Comparing the instruments

ed packard · From: Show Low AZ

There are lots of hand waving type discussions re the basic sound produced by a steel guitar. It seems about time to produce a documented comparison.

Jim Palenscar has a shop full of lap steels and PSGs in Oceanside Ca. We have had discussion on the subject. The net result is a plan for me to go to his shop, and we will do a repeatable instrumented evaluation of the available instruments.

The rough plan is to address harmonic content, and harmonic content vs time for the available instruments. These will be compared as a function of load. The evidence will be Frequency Spectrum Analyzer output in chart (graphic) form, hopefully an audio recording of what is submitted to the FSA, plus photos of each end of the instrument,including the pickups.

Maker, model, ser #, pickup resistance, neck length, string type, and such other info that can be obtained will be included.

If we can figure a way, a CD will be made that allows hearing the audio sample while looking at the FSA graphics.

This is a reasonably ambitious project and may get modified as we go along.

This is my excuse to leave the high country snow and get some California sunshine this winter.

Bob Kagy · From: Lafayette, CO USA

Good for you Ed; I think this kind of info would be great.

BK

Jim Cohen · From: Philadelphia, PA

I wish you all the best with this ambitious project, Ed. Please don't take offense at this, but I'm suddenly reminded of a wonderful quote by Albert Einstein:

Bobby Boggs · From: Upstate SC.

I have 3 Emmons guitars.All hand picked by me.Same models,same pickups.Yet they all sound just a little different.Unless you try 20 of each brand. What can you really prove?? Respectfully Bobby Boggs<

[This message was edited by Bobby Boggs on 14 October 2005 at 08:01 AM.]

Per Berner · From: Skovde, Sweden

This seems like a very good idea, to be able to hear several guitars recorded under identical conditions, AND with the graphs to clarify why they sound like they do. Somewhere down the line, the holy grail of tone may come within reach.

On a similar note: Carter's website has samples of Billy Phelps playing the same song with different pickups, also very interesting. The differences are subtle at first, but after listening to them side-by-side a couple of times the different personalities shine through.

ed packard · From: Show Low AZ

This little project will take a while. Looks like the idea splits the readers into at least two
groups…that is good, from the discussion should come ideas and maybe inspiration.

I used the techniques to be employed in the design of the BEAST to verify that the
structural changes produced acceptable (to me) results. Those that responded to the offer
placed on the Forum were sent the charts showing the results of pickup loading on
harmonic content, harmonic content vs. time, frequency spectrum change as a function of
pickup loading, etc for the BEAST and it’s prototype. The changes were obvious, and
repeatable…the method has a degree of validity so we plan to repeat it on other available
instruments.

Bob K…Thanks for the encouragement.

Jim C…If I recall correctly, the context of the quote was sort of a Theosophical bent by
AE. Nice play on words.

Bobby B…It is exactly the differences to which you refer that show up in the FSA charts. Proving something is not my intent…presenting the results of the experiments/instrumentation is. One may then look at the patterns and decide for themselves what the meaning is. Proof is a matter of persuasion, I am just presenting.

Per…If I recall, you were on the mailing list for the FSA work re the BEAST. I hope to take this run one more step. The sounds that made the charts will be recorded in such a way that they can be played back while looking at the chart to see the correlation…moreover they can then be played back thru your rig to see how they sound when you tweak to suit.

Jim Hinton · From: Phoenix, Arizona, USA

Hello Ed:

That's quite an interesting undertaking, and I wish you much success on it.

My concern for your project is this: can there be a corellation between the dynamic spectrum and a great sound?

I am assuming that one wishes to identify characteristics that are most pleasing to "their ear", and with the knowledge that you provide be able to select optimum brands of equipment that they would like before ever playing them.

The missing link to all this of course is the "master's touch". The skill and technique that the player provides.

We have all heard Jeff Newman play a dozen steel guitars, and get a great sound out of all of them.

In short, I believe you may be trying to quantify something which is so subjective that your results might not be as useful as you would hope.

But, being a technogeek myself, I hope you "Kick Ass and take names"!!

Jim

ed packard · From: Show Low AZ

Jim...The basic plan is simply to present the information re what the instruments/pickups do when excited in a given way. The information will be comparative. Judging re "goodness or badness" of the sound will be left for the user to do.

Might even provide an approach/information for those that would want to do PSG/pickup modeling software (sounds like a .......).

I was thinking more of plucking strings and taking note.

