Intonation

[femptofarad]

I bought 2 expensive DDS chips ($15 a pop, MSOP10's? sheesh) with the grand idea i might be able to do this; take an audio signal and filter out two fundamentals, then take those fundamentals and plot them. Lissajous pattern. When you "set up" a musical instrument that has a movable bridge (e.g. some guitars, violins, violas). When you "set up" a guitar, you use your ear to track beats created by the slight difference in frequency of the notes you are comparing. My theory is if you can SEE a pattern your brain can react/remember better than listening for beats. I can discern beats down to around 10 seconds, then my ears/brain give up.  If I can plot them on a display like a Lissajous pattern? Immediate feedback!  I am NOT trying to build a guitar tuner, you can get excellent guitar tuner on your smart phone for free. The job to be done here is to adjust the instrument so that the "intonation" is as correct as possible. This is intended to be a professional measurement instrument (pun) for serious musical instrument maintenance professionals. I want to build a circuit that compares the two signals (after really tight bandpassing) and draw a Lissajous pattern, the DDS chips will be used to produce good sinusoids, one each for the two filtered signals. I am reading up on DSP just in case this idea does not work. I think I can do it without DSP. The DDS device I like is the Analog Devices 9833, check out the datasheet. Thoughts?

[ataradov]

I think you need give us more information.

What are "two fundamentals" of the same signal?

[IanB]

I think the idea is you have two strings and you want to tune the second string to match the first. When there is no beat between the two strings the relative tuning is correct.

Where I think more information is needed is where the band pass filter comes in. If you apply a very narrow notch filter it is going to select for a particular frequency. However, if you are adjusting the string tension the frequency will vary, so it will not line up with the notch.

What I think is needed is for the OP to assume nobody here is a musician and to explain in layman's terms exactly what job the circuit needs to do. Proper advice cannot be given without proper understanding of the problem statement and the requirements.

[Brumby]

I've had several 'interesting' conversations (with one guitarist in particular) about 'tone'... guitars, amps, speakers and everything in between.  (Anybody who's ever come across a serious guitarist will know exactly what I mean.)

Nevertheless, I too am a little lost as to what the objective is.  You seem to have jumped right over fundamental theory and just gone for a solution - but, from what you have supplied, the concepts in that solution don't make sense.  It may well be you have simply not expressed yourself in a way we can understand - or it may be you have formed an idea of something that either won't work or won't do what you want.

This:

What I think is needed is for the OP to assume nobody here is a musician and to explain in layman's terms exactly what job the circuit needs to do. Proper advice cannot be given without proper understanding of the problem statement and the requirements.

[jwm_]

So, it sounds like you want to take the two strongest peaks from the spectral analysis of the signal, recreating them as sine waves (preserving phase?) then plotting them x/y on an oscilloscope, to get instant visual feedback of the relationship between the two strongest signals?

You will probably need a DSP for this but perhaps there is a cool analog trick i don't know about, but you may want to just brute force it on a computer with an FFT to see how it looks before doing hardware. I think the algorithms might be a little complicated, when the two main frequencies get close enough together, suddenly they will seem like one peak to the algorithm and some lesser harmonic will take over as the second. Not an issue if your notes are not the same but rather one is a harmonic of the other, like the 3rd or something so perhaps not important to you.

Hmm.. thinking about it, the two strongest signals probably isn't great, a harmonic of a louder instrument might be stronger than the primary of a more quiet one.. perhaps the two frequencies that show interesting yet distinct "harmonic structure" so likely to be different instruments ... not sure what that would mean actually.... Perhaps there is existing research on this.

[john_p_wi]

How would this be any different than a good strobe tuner that has been used by professionals for years?  It is certainly "visual" and portable too.  Even something as simple as this works very well from my experience:

https://www.turbo-tuner.com/

[RobertHolcombe]

I've only skimmed over this briefly, but I didn't see much discussion of practical intonation measurement and/or why you would make such adjustment.

For a stringed instrument, intonation is the accuracy of pitch at any given point along the scale length (maximum open string length from bridge saddles to nut), but when dealing with fretted instruments it is specifically referring to accuracy at each fret. Once an open string has been brought to its intended pitch, intonation is measured by comparing the pitch of the fretted note at the 12th fret (half the scale length), to the 2nd harmonic of the open string (12th fret harmonic). If intonation is correct these should be the same pitch - intonation adjustment is performed to achieve this by aligning the center of the string with the 12th fret. Adjustment is usually made at the bridge saddle, either shortening or lengthening the string to achieve correct intonation at the 12th fret.

The measurements cannot be made simultaneously because they are performed on a single string, and you can't play the fretted note and open string natural harmonic at the same time. I'm not sure what OP intends to do 

[Brumby]

I'm not sure what OP intends to do 

I think that's something a few of us are wondering as well.

[ataradov]

For a stringed instrument, intonation is the accuracy of pitch at any given point along the scale length

This is not true. Most classical instruments let you tune only one parameter - tension of the string and the rest of them are defined by the design.

If you look at virtually any electric guitar, you will see classical tuning pegs, of course. But there are also "intonation" tuners at the saddle (bridge) that let you adjust the length of the string.

In theory, with ideal strings, it does not matter. In practice, it does.

[IanB]

intonation adjustment is performed to achieve this by aligning the center of the string with the 12th fret. Adjustment is usually made at the bridge saddle, either shortening or lengthening the string to achieve correct intonation at the 12th fret.

If you look at virtually any electric guitar, you will see classical tuning pegs, of course. But there are also "intonation" tuners at the saddle (bridge) that let you adjust the length of the string.

