio for GM-800 .

GK-5 > GM-800 was definitely  not "plug & play" for me - which ultimately I feel will limit the GM-800 appeal for most guitarists. 
( complete opposite of my excellent 1st day VG-99 experience )

adamqlw

#31
Personal anecdote: I am super happy with this.

I had a VG8, GI20 (with an XV 2020), VG99, and Tripleplay (wireless), Jam Origin, and an SY200 over the years.

I really dislike the 13-pin cable, and when the output on my graphtech ghost/13 pin started to give random chirps, I bailed on the 13-pin system.

I tried the Tripleplay when it came out, the tracking *seemed* glitchy to me but that could be user error and my bad playing.

The Jam Origin was like black magic, in that it didn't require a special pickup and seemed pretty fast, but I could never get the glitching down to an acceptable level, not at all like some of the demo videos I have seen.

For the Tripleply and Jam Origin, I was running softsynths, and that adds the output latency from the audio interface drivers. The end result was for me, distracting.

The SY200 was fun but limited. It's now up for sale.

Fast forward to the GM800. Maybe my playing is cleaner after the years of practice, maybe my understanding of the limitations is better, maybe the hardware synth overcomes the issue of having to use audio interface, but the net effect is I am really enjoying it over the past few days. It now sits on a Pedaltrain Nano+ with my HX Stomp, and I feel confident enough to bring it out and use it in a live setting. It would be interesting to add a Tripleplay into the mix to trigger the GM800 via the MIDI in, but that's a heavy investment to add to my rig. So perhaps it doesn't track as well as the Tripleplay, or as flexible as a full suite of softsynths. But it's been comparatively easy to just plug in and have some fun, and reach the point where I think I could be using it a lot.

re: plug and play, there's certainly a learning curve, but I don't feel it's any deeper than what came before. Also, for all the shortcomings of Boss software (including the Waza-Air editor)... it's still a whole lot better than paging through the menus on the hardware

germanicus

Quote from: Bluesbird on August 08, 2023, 08:35:18 AMPeople talk about how glitchy the Tripleplay is, but they added a "monitor hand position" feature a few firmware's ago and it is the most rock solid tracking of anything I've tried in the last 25 years. Solely as a pitch to midi converter it is my favorite—the Roland GI-20 would be second. Jam Origin MG2 is great in monophonic mode playing single note lines.

Yeah I don't get much glitchy stuff with the Tripleplay.
The hand monitor position is cool, BUT you cant then do contrapuntal tapping across a wide range on the neck (it will ignore notes that are very far away).
So if you want to be stanley jordan on your midi guitar, use a dedicated guitar shaped controller like the YRG.

Im glad to see that adamqlw redid his tests. Those initial numbers were bad.

Those numbers you got on the second go round seem very useable.

I hope people enjoy the unit, and that we see a new unit in the future which focuses on alternate tunings/guitar modeling/HRM stuff.

Read the manual and learn the unit!
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

Quote from: jwhitcomb3 on August 07, 2023, 08:02:42 PMPatent protection only extends 20 years from the filing of the patent application. Once the patent term ends the technology becomes public domain. 

Right so 2034 for fishmans relevant patents. There is prior art in the description, so its not clear to what extent the algorithms themselves are covered by their 2014 application, but its clear they have a solid claim on the wireless transmission aspect. So don't expect to see any roland/boss units sending PTM data wirelessly for 11 years or so.
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

Quote from: germanicus on August 08, 2023, 08:56:12 AMRight so 2034 for fishmans relevant patents. .  .  .  but its clear they have a solid claim on the wireless transmission aspect. So don't expect to see any roland/boss units sending PTM data wirelessly for 11 years or so.
True- confirmed

jwhitcomb3

#35
Quote from: germanicus on August 08, 2023, 08:56:12 AMRight so 2034 for fishmans relevant patents. There is prior art in the description, so its not clear to what extent the algorithms themselves are covered by their 2014 application, but its clear they have a solid claim on the wireless transmission aspect. So don't expect to see any roland/boss units sending PTM data wirelessly for 11 years or so.
Which patents are you referring to? US Patent No. 9,460,695 claims priority to a January 18, 2013 filing (only 19 days of patent term adjustment (PTA)). So yes, just under 10 more years of coverage (assuming the maintenance fees are paid). However, that patent claims a very specific implementation, with limitations in the independent (broadest) claims (1, 11, and 17) where there must be two wireless transceivers, one on the instrument and a second remote, where the first transceiver sends "music control message data" to the second transceiver. This generally protects sending wireless MIDI from the guitar to a remote transceiver (see dependent claims 2 and 12 and independent claim 17). But apparently these claims may not prevent any implementation of pitch detection (including pitch-to-MIDI) that does not involve the wireless transmission of music control message data from the first transceiver to the second transceiver. Are there other patents you know of that protect Fishman's fast pitch-to-MIDI method?


germanicus

#37
It seems it was developed by Blue Chip, possibly jointly with Yamaha in the 90's.
"Transient Early Recognition"

Down the rabbit hole we go.

