Melodeon.net Forums

Please login or register.

Login with username, password and session length
Advanced search  

News:

Welcome to the new melodeon.net forum

Pages: [1]   Go Down

Author Topic: Synthetic Dedic?  (Read 3262 times)

0 Members and 1 Guest are viewing this topic.

forrest

  • Other Places::
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1009
  • digits still functioning
Synthetic Dedic?
« on: January 27, 2011, 08:06:08 PM »

   This thread is to continue a thought generated on another thread having to do with the shift in pitch of reeds.
 
    After much head-scratching about the why's and wherefore's of the Dedic or Viennese tuning model, and having successfully tuned a box in this fashion, it ocurred to my simple mind that this could be approximated, for the simplicity of using the convenient layout of tuning software. It also presupposes that at a small level, errors and inaccuracies will be minute enough to pass unnoticed.
   The premise of Dedic tuning is that it's result is more accurate and true to pitch, given that we are tuning two or more reeds in tremolo fashion. The reason being, that if one reed is tuned at pitch and one off pitch (either + or -) then the resultant tone will be pulled away from pitch. Therefore, we must establish a median where the two reeds are tuned exactly X+ and X- the median, the median being zero, or pitch.
   This means that if we tune one set of reeds conventionally, (at zero), then tune the other set in varying degrees sharp (+) for tremolo, the resultant pitch perceived will be sharp by the amount of 1/2 of the difference between the zero and + reed. On a wet tuned box, this could be quite significant, on a dry box, it's a non issue.
  
  To the point: Let's say we have a linear beating pattern, with C4= 12 cents, and C6= 7 cents (the approx.range of the 'C' row on my G/C pokerwork). Assume that chords and basses are all tuned to zero, and one set of treble reeds are also set to zero.  If I play C4, it will sound [1/2 of 12 cents] sharp,  if I play the C6, it will sound [1/2 of 7] cents sharp. You can see that with this beating pattern, the error becomes smaller as we ascend the scale.

  If the reeds were to be tuned to the Dedic tuning model, then basses and chords would stay the same, and C4 would become -6 cents and +6 cents, C6 would become -3.5 cents, and +3.5 cents, etc. We're all still aboard are we?

    Consider the center of the range, C5, with an interval of approx. 10 cents. With conventional tuning, it would exhibit a 5 cent error compared to basses and chords. If we were to choose the value of 5 cents as an offset, and readjust the zero reeds all by -5 cents (basses and chords remain at zero), then our error at C4 is now only [1/2 of 12]-5 , and at C6 would be [1/2 of 7]-5. Please note that now, instead of having an error of +6 at C4, and +3.5 at C6, you have
+1 at C4 and -1.5 at C5. These are closer to the Dedic model, with less error than conventional tuning, and all we did was set the median to -5 cents. It's a compromise, but one that could easily be programmed into something like Dirk's. The value of the offset could be calculated and tweaked for varying beating recipes.

   Please feel free to shoot plenty of holes into this, because I have not attempted it, but am considering it for my Pre WW2 Erika. Intuitively, it seems sound. Feedback more than welcome ;D
« Last Edit: January 28, 2011, 03:34:32 AM by j.w.forrest »
Logged

Telemorris

  • Respected Sage
  • ****
  • Offline Offline
  • Posts: 405
  • C/F Preciosa
Re: Synthetic Dedic?
« Reply #1 on: January 27, 2011, 09:17:16 PM »

Thanks for starting this thread. I am very interested in the discussion!

I'm going to start some record keeping of my tuning efforts with my D/G Erica to generate some decent data that might help the discussion.
 
One observation about your compromise tuning of lowering the zero reeds by 5 cents. I think you're implying that both the zero reed is adjusted by -5 cents and the upper reed is adjusted down by 5 cents. Is this correct? Otherwise, if you simply adjust the zero reed down by 5 cents you'll also be increasing the tremolo by widening the gap between the two reeds.

One issue I have with Dirks is that the tuner reads the pitch of the lowest reed as the pitch of the note when playing both zero and the sharp reed at the same time. Dirk's Tuner does not "average" the pitches of the two reeds together and derive a pitch between the two, the pitch that we would "hear". I do wish there were a setting in Dirk's that allowed this since that's the assumption behind Dedic tuning. Hope that makes sense.

If anyone has figured out how to set the tuner to read a two reed note as 1/2 the distance between M- and M+, please let me know.

Logged
Ed Stevens
Bridgetown Morris Men
Portland, Oregon, US
Erica D/G, Presswood Pokerwork G/C, Preciosa C/F

forrest

  • Other Places::
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1009
  • digits still functioning
Re: Synthetic Dedic?
« Reply #2 on: January 27, 2011, 11:14:09 PM »


......................I think you're implying that both the zero reed is adjusted by -5 cents and the upper reed is adjusted down by 5 cents. Is this correct? Otherwise, if you simply adjust the zero reed down by 5 cents you'll also be increasing the tremolo by widening the gap between the two reeds.