This is exactly what RobertHolcombe said. I am not sure why you are disagreeing?

[ataradov]

I am not sure why you are disagreeing?

I guess I just got lost in the text. In that case we are in agreement and OP is nowhere to be found, so mystery of plotting things on the scope and why such expensive components are involved, will remain unsolved.

[Cerebus]

... I can discern beats down to around 10 seconds, then my ears/brain give up. ...

I'm quite impressed if you can; I reckon I can't pick out a 'beat' of longer than 3 seconds. Either way that's a 0.1 or 0.33 Hz difference.

I'm an amateur acoustic guitar builder so I know exactly what you're trying to do. The thing is I don't think you've fully thought through whether you need to do it. I'd have to do away and do some maths or make some measurements to be sure but instinct suggests that the whole intonation process has more 'slack' in it in terms of pitch differences than the precision that's achievable by ear.

What I mean is probably better explained with a more concrete example, but without real numbers. Let's assume a guitar. You start off by adjusting the intonation so that the 12th fret gives an exact octave when fretted - compared either to the open string or a plucked harmonic at the 12th fret. Now if you fret that string at the 5th fret you should get a perfect 1/3 but in practice the intonation is a little off. You re-adjust the intonation to balance the error at the 5th and 12th frets to be acceptable. What I'm calling 'slack' above is the difference between the exact results for the 5th and 12th frets and the compromise results for both. I strongly suspect that the frequency error induced by the compromise intonation (say 2-3 Hz at concert A) is greater by an order of magnitude that the error that is introduced by inaccuracies of tuning by ear and beats. Even if the compromise intonation introduces only a 1 Hz error that's still much more than the 0.1 to 0.33 Hz error introduced by tuning by ear.

[Cerebus]

For a stringed instrument, intonation is the accuracy of pitch at any given point along the scale length

This is not true. Most classical instruments let you tune only one parameter - tension of the string and the rest of them are defined by the design.

Erm, no. If you actually look properly at members of the violin family you will find that the bridge is not fixed to the instrument, it is held in place by string tension alone. The angle the bridge makes with the strings can be adjusted and this is what is done to correct the intonation. Also the height of the bridge under each individual string can be adjusted to compensate for the varying diameters of the strings (also at the nut); this is done by the more drastic move of filing away material from the saddle. The saddle is a piece of, typically, bone that sits on top of the wooden bridge. I have a little (but expensive) set of Japanese files designed expressly for this purpose. On members of the guitar family the basic angle of the saddle partly sets the intonation but precise intonation is set by, again, filing away material but this time longitudinally as well as vertically. On electric guitars the intonation is set by adjusting the guitarist with a 4 lb lump hammer.

In theory, with ideal strings, it does not matter. In practice, it does.

It matters in theory too. A simple model accounting for the tension and elasticity of the string along with the height of the string above the finger/fret board and string diameter will quite clearly show that an adjustment for intonation will be necessary. Then you can add nasty variables like movement over the bridge and nut, effect of wound versus plain strings, fret widths, fret heights, finger pressure, volume of the note being played etc. etc.

[Brumby]

It matters in theory too. A simple model accounting for the tension and elasticity of the string along with the height of the string above the finger/fret board and string diameter will quite clearly show that an adjustment for intonation will be necessary. Then you can add nasty variables like movement over the bridge and nut, effect of wound versus plain strings, fret widths, fret heights, finger pressure, volume of the note being played etc. etc.

I always appreciated the mathematics of a fretted instrument and the importance of precise placement of the frets being a fundamental requirement, but while some of these other parameters weren't really in my conscious appreciation, when you mention it, their contribution is somewhat obvious.


However, are we any closer to a solution for the OP? ... and has anyone seen them?

[SL4P]

I know bugger all about all aspects of this discussion, but if I read the OP correctly, he's looking for a way to identify the tonal variations caused by EITHER adjusting the string tension... OR bumping the movable bridge position, and the associated tonal, resonant, or harmonic influences those changes may introduce which are beyond simple pitch.

Am I close?

[Cerebus]

I know bugger all about all aspects of this discussion, but if I read the OP correctly, he's looking for a way to identify the tonal variations caused by EITHER adjusting the string tension... OR bumping the movable bridge position, and the associated tonal, resonant, or harmonic influences those changes may introduce which are beyond simple pitch.

Am I close?

As I read it, it is all about simple pitch. He uses the word 'intonation' which is all about the instrument or player's pitch accuracy versus the desired pitch. It's a pretty specific word and widely understood. The terms I'd associate with things beyond simple pitch are things like 'timbre', 'resonance', 'tone', 'colour', 'tonal colour' or even just 'sound' (as in the 'sound' of a Les Paul guitar). Music is full of highly subjective adjectives and nouns but 'intonation' is one of the most precisely defined and largely objective terms.

Don't underestimate how important skilful setup of the intonation on a guitar is. If you pick up an average £200 acoustic guitar and a £1000 one the difference in tuning accuracy is sometimes quite painfully obvious and the difference is the time and skill that went into the 'setup' of the guitar, which is a post manufacturing stage. Guitars are made of wood and wood varies from plank to plank so it doesn't matter how exact the manufacturing process is, some adjustment is needed once the whole guitar has been made, strung and given a while to settle down. The typical acoustic guitar soundboard is about the thickness of a couple of matchsticks or less and can be snapped with the fingers of one hand but has around 120 lb of force placed on it by the strings. That makes things move and each piece of wood moves differently.

[Brumby]

OK then...

Question: What is the specific source of each of the two signals mentioned by the OP?