Looks like 5780759, applied for in late 1995. Assigned to Yamaha in March or 96.

https://ppubs.uspto.gov/pubwebapp/external.html?q=(5780759).did.&db=USPAT

Its expired at this point. As I stated above, fishman have a claim regarding wireless transmission of PTM.

Whether roland has decided to incorporate this transient early recognition is unknown.

Also see:

5,824,937
5,717,155

For more fun, peruse these:

https://patents.justia.com/patents-by-us-classification/84/654

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

#38
Quote from: germanicus on August 08, 2023, 11:26:54 AMIt seems it was developed by Blue Chip, possibly jointly with Yamaha in the 90's.
"Transient Early Recognition"

Down the rabbit hole we go.

Looks like 5780759, applied for in late 1995. Assigned to Yamaha in March or 96.

https://ppubs.uspto.gov/pubwebapp/external.html?q=(5780759).did.&db=USPAT

Its expired at this point. As I stated above, fishman have a claim regarding wireless transmission of PTM.

Whether roland has decided to incorporate this transient early recognition is unknown.

https://www.patentguru.com/search?q=5780759

http://www.panda-audio.com/midiguitar.php
Andreas Szalay wrote:

QuoteMIDI Guitar FAQ
I spent most of the last thirty years designing guitar synthesizers. I developed several generations of devices from the Shadow GTM-6 / SH-075 through the Blue Chip Music / Terratec Axon until the Fishman Triple Play today. Over this period I answered thousands of questions from users who had some difficulties to understand certain features or behavior of their devices. Now, as I have here a web page of my company, Panda-Audio, I want to take the opportunity to summarize this information in the form of an FAQ. I hope I can help users of guitar synthesizers a better understanding of their devices with this information. While these answers are based on my designs, I believe most of the information here is true for Roland or other devices too. I will update this page time to time with questions whatever I recall from my memories, or whatever new questions arise.

Andras Szalay


Mr Szalay's Patents
https://www.patentguru.com/inventor/szalay-andreas

jwhitcomb3

#39
Quote from: Elantric on August 08, 2023, 10:56:23 AMSearch on " Patents Andreas Szalay"


https://www.patentguru.com/search?q=Szalay,+Andreas

https://patents.google.com/patent/US9460695B2/en

https://www.google.com/search?q=Patents+Andras+Szalay&oq=Patents+Andras+Szalay&gs_lcrp=EgZjaHJvbWUyBggAEEUYOTIHCAEQIRigATIHCAIQIRigAdIBCDEzNjFqMGo3qAIAsAIA&client=ms-android-samsung-gs-rev1&sourceid=chrome-mobile&ie=UTF-8


https://patents.justia.com/assignee/fishman-transducers-inc


And read Andras Szalays GUITAR MIDI FAQ
http://www.panda-audio.com/midiguitar.php

Nope, I didn't see any either. Szalazys has lots of patents, but I only found one of his active US patents/applications assigned to Fishman. I found no Fishman patents claiming guitar pitch detection method/system. Based solely on this, I don't see support for the argument that Boss' guitar synth latency challenges may be attributed to patent matters.

Ludek

My info is based on simple aural experience. Though playing gm equipped with GK-5 brings quite good but not perfect tracking / occasional glitches as I experienced with GR55 and SY 1000/ using guitar to midi /in Reason/ is a real joy. No glitches even if I'm not as precise as I must be with GM itself.

gumtown

Here is a latency test I did, using Cakewalk DAW, UA101 USB interface, inputs are
1: normal pickup
2: GM-800 L/Mono output
3: external midi Roland Fantom Xr module L/Mono output A
Guitar with a GK-2A pickup
Not taking into account any USB/processing latency, these values are referenced to the normal pickup initial transient start.
A single one time test, no averaging, finger picked notes.
Maybe I could have tied a few times, used a pick, juggled the results, but no, this is straight off the bat tests.

