Install loopMIDI
In the MG3 Chain, open MIDI MACHINES/MIDI OUTPUT/loopMIDI
Open Guitar Pro 8
From the File menu, select Preferences then click the Audio/MIDI tab
Select your loopMIDI port
BTW. If MIDI data is not being received by Guitar Pro 8 (GP8), make sure the GP8 window is active (i.e., selected). I ran into this as a solution when I couldn’t figure out why loopMIDI showed MIDI throughput but nothing appeared in the inactive (unselected) GP8 window.
Thank you for showing the setup to allow midi to be sent to GP8. Single notes are fine but two or more notes are not recognised by GP8. Can you offer any solution? Cheers
I haven’t tried sending chords to GP8. But I’ll give it a try. It may end up being a more practical workflow to capture the MIDI first in a DAW. Review/edit it. Then export it for later import into GP8.
Thanks. Yes, that approach might be easier. Another thought, maybe using hexaphonic pickup via Boss GP-10 into a DAW on six separate channels outputted to Guitar Pro 8. More complex.
The next version of Guitar Pro might address midi controller issues. Real time midi recording is what’s needed.
Cheers
I agree that there is a need for GM to support real-time MIDI recording. Has anyone posted an enhancement request for this? It’s needlessly complex to have to use a DAW setup for recording GM’s MIDI output.
Re a hexaphonic pickup. I’ve used the Roland (Boss) GK protocol for several years on both guitar (Godin) and violin (Fourness). How many GM users do you think would benefit? I believe it would be a major software development effort.
Since GK pickups do not convert audio to MIDI (they provide 6 separate audio streams), I’m not sure how the GK interface to GM would be accomplished. Today I was wondering if the Boss GKC-AD might help.
To make it all work, GM would require (1) an appropriate GK physical interface for input, (2) software to decode the GK-coded audio channels it received, (3) conversion of each of the 6 audio channels to MIDI, then finally (4) use of that 6-channel audio in some beneficial way.
Bottom line. I don’t think it would be worth the effort! What’s great about MG is that it doesn’t require ~$1300 of hardware. I used the old Roland GR-55 GK synth for many years but feel that GM does a better job in a simpler, cheaper, and more elegant manner.
the setup is time consuming, especially for some daws. but it’s a one time deal, once you have it working it’ll work forever.
the gkc-ad is used to connect an older style gk pickup to newer boss gear. the gkc-da connects newer gk pickups to older boss gear. neither provides what you want.
hex processing doesn’t mean roland/boss/gk solutions only. mg3hex would be unnecessarily burdened in accommodating this now clearly defunct technology.
and no version of mg3 could ever provide quality audio from a boss pickup (they are filtered to optimize note detection over fidelity).
however, if tablature is the goal, any gk guitar with a separate strings breakout cable and a 6+ channel audio interface will deliver the goods.
This can be done either with the GP-10 or the SY-1000. Then you need 6 instances of MG running in parallel or 6 instances of an instrument in a sampler like Kontakt. As Kimyo mentioned this can also be achieved using a breakout cable and a 6-channel interface.
There I see some potential for a future development of MG3. Give it more inputs that can be assigned to chains, and add 3 more chains. So far you can connect 2 inputs but the incoming signals go down the same route which makes it a sort of useless.
I can see further developments coming from MG3 with regard to use of hexaphonic outputs, etc. But to improve scoring notation, Guitar Pro, and similar, would need to improve.
With regard to the most successful of GK midi interfaces, the Roland GI-20 has been the best with Guitar Pro. But it’s not fantastic.
The future lies with software developments.
Great discussion. I agree with all points made. E.g.,
Getting back to MIDI recording. Since MG3’s modular design allows for the use of loopMIDI as a MIDI OUPUT channel, perhaps there’s a simple MIDI recording plugin that could be added to the chain to get the job done 
Re all of the hardware solutions involved with hardware audio-to-MIDI-to-synth solutions, I’m at the been-there-done-that point. My pedalboard just kept growing – and creeping toward obsolescence! I love the simplicity, modularity, versatility, and low cost of the software solution provided by MG3.
I’m going to have a look at Sibelius again as it’s real-time recording for midi. I will used it years ago.
