Welcome to our forum. A Message To Our New and Prospective Members . Check out our Forum Rules. Lets keep this forum an enjoyable place to visit.
Private messaging is working again.
Log In
Register
Home
Topic RSS
(2 votes) 
Offline
Did someone say "theremin" @Peter LOLOL
Aye, indeed then, we'll see how this first simple mini-project on the Arduino goes (well, I know already, it'll work - I just need the time to do it - I mean I'm only talking a morning to do it and get it installed where it will do its job - but we're in apparent lock-down and my good lady insists on getting the re-decoration done LOL )
Well, my plan now is for a second little project, and I might just build it from the ground up using the DIP Atmel device rather than a prebuilt Ard board, but we'll see....
It will use a couple of light-dependent-resistors - the first to vary frequency, the second - well - some realizable effect - maybe volume, maybe (if there's enough RAM) an echo/reverb, maybe something even more interesting !!! I've messed with LDRs way back for something "vaguely" similar (not music related) and had a lot of trouble due to the highly non linear - well logarithmic really I guess - response. But I have a plan for that .. LOL
Oh - something else came to min - I think it was @ABitRusty, yourself Peter and @Irv - yeah the Intonation Correction Device ( heck, I just love the idea !!! Evil grin ) - yeah - I've looked at a lot of violin sound samples (many on the Octave strings investigation, and some on a normally strung violin) - and - yes - I forget who mentioned this - but sure, the signal is very complex with many many harmonics -so outright "level/peak" detection etc (as a starter for identifying the frequency of interest) must be real difficult especially on the G - on ALL of my instruments, looking at an open G (and a few tones up to D) the fundamental is generally of lower amplitude than the first harmonic ( and on the octave stings, the fundamental was TINY, it being the fourth or fifth harmonic that had peak amplitude ) I can see a lot of DSP issues sorting that sort of stuff out - and I won't be going there LOL - hang on - let me see if I still have a trace from one of my investigations - yah found one - this was using octave strings with a "cable tailpiece" from Irv - part of the reason for the amplitude variation of the harmonics comes down to the nature of the resonance points of the instrument itself - the violin is not designed to have body resonances down at G2 ( around 98Hz ) and the level down there is minimal. But, as you can see, even the G3 is significantly lower than it's harmonics (which is what I see on a normally strung fiddle - i.e just ignore the tiny tiny 98 Hz G2 - the rest of the trace is pretty similar !
I seriously recommend not copying my mistakes. D'oh - 
Please make your own, different mistakes, and help us all learn :-)
Regulars
Online
Some genuine theremins can be exceptionally expensive, I'm told.
But you can get them for what's probably a reasonable price on Amazon (I'm not especially tempted, though).
Perhaps my informant felt the Etherwave one was exceptionally expensive. Perhaps it is.
Andrew
Verified human - the ignominy!
Regulars
Offline
BillyG said
Did someone say "theremin" @Peter LOLOL...
It will use a couple of light-dependent-resistors - the first to vary frequency, the second - well - some realizable effect - maybe volume, maybe (if there's enough RAM) an echo/reverb, maybe something even more interesting !!! I've messed with LDRs way back for something "vaguely" similar (not music related) and had a lot of trouble due to the highly non linear - well logarithmic really I guess - response. But I have a plan for that .. LOL
Mine uses an LDR for the volume, pinching off the gate of a FET to do the job.
The pitch generator is (relatively) faithful to the Lev Termin original in that it mixes a pair of HF oscillators and take the beat note (audio frequency) and sends it to the amplifier. However, the oscillators are 24 MHz crystal-controlled, and in spite of this the antenna pulls the control oscillator sufficiently to give 2 Hz to 5 kHz range. Crystals are proportionally more pullable as frequency increases, rough maximum is 1000 ppm.
The frequency / distance response is logarithmic, and I'm OK with that because it feels the same as a boatswain's call's response to the player's right third finger. Conventional theremins have linearising reactances in series with the antenna to straighten the curve.
Have fun with it: theremins are nothing if not fun.