Yes, that was the intent. I considered revising the post, but it had already been read several times.



One issue I have with Dirks is that the tuner reads the pitch of the lowest reed as the pitch of the note when playing both zero and the sharp reed at the same time. Dirk's Tuner does not "average" the pitches of the two reeds together and derive a pitch between the two, the pitch that we would "hear". I do wish there were a setting in Dirk's that allowed this since that's the assumption behind Dedic tuning. Hope that makes sense.

If anyone has figured out how to set the tuner to read a two reed note as 1/2 the distance between M- and M+, please let me know.



  If you observe the graphic depiction on the bottom of Dirk's display, it shows the relative positions of two or more notes relative to zero. If you wish to use Dirks for Dedic tuning, simply load your values on the beating chart with column one as the minus, column two at zero, and let column three be the actual reed two.
 
     Try this experiment: use a guitar tuner or some other simple single pitch tuner.
Play a treble note. Observe what the note indicator does. My Korg guitar tuner shows an oscillation, with the average being sharper than pitch, roughly halfway between notes. Not scientific, I'll agree.
« Last Edit: January 28, 2011, 03:35:34 AM by j.w.forrest »
Logged

Telemorris

  • Respected Sage
  • ****
  • Offline Offline
  • Posts: 405
  • C/F Preciosa
Re: Synthetic Dedic?
« Reply #3 on: January 27, 2011, 11:55:39 PM »

I've been setting reed one to M- and reed two to M+ with reed 3 set to zero. Then I set the Asymmetry to -150 to put reed three actually on zero in the graph. I'm basing that choice on the fact that I'm tuning a two voice instrument and need reeds one and two to be equally distant from zero. This way I set the tuner to read two reeds only, not three. I think your setup requires the tuner to be set up for three reeds but it only registers two reeds. Haven't tried that yet.

I'll go back and look at what happens to the readings when you set up the way you suggest.
Logged
Ed Stevens
Bridgetown Morris Men
Portland, Oregon, US
Erica D/G, Presswood Pokerwork G/C, Preciosa C/F

forrest

  • Other Places::
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1009
  • digits still functioning
Re: Synthetic Dedic?
« Reply #4 on: January 28, 2011, 03:28:40 AM »

You might try pretending that you have a three reed box, M-, M0, and M+. Set your beating pattern just using M- and M+ on columns 1 and 3. Leave column 2 at zero. Set the tuner to only hear two reeds. The values relative to zero of M-, M+ should be shown. Make certain the beating interval between reed 1 and reed 3 are as you wish, in other words, not doubled.

 Also keep in mind that you may wish to temper certain notes of the scale, which adds yet another bit of detail...... ::)
Logged

forrest

  • Other Places::
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1009
  • digits still functioning
Re: Synthetic Dedic?
« Reply #5 on: January 28, 2011, 06:21:14 PM »

After re-reading my posts, I realized that they rambled on a bit, so here's an attempt to simplify:

An approximated form of Dedic or Viennese tuning could be constructed using the conventional tuning method (one reed at zero; one reed sharp), by applying an offset value (flattening) both reeds by the amount of: (the beating interval in the middle of the scale divided by 2). This would actually decrease the perceived error on the bottom of the scale to a small plus value, and render a small minus error at the upper end of the scale. Mid scale notes would show very minute error.

   Advantage: Perceived errors at hi and lo end would be diminished significantly, and the body of the scale would approach Dedic or Viennese tuning, without the necessity to maintain the symmetry of that tuning model, and could be more easily employed by tuning software.
 
   Please remember that this is only a hypothesis, has not been tried out to my knowledge. Comments and criticisms welcomed! If this is idiotic, I wouldn't want to be the last one to know! ;D
Logged

Theo

  • Administrator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 14475
  • Hohner Club Too
    • The Box Place
Re: Synthetic Dedic?
« Reply #6 on: January 28, 2011, 06:36:34 PM »

If I understand you correctly your proposal is to use the standar method, one reed at pitch , one reed sharp,  but with the whole lot dropped in pitch by a constant amount?
Logged
Theo Gibb - Gateshead UK

Proprietor of The Box Place for melodeon and concertina sales and service.
See my YouTube channel  for videos on repair and tuning

forrest

  • Other Places::
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1009
  • digits still functioning
Re: Synthetic Dedic?
« Reply #7 on: January 28, 2011, 07:33:10 PM »

If I understand you correctly your proposal is to use the standar method, one reed at pitch , one reed sharp,  but with the whole lot dropped in pitch by a constant amount?

Yes, Theo. Would probably be more easily visualized by use of a graph, which I am not capable of furnishing :'(
Logged

Theo

  • Administrator
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 14475
  • Hohner Club Too
    • The Box Place
Re: Synthetic Dedic?
« Reply #8 on: January 28, 2011, 07:46:03 PM »

In which case the tuning errors on intervals played on the treble side would be exactly the same as in standard tuning.

