Missing midi when two identical notes played sequentially on different strings

First let me say, I love this software, and use it live in my solo performances. But I came across this ‘artifact’ and just wondered if it’s a known limitation that might possibly be something that could be improved, or if anyone has any suggestions.
I do some finger picking, and have found that under certain conditions some midi notes were being skipped, delayed or multiple retriggered. I condensed the problem down to the following situation: if I continuously pick an open B at somwhere around 140 bpm, I get a stream of midi notes, no problem. Next, if I play the same B but at the 14th fret on the A string, same rate, I likewise get a reliable sequence of midi notes. However, now if I alternate between each string in turn between the two identical notes (so same overall rate of notes, like a very basic two finger picking), it intermittently misses the 14th fretted B. I assume this is because the open B is still ringing undamped, and the software is finding it difficult to detect the new note being struck in the presence of a still ringing identical note. However, this is a pretty common occurrence in picking where often ‘drone’ a note is being playing. Interestingly, if you move down, and alternate between say the 10th fret (G) and the open B, there is no problem. Move up one fret, and it’s usually ok. Keep moving up and you start to get issues. When you get within one fret either side of the14th (Bb or C), the midi stream gets very erratic, with missed or multiple triggering and even incorrect notes. By the time you get to about the 16th or 17th fret, the issue goes away.

I highlight the problem with the open B and the 14th fret B, but it likewise occurs when any two identical notes are played sequentially on virtually any other strings as far as I can tell, as long as one is still ringing when the other is played. I would suggest it doesn’t occur when you just play the sequence of notes just on one string because the string is momentarily damped when either the pick or your finger hits the string, thus allowing a clear new note to be differentiated?

Details of setup: clean 6 string Ibanez electric signal going directly into my Firewire 24 MKII Phonic mixer, lots of headroom. No Firewire dropouts detected by the driver. Win 7 PC, MG2 either standalone, or in Samplitude 11 DAW. Noise gate adjustment set fairly low (about 20% I’d guess) for good normal playing (chords and single note etc), but in any case, varying this doesn’t seem to affect the issue. Other than this problem, no other issues with MG2, so I believe it’s a real software issue, rather than setup. Maybe I’m expecting too much? Any thoughts?

Retriggering a note is not trivial: it is not easy for a monophonic system. In a polyphonic system it is far more difficult even.
The stuff that prohibits false retriggers may be killing your unisono. The 2 notes are different in their spectrum, MG most likely then thinks it is not a retrigger of the same note.
We’ll take this issue to the next version of our tracker.

2 Likes

Thanks Paul, I can certainly see why this is a challenge to detect. It would be great if you can find some way to address this in the next or future versions, I’m sure it would make a great enhancement to the MG2 capability and tracking.

1 Like

Ah an update to add here… I actually have two DAWs set up, my Dell M6800 laptop, which I use for live performances with a StarTech Expresscard Firewire adapter, and my trusty ~7 year old Zt desktop in my studio, which uses a soundblaster card with Firewire support, both i7 running Win7, and running with my Phonic Firewire 16 channel mixer, using Samplitude 11. It turns out that the missing notes problem only occurs on the laptop. I’ve just being trying it on my desktop, and it tracks great, very few missing notes, if any. Which is a bit of a PITA, cos I need the MG2 more for live applications, than the studio. Anyway, bottom line is, I don’t think it’s a MG2 issue, more likely some issue in my laptop hardware/software. I have no problems when using the laptop for recording up to 16 channels live (no dropouts detected, or problems with hitting the CPU max) but obviously something is getting screwed up somewhere - my suspicion is the StarTech firewire card. If anyone has any further thoughts or suggestions, I’d appreciate it, but other than that, at least I known MG2 will do what I need it to do and does track ok, given properly working hardware! If I ever manage to isolate the issue, I’ll update further.

Drewmo,

You certainly provide a clear and concise description of the issue you’re experiencing. It appears you have the developer’s attention and it is an issue they will address. That’s great for everyone.

Since it appears you can reproduce the issue on demand easier on one computer than another I’m wondering if the issue might be a difference in the digital audio stream presented to the program.

One exercise that might be helpful is to record the guitar digital audio input to MG2 and the resulting MIDI data stream output while attempting to induce the issue. Record a performance on both computers with each computer set up as you normally use it. Hopefully the issue will occur as you described on one computer but not the other.