Peter
"It is vain to do with more that which can be done with less" - William of Ockham
"A crown is merely a hat that lets the rain in" - Frederick the Great
Offline
Nor me @Gordon Shumway 
I seriously recommend not copying my mistakes. D'oh -
Please make your own, different mistakes, and help us all learn :-)
Regulars
Offline
@Billyg @Peter Electrically speaking because I haven't investigated myself..Will a Mic say a sm57 (dynamic instrument type) help with knocking down sone of the amplitude of the harmonics? I know we can hear them and its part of the sound a violin generates but ..and for this youd probably want the worst mic you could get for this reason.. maybe the Mic won't pick those up as strong as the note youre trying to play. Also..and this gets into what @peter was saying about needing extra chips and hardware to do this justice..maybe hack a tuner and use what it is picking up instead of a Mic. I mean..the chromatic tuners will show what note youre playing.
Regulars
Offline
I want to expound on a comment by @ABitRusty (which I guess is another way of saying he beat me to it). Harmonics do not sway my ability to tune bass notes as long as I use a pickup as part of the tuner circuit, even on the c note of a cello. So the electronics of a tuner must be able to do that filtering. Perhaps a little reverse engineering on the chip types of the circuit board is reliant.
Success is the progressive realisation of a worthy ideal. —Earl Nightingale.
Regulars
Offline
GregW said
@Billyg @Peter Electrically speaking because I haven't investigated myself..Will a Mic say a sm57 (dynamic instrument type) help with knocking down sone of the amplitude of the harmonics? I know we can hear them and its part of the sound a violin generates but ..and for this youd probably want the worst mic you could get for this reason.. maybe the Mic won't pick those up as strong as the note youre trying to play. Also..and this gets into what @peter was saying about needing extra chips and hardware to do this justice..maybe hack a tuner and use what it is picking up instead of a Mic. I mean..the chromatic tuners will show what note youre playing.
Cleaning up a violin signal is as complicated (and contentious, no doubt) as the work done on electric guitars. If you cut off the harmonics of your first octave, you lose signal in the subsequent octaves. A violin has an operational compass of a little over three octaves, so unless you're happy to mumble away on the G and D in first position, filtering the signal will make the upper register sound thin and woolly.
<sacrilege>Perhaps what's needed is a segmented bridge, with a piezo pickup for each string and strain gauges under the fingerboard to detect where the player is stopping notes. The individual pickups for each string feed four individual channels which have their own baseline response, and the strain gauges inform the filters to move their cut-off.</sacrilege>
I agree with @Irv on this: how about sacrificing a tuner on the altar of reverse engineering, and see how it's done? My bet is that it's a DSP filter in an AVR or ARM device, and we'll have to do our own coding (they'll have locked the chip...).
Peter
"It is vain to do with more that which can be done with less" - William of Ockham
"A crown is merely a hat that lets the rain in" - Frederick the Great
Offline
Yes @ABitRusty and @Irv - my best guess is that the specific tuner devices I have seen (like the Snark- the one I have seems to be violin specific - i.e. it only shows G D A E and NOT G3 D4 A4 E5) are probably pretty simple with no DSP involved. For the violin tuner I would *guess* that the circuitry is really simple - a little bit of analog op-amp band-pass filter (say from 150 Hz (a reasonable bit below G3 ) to 990/1000 Hz (a reasonable bit above E5), largely eliminating most harmonics - followed by a simple zero crossing detector which just triggers a counter. Dependent on the counter frequency, the count between zero crossings will be detected, leading to a direct conversion of the measured interval between crossings to frequency. Not a DSP device to be seen..... (well, my guess on that one, for that type of tuner - there are of course other "smart types" in mobile devices and they almost certainly do some form of dsp - in fact maybe even as little as some dsp filtering to replace an analog bandpass filter, to get clean zero-X and may still eventually just measure the interval between zero cross. LOL that's a guess as well, but sounds reasonable.... !)
I seriously recommend not copying my mistakes. D'oh -
Please make your own, different mistakes, and help us all learn :-)
Offline
ROFL @Peter - we responded together ! Interesting what you say. The rationale for my guesswork was based on two separate actual tuner devices I have seen and used. The first was Korg (which I no longer have) and was intended as a guitar tuner - but - that was way before the days of ARM or other highly integrated electronics (in fact, at the time, there was little in the way of LSI at all). But yes, the approach you mention is of course possible today!