My feeling is that this is a solution looking for a problem.   Viennese tuning takes a little longer to do than "standard" tuning, and requires a little more care, but its not inherently problematic. 

What Telemorris reported earlier in this thread (reeds appearing to change pitch some hours after being tuned) seems to be a problem of either (a) reeds actually changing in pitch over a period of hours after tuning, or (b) reeds appearing to change in pitch when measured in a particular way. 
Logged
Theo Gibb - Gateshead UK

Proprietor of The Box Place for melodeon and concertina sales and service.
See my YouTube channel  for videos on repair and tuning

Telemorris

  • Respected Sage
  • ****
  • Offline Offline
  • Posts: 405
  • C/F Preciosa
Re: Synthetic Dedic?
« Reply #9 on: January 28, 2011, 08:31:43 PM »

Last night I gathered some data about how my reeds are tuned at the present moment. In the process I discovered an interesting phenomenon that might be contributing to my problem of drifting pitch. I discovered that I needed to sound the reed I was testing repeatedly before it settled down to a reliable reading. Almost as if the reed needed to warm up its voice for a bit before it would "sing" properly.

Dirk's tuner is so sensitive that it reads changes down to the tenth of a cent. My Korg OT120 tuner did agree with Dirk's tuner quite well but the accuracy of the reading on the Korg was to within plus or minus one or two cents. I'm starting to suspect that when the Korg tells me that the concert pitch reed is in tune and Dirk's is +0.8 cents that I'm close enough. Shaving 0.8 cents worth of metal off the reed is probably going to be difficult to achieve.

Theo, I decided to try to tune this instrument in standard Hohner tuning, one reed at concert pitch (M0) and the other sharp by the amount of standard Hohner tremolo for that pitch (M+). I am very interested in Dedic tuning and plan to migrate the tuning of this instrument to Dedic, however I wanted to give myself some practice at tuning with an ear to accuracy. Given that the reeds I just installed were pitched slightly above the standard Hohner tuning, I can always tune the reeds lower without compromising the reed thickness.

Thread Drift...

Is there any information about how accurate the average human ear is when listening to two notes that are slightly out of tune? Can the average person actually hear a difference of 1 or 2 cents between two notes? I'm trying to understand the necessary tolerances here. When I was making telescope mirrors, I had a given tolerance for how good a mirror had to be. Beyond that wave rating, only super human eyes could see a difference and the work load would go up exponentially. It's the difference between "excellent" and "perfect," where perfect is impossible to obtain.
Logged
Ed Stevens
Bridgetown Morris Men
Portland, Oregon, US
Erica D/G, Presswood Pokerwork G/C, Preciosa C/F

forrest

  • Other Places::
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1009
  • digits still functioning
Re: Synthetic Dedic?
« Reply #10 on: January 28, 2011, 09:00:25 PM »

In which case the tuning errors on intervals played on the treble side would be exactly the same as in standard tuning.
my thinking is that they would be shifted more closely to the position held by the bass and chords, (which would remain at zero), therefore the error margin is decreased


My feeling is that this is a solution looking for a problem.   Viennese tuning takes a little longer to do than "standard" tuning, and requires a little more care, but its not inherently problematic.  
 
 

Maybe I am overthinking this. I just thought it could perhaps be useful as a shortcut.






Logged

Gandy

  • Respected Sage
  • ****
  • Offline Offline
  • Posts: 407
Re: Synthetic Dedic?
« Reply #11 on: February 02, 2011, 02:19:06 PM »

my thinking is that they would be shifted more closely to the position held by the bass and chords, (which would remain at zero), therefore the error margin is decreased

But there would still be a difference across the treble range.  So for example if C4 was true to the bass and chords, then C6 would be flat.
Logged
Tony S

forrest

  • Other Places::
  • Hero Member
  • *****
  • Offline Offline
  • Posts: 1009
  • digits still functioning
Re: Synthetic Dedic?
« Reply #12 on: February 02, 2011, 06:43:21 PM »

my thinking is that they would be shifted more closely to the position held by the bass and chords, (which would remain at zero), therefore the error margin is decreased

But there would still be a difference across the treble range.  So for example if C4 was true to the bass and chords, then C6 would be flat.

I took Telemorris' lead, and picked up a humble pencil and paper to illustrate the idea. Here are some comparative graphs. They all show the same beating intervals; the chords and basses in each example would be tuned to zero. Temperament is not considered here. Note that I have arbitrarily located the center of the scale, C5, to be true, thereby dividing the perceived error and mitigating it significantly.
Compare with perceived error on conventional tuning.



« Last Edit: February 02, 2011, 06:47:20 PM by j.w.forrest »
Logged
Pages: [1]   Go Up
 


Melodeon.net - (c) Theo Gibb; Clive Williams 2010. The access and use of this website and forum featuring these terms and conditions constitutes your acceptance of these terms and conditions.
SimplePortal 2.3.5 © 2008-2012, SimplePortal