Wish "TUNING background Assistant" - new tuner Concept

BobXX

BobXX

Inspired
Let's discuss about a concept for a new tuner type, don't think it exists yet, theoretically conceivable on systems like FAS ones.

GOAL: a FAS tool to tune a string without stopping to play, like more talented guitarist do by ear
For some musician it may happen that there are rare moments in gig to stop and check tuning,
I think this concept will help the musician to remain more focused on the performance, not on guessing the string out of tune.

The Concept.
Imagine an optional tuner (in addition to the real time one) working in the background while the musician normally plays,
and doing a smart analysis of all the strings played while the musician plays normally and trying to work out if a string is out of tune, maybe beyond a deviation limit set by the user.

Whenever there is sufficient 'proof' of which string is, the musician receives a pop-up message with big font on the big display and/or a message on FC6 FC12 about the string n. to be corrected and how many cents is currently detuned:

5 ▲ 25 (meaning that 5th string is currently 25 cents higher)
or
A2 ▲ 25 (meaning that A2 string is currently 25 cents higher)

6 ▼ 10 (meaning that 6th string is currently 10 cents lower)
or
E2 ▼ 10 (meaning that E2 string is currently 10 cents lower)

N. of string is independent from non standard guitar tuning although maybe (I don't think so) it could be necessary to declare the tuning of each string for an easier processing.

In this way, the musician can quickly correct as much as possible that string tuning, in some case even while playing. like most gifted guitarists already do "by ear".

Edit:
That Tuning Assistant shouldn't react in real time, but it prompts the worst tuned string only after enough confidentiality being reached.
Confidentiality and tuning avg. deviation could be stored simply with a triplet of data for each string:

String no.n. of measurementsaverage tuning deviation
14+1.02
210- 1.85
312- 0.23
417+ 5.45
514+ 7.12
620+10.78 (above the user's set threshold of 10)
(7)......
(8)......

In that case it will prompt: 6 ▲ 11 cents

Some hint:
  • ignore analisys while bending notes, vibratos, strong hits, only stable notes have to be considered
  • After every reliable measurement it adds it and recalculate a weighted average deviation
  • Whenever it reaches a set number of consistent measurements and the deviation is over the set cents limit it prompts the worst case

  • in a standard 6 strings guitar tuning, it's easy to detect E2 F2 F#2 G2 G#2, only one string can play these notes
  • when playing a chord with all 6 strings is probable that the lower note is the E string, the slightly higher one is A strings etc etc.
  • when playing a 3 strings power chord and the lower one is E2 F2 F#2 G2 G#2, the other two string are probably A and D
  • The user must probably set the n. of strings, I image from 4 (bass) to 8, and the minimum deviation that will cause the prompt on the display
  • ...

This type of pitch detection will open the possibility also to develop an additional box... GUITAR to MIDI, allowing the FM3/FM9/AXE to send MIDI notes to the MIDI OUT connector and drive external devices (synthetizers, vocal harmonizers...)
https://forum.fractalaudio.com/threads/guitar-to-midi-block.207004/

NOTE:
TC ELECTRONIC "Polytune" works strumming all strings loose, I don't think this concept will violate TC ELECTRONIC patents:
this one doesn't show the tuning of all strings and works normally playing even solos and/or rhythm parts, not necessarily of all strings, finally it shows only the eventual tuning problems of one string at a time.
 
Last edited:
unix-guy said:
Aside from probably consuming a fair bit CPU, I suspect it would need a lot of additional storage to contain the analysis data.

Also, I think it would be incredibly complex to filter out pitch bends, vibratos, aggressive picking, etc.
Click to expand...

I have edited my post considering your doubts and added an algorithm hypothesis of how it might work in order to save memory (just 2 bytes and 1 integer per string).

Obviously only stable notes have to be considered, I think while playing there will be enough ones for our analisys.

I agree that the big difficulty is the polyfonic analisys of more strings in the same time, but in that case isn't necessary to be so precise as we could think, since the results considered will be the average of several measurements and there's no need for an immediate prompt.
 
Last edited:
sedorjulian said:
This all sounds kind of cool, but It’s kind of like you’re inventing a problem to solve.
Click to expand...