I seriously recommend not copying my mistakes. D'oh -
Please make your own, different mistakes, and help us all learn :-)
Regulars
Offline
It's ironic.
I left radio & electronics behind to become a fiddler, and now I suppose the old adage is so true:
"You can take the boy out of the radio shack, but you can't take the radio shack outta the boy."
Peter
"It is vain to do with more that which can be done with less" - William of Ockham
"A crown is merely a hat that lets the rain in" - Frederick the Great
Regulars
Offline
I have a couple of cheapy clip on tuners that will read any note..not just the open string.
as far as fundamentals and harmonics..we are talking about reading 1 note..not faithfully reproducing the note for a listener.
Regulars
Offline
it suddenly occurs that our device is this..ROFL
Offline
Aye @ABitRusty - as you say
I have a couple of cheapy clip on tuners that will read any note..not just the open string.
Yeah - I didn't mean to mislead by mentioning just the G D A E - what I was really getting at - it shows no indication of WHICH G D A E octave. Yes the ones I mentioned are chromatic, with indications of cents-high or low - but tailored to measure across a specific range - in this case the G D A E for violin... My bad...
I seriously recommend not copying my mistakes. D'oh -
Please make your own, different mistakes, and help us all learn :-)
Regulars
Offline
@Billyg the models I have put the number e.g. G3 A4..etc.. which practically speaking is useless when you play up to speed..but they can do it. Here is another thought on the software end of things. programs such as melodyne are very good at detecting notes and then showing how far off in cents that you are. Even can tell what key you are likely in. I wonder if there is a developers SDK that unlocks a little of how they do it. Im surface level familiar with coding but you might have a better shot at exploring that and getting some deeper insight on whats going on under the hood so to speak. for what its worth I believe they have some trial versions of it.
Regulars
Offline
@peter et al. I created a thread here a while back on mutant bridges. There was one that would likely isolate between bass and treble strings. I reproduced the pictures here so that you could avoid use of search feature.
Success is the progressive realisation of a worthy ideal. —Earl Nightingale.
Regulars
Offline
@Irv - I have no idea what my 'segmented bridge' would eventually look like (or behave), but unlike your mutants, they would most likely have four individual feet, with the violin itself being necessarily modified to accommodate this (different bass bar and sound post designs). At this point, it's arguably no longer a violin, much like my solid fiddle (it has a monopod bridge with a single large piezo pickup). It may look like a violin, it may even sound like one, but it would have four separate, disconnected personalities in its several and severed strings.
If we move too far from Cremona, I guess we're not in Kansas anymore.
What happened to the sound of the violin which had the bridge lobotomy? Did you make recordings and do you recall anything about the sound? I imagine the changed moment between the two sides had an influence.
Peter
"It is vain to do with more that which can be done with less" - William of Ockham
"A crown is merely a hat that lets the rain in" - Frederick the Great
Regulars
Offline
using reverse engineering seems useful for this doesnt it? why I brought up melodyne. at some point coding the logic will be needed. not saying its impossible from scratch, especially for pros..just something to consider for the time I suppose. although by the time you figure that out a person could be further down the line just doing it.
Regulars
Offline
GregW said
using reverse engineering seems useful for this doesnt it? why I brought up melodyne. at some point coding the logic will be needed. not saying its impossible from scratch, especially for pros..just something to consider for the time I suppose. although by the time you figure that out a person could be further down the line just doing it.
Re-writing code is an old evil and we seem doomed to repeat it.
It may be best to look for an appropriate DSP library for the device we intend to use, and exploit it.
Peter
"It is vain to do with more that which can be done with less" - William of Ockham
"A crown is merely a hat that lets the rain in" - Frederick the Great
Regulars
Offline
Peter said
Perhaps what's needed is a segmented bridge, with a piezo pickup for each string and strain gauges under the fingerboard to detect where the player is stopping notes. The individual pickups for each string feed four individual channels which have their own baseline response, and the strain gauges inform the filters to move their cut-off.
Would this be close to the type of bridge you're looking for? Note that the bridge is segmented between strings.
Regulars
Online
@Peter - piece of friendly advice from someone who has similar tendencies - don't fall into the trap of letting theory override practice!
Andrew
Verified human - the ignominy!
2 Guest(s)
ELCBK
Mouse
Fiddlerman
Jols