Bundle each performance into an archive file and upload to cloud storage (Dropbox, OneDrive, Box, etc.). Then provide the links to the developers.

I can think of several ways this might help them. First they can easily recreate the issue on demand. They can compare the audio and midi files from each performance to dig deeper into the question of whey does the issue happen on one computer versus the other.

Yeah, since posting this update, this has been bugging me, partly because I need the system working on my laptop for mobile live application, but also just from an engineering perspective, it’s a puzzle (I’m an analog IC design engineer).
What I intend to do is record a problematic sequence on my looper pedal, then set up each system in turn, using the identical setup in each case, and record the streams from the identical sequence. That will confirm whether or not it’s something in my playing/setup or whether it’s really a difference between the two computer systems. This might take a while though, I’m going out of town for a few weeks, so don’t hold your breath :slight_smile:

Also, I’ve bought another Firewire Expresscard from this company: which I want to check out, in case that is a factor:
https://www.pimfg.com/product-detail/EXC-1394-2
Appears externally almost identical to the Star Tech Expresscard, but for less than $12 including tax and postage, seems worth checking out, considering the Startech card is about $90 now on Amazon! If that works out, I’ll post an update, as a good low cost 1394 solution for laptops (albeit older ones with Expresscard ) is hard to find.

I clicked on the link. I believe you’ll be disappointed but I hope I’m wrong.
A couple of things lead me to believe you’ll be disappointed.

Drivers supported MS Windows 2000/XP/XP 64-bit/Vista

For Driver Please use vt6130

No link to download the driver.

In Microsoft Windows the type of ic chip the Express Card uses makes a BIG difference. Texas Instruments ic chips were the only ones most people could get to work consistently. You may find this thread in a Windows 10 forum interesting because it does provide the names of several Firewire adapters that work in Windows 8 or 10 while using a Windows 7 driver.

Windows Ten “Is Firewire IEEE 1394 Supported?” Thread

OK, first, unfortunately, I’ve checked this out thoroughly on my laptop and desktop, and under careful test conditions, I’m afraid the problem is still apparent on both machines, and there is no significant difference when all conditions are identical. So to check this out further, I recorded (in Samplitude) a sequence of alternately plucking the same note on an open string (G) and on the A string fretted at a G (10th fret). From that, I carefully chopped out two bars (total of eight notes), and then copied them 8 times over, using snap to bar to ensure each set of eight notes lands in the same place relative to the bar. I then dropped this onto the track with MG installed, and recorded the midi output. I repeated this 3 more times with bend note range 2 enabled, and then 2 more times with bend disabled. I have attached an mp3 of the looping guitar recording, and also the score of all 6 midi outputs together, which should just show an identical train of single notes. I also output an example recording of the midi (piano), one for each case of with/without bend enabled.

If you listen to the midi output, you can hear the missing notes. With bends disabled, you can hear the additional passing notes. If you look at the score, where there are tied notes is where a note is missing
With bend enabled (last two lines), usually I get a couple of errant passing notes as my plucking finger settles on the string and then plucks. If I turn off bending, then often it misses that note altogether instead. Now the strange thing is that even though this is the same recording looped 8 times, the result varies throughout the midi output. Sometimes notes are lost, sometimes delayed, sometimes not. Even stranger, is when I repeat the midi output, then the results are different each time. Very unpredictable.

Not sure if this is any use to anyone, but hopefully it might give you guys at Jam Origin something to think about and maybe see if any improvements can be made sometime down the road :slight_smile:

https://jamosapien.com/uploads/default/original/1X/6d616b202d9bdd831688fa2c7070681c5b879a54.mp3 https://jamosapien.com/uploads/default/original/1X/862d0e43c0159d78cf3fe76e729f75968c3ca122.mp3 https://jamosapien.com/uploads/default/original/1X/047546b8db2181d6d5c603616b1e4dc0fba8be7d.mp3

thanks for the data!
It is unpredictable whether we can fix that soon: MG2 was just not made to detect a new tone that is the same as a note that is allready sounding. Changing that behaviour might lead to more new problems than that it fixes…

@Drewmo, Great troubleshooting. It looks like the developers have a better understanding of what might be happening. While positive results are not promised I’m sure their understanding of what is causing the dropped notes issue will help them make Midi Guitar 2 more resistant to dropping notes. Thanks for taking the time to help clearly define the issue. That is helpful to everyone that uses the product.