Bob Hoffnar · From: Austin, Tx

Ed,
The measurement that means the most to my ear is the overtone series. I can "see" the levels of the upper partials in my head when I hear a tone. Its the even overtones that need to be stong with the odd ones needed for color and texture. If a steel does not have stable enough even overtones the guitar is no good. I'll go out on a limb a little here and admit that I think most of the steels made really don't cut it. Then there are a bunch of steels that are strong with the even overtones but sound sterile because they don't blend the odd overtones in well. Interestingly enough it seems like over the years steel builders have gotten much better at stabilizing the even partials. Getting the right blend seems more elusive.

My opinion is not based on any scientific anything. I could be completely mistaken. Its just how I hear it. I'm interested in your experiment.

Bob

Jim Hinton · From: Phoenix, Arizona, USA

There are a number of knowledgeable audio engineering types on this forum that may have some observations of interest in this regard.

Sam Marshall, Brad Sarno, and Mike Brown come to mind. I'm sure there are a lot more, but I'm not as familiar with their names.

How about it "audio guru's"? Got any good ideas for our colleauge here?

Dave Boothroyd · From: Staffordshire Moorlands

I know what you are saying Bob. I have seen results comparing perceived "tone" with harmonic content. You are right in saying that a good blend of even harmonics sounds smooth, but that very few people want smooth. Odd harmonics add that edge- a touch of savoury with the sweet, but it's even more complicated than that, becuse the majority of listeners do not want the blend of odd and even harmonics to be the same across the whole pitch range of the instrument. Think of a Clarinet- it has a completely different tone in its upper register to it's lower one. Human voices are even more variable.
It will take a lot of research to find out why some instruments pull at your heartstrings and some leave you cold- but I think it is the change of tone across the pitch range which is an important factor.
That is why the player's style is so distinctive.A good player will adjust the tone to suit the note and the song. To get it down to the instrument alone, you'll have to have it played by a machine.

------------------
Cheers!
Dave

[This message was edited by Dave Boothroyd on 14 October 2005 at 08:50 AM.]

James Quackenbush · From: Pomona, New York, USA

Ed,
I think this is a GREAT idea , and something that has needed to be addressed for a long time ....I can remember early on when first playing pedal steel asking questions about the various tones and timbre of various steels and pickups ....I would say that 85% of the answers I got were "the tone is in the hands " ....While the hands do play an important role in the overall sound of the instrument , the construction, and the electronic's weigh in just as heavily ...There are players out there that can make any steel sound good, and can play just about any steel to sound like another , but for the most part, there are more of us that can't ..... Again , there are different sounding instruments of the same brand, but there IS a basic overall tonallity that is common between all of them ...For example if I asked which steel is briter in tone, with the same settings on the amp to your ears, a Mullen, or an Emmons P/P ? ..... If there was no difference in pedal steels, and no difference in pickups, there would be a lot of people out of business !
Good for you Ed, I look forward to your results , and commend you for taking on this project ......Sincerely, Jim

[This message was edited by James Quackenbush on 14 October 2005 at 08:58 AM.]

Donny Hinson · From: Glen Burnie, Md. U.S.A.

I think it's a great study, and I'm curious as to just how much harmonic content there can be in an analog signal. I'm also curious to know how a player (Buddy, for example) seems to have a consistent sound, no matter what brand he plays. In other words, can the player (by virtue of his technique), affect the harmonic content? What about the amp? Does that affect what we hear, harmonically? (The "tube heads" swear that's really significant.)

Lastly, if we show that two signals are identical, or can be made identical, what do we do when someone says they still sound "different"?

Food for thought, anyway.

Jim Hinton · From: Phoenix, Arizona, USA

This is starting to sound like quantum physics.

It's gonna get deeper and higher before it's all over I suspect.

ed packard · From: Show Low AZ

The "human factor" in the sound of the instrument is a large variable. How and where the strings are excited/picked, along with the material/gauge of the picks change the way the strings vibrate. Pickup type and location further interpret these vibrations. Then pickup load further modifies the spectrum.

The approach to be taken in gathering the data will consider these issues by exciting the strings in a specific manner at chosen locations with a specific material.

The amount of odd and even harmonic content in a strings vibration can be controlled to a large extent by picking location and method. An on line search via Google or sim' will produce a wealth of info on string vibration.

Scott Swartz · From: St. Louis, MO

Another important factor is the decay of the harmonics relative to each other as the note is fading out, I'm sure Ed will be measuring that also.