Have a look on my other posts: you will find another concept I described in details to FAS. Firsts reactions were "do you really need it? Don't seems it's a big issue that needs to be solved"
Now is considered one of the biggest improvements done recently by FAS. :D

The same happened with Korg, Roland, Creamware, SonicCore....
 
sedorjulian said:
I see a lot of wishes. I'm too lazy to search your profile. I feel like this is just unnecessary and complicated.
Click to expand...

I don't think it's complicated to use (just a prompt of very few BIG characters),
the algorithm design could be a little tricky, but not so much in the way I proposed it.

I'm confident it will be used by the more demanding people, wishing to tune strings without stopping to play.
I sometimes do it when I understand a string to be out of tune, just using my imperfect ears.

Among all improvement whishes I've got, I wouldn't put this in the first places of the road-map...
 
This idea is like an improved version of the Polytune, which graphically indicates individual strings' tuning for ?quick(er)? adjustment. Not to say that OP isn't already doing this but instead of relying on tech you could do some aural training and get that up to scratch.
 
NattyBar said:
This idea is like an improved version of the Polytune, which graphically indicates individual strings' tuning for ?quick(er)? adjustment. Not to say that OP isn't already doing this but instead of relying on tech you could do some aural training and get that up to scratch.
Click to expand...
It is similar to Polytune only in that it analyses all the strings,
but does not require all the strings to be strummed free, in fact it lets the guitarist play his parts normally:
in the meantime, the FAS device takes all the time needed to analyse and detect any out-of-tune strings and prompt the worst one.

I my gigs I have very rare moments in ~2 hours concert when I can activate the tuner, so this concept will help musicians like me in doing a "while playing adjustments".
Electronics can be much faster than my ears+brain 100% focused on the performance.
 
Last edited:
Kudos for the extensively thought through post :)

It would be a cool feature, but it would probably take quite a lot of work to implement, test and perfect something like this, in order to be sure that it was truly useful, reliable and didn't significantly impact existing functionality. I don't know what dev resources are like at FAS, but my guess is that this would probably get pushed a fair way down the list unfortunately.

My ears are pretty good, but I still use a standard digital tuner on the headstock and in combination that is enough for me. I just glance every now and then to confirm that my ears are "correct" and if I'm a fraction too far out visually, I silently chastise myself and make the required adjustment :laughing:
 
BobXX said:
It is similar to Polytune only in that it analyses all the strings,
but does not require all the strings to be strummed free, in fact it lets the guitarist play his parts normally:
in the meantime, the FAS device takes all the time needed to analyse and detect any out-of-tune strings and prompt the worst one.

I my gigs I have very rare moments in ~2 hours concert when I can activate the tuner, so this concept will help musicians like me in doing a "while playing adjustments".
Electronics can be much faster than my ears+brain 100% focused on the performance.
Click to expand...
Challenge with this as I see it however is that the tuner has no ability to understand what you’re doing in order to actually identify an out of tune string.

Some cases
  • Bends from micro to semitones
  • Vibrato fretting or whammy bar
  • Pushing strings out of tune from a heavy fretting hand

This also would assume that our guitars are in tune across the neck which we know isn’t true. What would the tuner consider its reference for being in tune?
 
NattyBar said:
Challenge with this as I see it however is that the tuner has no ability to understand what you’re doing in order to actually identify an out of tune string.

Some cases
  • Bends from micro to semitones
  • Vibrato fretting or whammy bar
  • Pushing strings out of tune from a heavy fretting hand

This also would assume that our guitars are in tune across the neck which we know isn’t true. What would the tuner consider its reference for being in tune?
Click to expand...

Thank you NattyBar for your comments, I considered the three points in my concept description, please have a look at it. Also consider that in this concept the tuner doesn't need to react in real time.

This means that:
1- it has plenty of time to gather information
2- enough time to detect real stable notes, ignoring bending, vibratos, vibrations and so on
3- have multiple confirmations that that string is out of tune before prompting it
4- perfect accuracy is not necessary, I agree that it's reached only by stopping to play and using the a standard real time tuner

You're right that some guitar may my be slightly out of tune across the neck, but considering all the points above it can be sorted out.

I know there are other challenges, but it is finally doable, strange and innovative. I think demanding musicians will appreciate it and competitors will then start copying it.

I had counter-reactions to obvious needs like my gapless switching concept (now done by FAS 🖤), in this case, with a much less obvious concept, I am prepared it won't be understood by the majority of people, until they see it in action. :tearsofjoy:
 
Last edited:
You must log in or register to reply here.
Back
Top Bottom