Results:
LOW open E
1: npu = 0.068ms (start offset time)
2: GM-800 = 0.126ms
3: external midi = 0.126ms
difference (latency) = 58ms for both

High open E
1: npu = 0.565ms (start offset time)
2: GM-800 = 0.590ms
3: external midi = 0.590ms
difference (latency) = 35ms for both

High E 12th fret
1: npu = 0.725ms (start offset time)
2: GM-800 = 0.746ms
3: external midi = 0.748ms
difference (latency) = 21ms for GM-800, and 23ms for the externa midi





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

gumtown

Quote from: Recovering Roland User on February 24, 2024, 08:02:10 AMKeep in mind, this is from plucked string to audio waveform, and not plucked string to midi note production, which would be somewhat less, but yeah the GM-800 sucks. Hey Roland, the 80s called--they want their technology back!  As long as you play on the top four strings around the 10th fret and higher, you should be good--ha, ha. It's also great for ambient pads!! Can't believe all the people that shilled for this piece of crap. Apparently, being a nostalgic Roland fanboy can cloud reason and good judgment.


I guess you have your reasons for not liking Roland/Boss gear anymore, but for the purpose I use the GM-800 for, for me, it doesn't suck.

I don't use guitar to midi, and am happy with the internal tones of the GM-800, a big improvement on the GR-55, and you can add more wav expansion packs for new tones.
I also like that it is compact.
I use it in tandem with the SY-1000, which I guess both are a replacement for the GR-55 (x2).
I use the SY-1000 mostly just for the instrument and amp modeling.
Free "GR-55 FloorBoard" editor software from https://sourceforge.net/projects/grfloorboard/

admin_shawnb

I use the midi out to drive a Blofeld.

Honestly, the GM-800 is the best guitar-to-midi I've tried.

I even favor it over the FTP - in my experience the tracking & dynamics are better.  Feels more natural.

BROCKSTAR

I love my gm-800 and sy-1000 combo... Best guitar synth and virtual instruments/amps/effects combo on the market right now!

whippinpost91850

Gumtown what Rock sax are you using

arkieboy

Quote from: BROCKSTAR on February 24, 2024, 05:34:51 PMI love my gm-800 and sy-1000 combo... Best guitar synth and virtual instruments/amps/effects combo on the market right now!
Couldn't agree more.  I've been waiting all my life for a rig like this.
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

IMH1234

Whilst I am on the record as being on the fence about the GM800 to say it sucks because pitch to midi inherently has latency, and the Roland approach has marginally more latency than some other imperfect products that also have latency seems to miss the point - the logical conclusion is that given they are equally imperfect in their own ways, FTP and Jam original also 'suck' and are 'shilling' users?

The physics of triggering any synth - whether digital, sample based or analogue - from a guitar means latency is part of the deal. The only way this will be addressed with current technology is by the front end of the note triggering a COSM/HRM or GR300 type wave and morphing to the synth voice when the pitch detection has identified a stable pitch. Potentially simple enough for Roland to implement but for whatever reason they have not done so. In the meantime they have introduced a product that gives a solution that works for many users for whom the benefits outweigh the limitations and committed to the platform - seems like there is a lot to like about Roland and their guitar synthesis line up right now.

BROCKSTAR

Quote from: Recovering Roland User on February 25, 2024, 10:16:07 AMIt's funny that the FTP has a latency as low as 7 ms. on high e and 14 ms. on low E with superb tracking, dynamics, and "natural feel" but people sing the praises of a device that has latency numbers (as reported by Gumtown) the same as the Ibanez MC-1 and GR-700. The MC-1 and GR-700 met their demise largely because of their horrible latency, but all of a sudden the GM-800 with the same horrible latency is the best pitch to midi device in town. I find that laughable, and indefensible.

In addition, of my many different software synths, Zencore was at the very bottom. If horrible latency and dated 80s synth sounds is your thing, then I highly recommend it.

It's a pretty blue color though and the serial protocol is cool.

I realize fanboyism is a thing, so limit your vitriol towards me, please. 

 

If all you're getting is 80's sounds out of zencore, the problem isn't zencore, it's you....

AlakaLazlo

Haven't posted for a while. So here's my opinion. YMMV.
Initial disclosure - I have a pretty massive synth collection, as well as a custom Variax, a Helix and a ton of software instruments. Many years back, I was a professional synth programmer - what we now call a sound designer. 
I find it unnecessary to say something sux. Music making devices are tools and what works for one may not work for another. But it is undeniable that the guitar to midi latency being reported for the "new" Roland gm800 is significantly worse than my Axon AX50 which is 16 years old. With it's massive resources and market share, I think it's disappointing that Roland has never gotten close to the old Terre Tec or further developed Fishman systems on that functionality.
The SY1000 was underwhelming to me when I played it at NAMM. The onboard sounds of the SY1000 were just meh! compared to the sounds I can get from my other gear. As for Zencore, requiring a subscription to edit and load sounds on a purchased hardware device like the gm800 just feels sleazy to me. Yeah it's a fun little box. But I believe it should perform far better than it does.
That's just my $.02.
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.