That’s a good idea regarding a plugin to add into the chain on MG3. I’ll take a look at that as well.
Cheers 
Back to the topic of this thread
I originally did this MG3 to GP8 workflow experiment in support of my wife’s interest in using MG3 and GP8 for arranging and composing for guitar.
As a follow up, this morning I did an MG3 to GP8 chord test using a Taylor 814ce-N Grand Auditorium Acoustic-Electric Nylon-String.
This test involved playing a series of simple Am 4-note chords.
Result:
Though my technique was clean and clear, MG3 didn’t always successfully convert the 4-note chords to MIDI. Sometimes it got all or them. Other times only 2 or 3 notes. I based this conclusion on monitoring the MG3 keyboard display. This was cofirmed by what GP8 recorded.
GP8 appeared to record what GP8 sent via loopMIDI. However there was a significant lag and some innacurracy in how it stacked the notes received from MG3 (some “escaped” the stack as separately articulated notes).
Bottom line: MG3 didn’t always accurately convert the 4 note chords played, and GP8 does not appear to be flexible, performant, and “smart-enough” for real-time recording.
My recomendation to my wife is to use her DAW to (1) capture a clean audio recording, (2) convert the recorded audio track to a MIDI track using MG3, (3) clean up the MIDI result, then finally (4) Export the MIDI and import it into GP8.
That was a worthwhile test. Thanks. Plenty of food for thought. There are some variations that I will try based on your test. Will report back. Cheers
Thank you. Very informative and thoughtful thread!
MIDI-to-score and MIDI-to-tab is indeed non-trivial. We also did some experiments, some years ago (just after the MG2 release) and found GuitarPro, MuseScore, Sibelius to do worse than expected both with live inputs and when recording a MIDI file and converting it offline. The latter is better, but still leaves a lot to be desired.
For a year or so we were working on an “auto-tabber” and got much better results than GuitarPro and MuseScore. There were indeed two main technical challenges to this: training the transformer from MIDI to Tab or Score, and inferring string mappings. I think, for the former we had great results, and I thought we almost solved the second string mapping, 1) by using a slower “deferred” tracking, which can look forward in time and make better decisions, 2) by a reinforcement learning system to train the tracking to your guitar strings, and 3) not least by some smart UI tools that makes it very to tranpose and wrap chords around to new positions on the fretboard. Once you correct a few chords, the surrounding chords would automatically fall into place.
The main problem, and the reason we stopped, was that it also became clear that you really want a full notation app UI anyway, in order to make all sorts of small corrections. Even if the tracking, mapping and some high level tools makes a “perfect” transcription, you will still want to correct notes and make up for small variations in tempo or alignments, adding or removing a note to a chord, adding chord diagrams, keys, etc… So I think a satisfying solution actually comprise a full notation app, plus some clever tools, which would be an awesome project, but one that would be much safer with Tantacrul and the MuseScore team.
But… I’m just not entirely convinced that hex tracking/string detection is not a red herring, also for “auto-tabbing” purposes. Yet, as things look right now, the a hex tracker may happen before a notation app anyways 
Thanks for sharing these technical alternative strategies. Regarding transcription, it brings to mind a product I have some experience with, SmartScore.
My exploration of the use of MIDI Guitar as part of a notational workflow is entirely in support of my wife’s guitar-oriented composing and arranging interests. The ability to easily generate a MIDI file from Guitar MIDI is primarily what she wants. A MIDI recorder plugin or standalone MIDI recorder for Guitar MIDI output would be perfect.
What I like about MIDI Guitar is its use as a live performance tool with violin. Love it for that purpose
But I do wonder if the use of a polyphonic pickup (4-, 5-, or, 6-string) with Guitar MIDI would help in completeness and accuracy of note identification for each string and minimize false-note generation.
Here is a suggestion for a simple test: Use only the free strings and only fingered notes up to position IV on the fingerboard. Then there are no identical notes on different strings, i.e. the assignment of string and corresponding notes is unambiguous. Does the conversion to e.g. GPro 8 work without problems in this configuration?
Here’s the fingering I used in my simple chord test:
I simultaneously plucked strings 1, 2, 3, and 5. The resulting audio input was: