just found this application, and it seems to best accomodate my use case, out of all the current offerings: process guitar tracks (mostly acoustic/classicla) recorded into DAW projects to render useable notation that only needs to be finetuned.
My first attempts were promising, working with both the latest v3 beta and the last v2 official, inside Cubase Pro 14 (Win). I generally load arrangements that need to be complemented by a guit track, jam live into the DAW and then deliver parts to the guitarist that hit the readability/accuracy sweet spot (multi voice vs. tied layers, you know…)
Not so long ago i have tried Melodyne for that task (it does the audio to midi thing actually), but my attempts were frustrated, possibly due to user incompetence or lack of specific polyphonic accomodation on Melodyne’s side.
So all hopes on MidiGuitar! I cut my first teeth with the demos (see above), where v2 worked right away and v3beta gave some curious warnings (alledging Cubase tracks with 512 buffer–> too large, whereas I am on 256 actually according to RME settings)… with v2, everything from midiport recognition to buffer acceptance was working out of the box, not (yet) so much the 3 beta.
anyway. I’ll gladly purchase a license to pursue this, unless someone points me to yet another application that better serves this singular purpose (guit audio to notation). Happy to hear any thoughts on this, and apologies for the verbose ramble.
Although converting audio to MIDI for notation purposes is not as “flawless” as one might think, I don’t know of a better solution than using MG3.
There are many “helpers” and automations in Cubase 14 to clean up the recorded MIDI events before notation conversion. However clean playing and a perfect setup of the guitar is essential.
MG3 works well with Cubase and the buffer size message you received is probably due to the audio interface settings in the Cubase studio setup.
Thank you for your reply.
Regarding the last point, the audio buffer is set via the RME interface, even if it is accessed within the Cubase settings, and it is at 256… I am not aware of any other setting/channel within Cubase that observes yet another buffer setting, but I’ll gladly stand corrected. And as I wrote, MG2 does “accept” the guitar audio input at the same settings.
Besides all your comments re: cleanliness of playing, setup etc, do we think an acoustic (nylon) signal is particularly vulnerable to recognition. compared to an electric?
If your RME (instead of my Steinberg) interface is showing 128 or 256 samples buffer size on the studio setup page (and that is the only location to change the buffer) you should be good to go.
I have no experience with nylon stringed guitars / acoustic with a piezo PU works fine, however I prefer electric guitars but that is an individual matter.
all good, the erroneous buffer reading seem to have been a quirk of the trial version, the licensed v3 acknowledges the buffer set in the device driver.
I have been trying to skin the Audio->Midi Guitar transcription cat for years.
I have found no real easy solution until MG3 came on the scene, specifically MG3Hex.
With the one string per channel you can now get the notated and/or tab fingerings along with better separation. This is a god send along with the cleaner triggering that comes with MG3Hex.
You need a Hex pickup and breakout box/suitable Roland synth to allow 6 channel input and also pay for the MG3Hex upgrade. I’d been using Roland Synths (GP-10 for this purpose) for ages so only need MG3Hex.
I think it’s well worth it, for transcription essential even.
so it appears the hex system you suggest would only work with electric and steel string guitars, not concert guitars, unless one splurges for a Godin SA, correct? no real options to fit an aftermarket PU onto a normal classical guitar, or a Yamaha Silent (which I own)? that would be a bummer, as my notation chores lie mainly in the classical field. I may have missed something, though? thanks.
all of the roland/boss solutions require steel strings. i think the godin sa uses graphtech piezos, but that graphtech doesn’t sell these to end users.
so, unfortunately, there is no way to get hex output from a normal classical or the yamaha silent.
Thanks kimyo. This is getting expensive;) going back to my originally stated purpose of merely speeding up notation, I’ll ponder the options, and am grateful for all the pointers and comments so far.
The bitter irony is I now recall having owned a Godin Grand Concert at one point, and couldn’t even say for sure whether it had the 13pin jack or not, because it didn’t matter at the time… (sigh)
Do the aftermarket hex PUs not work on classical guitars just because of the steel string requirement, or because of the string spacing? if it’s the former, I’ve heard of (treble) steel strings suitable for concert guitars that could be mounted on a dedicated notation input instrument…
i was wrong, those are what you want. might be tricky to figure out how to replace the current bridge.
you’d also need the hexpander board, a splitter cable and an audio interface.
then you will have the tablature input system of your dreams.
i had very similar pickups in a steinberger clone. a vendor in new hampshire sold them to me, and included adjustable mounts which wouldn’t be too hard to fabricate on a 3d printer.
I have a midi-equipped (built) in Godin SA which works well in MG3Hex (and is a lovely guitar) — not fully evaluated everything yet but I think my steel string acoustic gives a better response/feel to acoustic and nylon guitar sample libraries (mainly Ample for me).
My lighter string electric (external GR pickup) and the Godin both make very decent more general midi controllers for electric guitar sample libraries and other instrument samples or synths.
I do not know of any external retrofit for nylon guitars.