If you want to hear how different partial combine, check out
http://www.blackstoneappliances.com/dist101.html

There is a cool sine wave addition tool there that allows you to combine partials any way you like.

ed packard · From: Show Low AZ

Scott...right you are...that is the "harmonic content as a function of time" test.

Inasmuch as I need to generate the test method(s), and documentation to go with it/them, I might as well manifest my ignorance here so that objections/corrections/alternatives can be brought up and discussed by the interested Forum members.

It is the intent that:

Each instrument will be identified by maker, serial#, body materials, tuning, neck length, pickup maker where known, pickup resistance, switch settings, string types where known, photos of each end of the instrument to show pickup location/changer fingers/rollers,keyless or key tuning structure etc.

The test instrumentation will be calibrated for a flat frequency spectral output and 0dBU before being attached to the instrument to be tested.

The test channel will have an input Z(impedance) of at least 1 megohm or more (I hope for 10 megohms). The output Z will be sufficiently low to avoid being loaded down by the signal recording/processing equipment.

The signal recorded will be the spectrally flat 0dBU referenced signal so that differences in output amplitudes are quantified between instruments/pickups.

Unless otherwise stated, the tests will be performed upon the open string(s).

The first test will be a single top E string excited at fret 12 with a plastic thumbpick. The resulting signal will be shown as a graph of output dB vs. frequency content, captured in the peak hold mode. This way the "harmonic content" of the string may be seen for each harmonic generated.

The second test will be like the first test, except using the middle E string.

The third test will be like the first two, except using a low E string where available.

These three basic tests may be repeated with the 500K ohm and 250K ohm resistive loads.

These three basic tests may be repeated with the string(s) excited at fret 24 and 36 to show the effect of picking location upon harmonic content.

Perhaps finger chimes/harmonics/flageolets(sp?) will be captured to show how the finger touch, and the picking location can effect the harmonic content. This last, will probably not be done on all instruments.

And then we will proceed to the strums, and total neck response; as a function of load, and as a function of time = 0,2,4,8 seconds.

What is missing that you would like to see?

More later.

Scott Swartz · From: St. Louis, MO

Ed,

How will you correct for the wildly varying frequency response of different pickups on the different steels, ie this one's resonance peak is at 3 kHz with a Q of .8 and the next has a resonant peak at 4.5 kHz with a Q of 1.9?

The freq response character of the pickup would not vary the harmonic decay data as long as its done in a differential fashion ( ie the 2nd harmonic was -10 dB at time zero, -16 dB at time = 2 seconds, -25 dB at time = 4 seconds and you look at the change between 0,2,4 seconds). However the pickup response peak could change the apparent spectral balance at time zero. Example - if the pickup has a very sharp peak at 4 kHz and you use a 1 kHz pitch, the 4th harmonic (4 kHz) will appear very strong due to the pickup, independent of the body or or changer or whatever else.

richard burton · From: Britain

It's pointless doing this test on open strings.
Results would be much nearer reality if the strings were fretted (at various points on the neck) and readings taken.
Obviously, the strings would have to be muted behind the bar, just as in real life.

Steels sound a whole lot different when the strings are barred !

Bob Hoffnar · From: Austin, Tx

Ed,
Unless I have missed something in your posts you might try testing the steels without plugging them in also.

One test I found was to strum the strings of an unplugged steel and listen to the decay. If the fading out ringing sound gets higher as it fades out that means the steel may not hold its own in a recorded track or pull you out of the mud on stage. Maybe that fits into your over time test.

Bob

ed packard · From: Show Low AZ

Scott, Richard, & Bob H…thanks for the inputs.
The issue being addressed in this round of tests is the instrument and its included pickup. I know of no way to “not plug in” and do the tests. The best I can come up with is to use the very HiZ in voltage follower circuit with a LoZ out going directly to the FSA instrumentation…no vol pedal, and no amp attached. These would modify the nature of the pickup response by damping the pickups Q and shifting its resonant point. Different setups (vped/amps/cables) would affect the same pickups differently. We are starting with the most simple of test arrangements. NO attempt will be made to correct/adjust the pickups response. The pickup type, resistance and any other defining information will be noted. It would be nice to be able to use the same pickup on all instruments = not practical as the only easily interchangeable units are on SIERRA, and now MSA. Then there would be the issue of which pickup(s) to use in the tests; let’s start with what is available, and if the technique is seen as a desirable approach, it can be extended to cover other variables.

Further, if an amp/speaker is active during the tests, the output from the speaker serves to excite the resonances in the body/mechanism of the steel, which of course would affect the test results (feedback). The amp/speaker approach can be used to define these resonances which play a strong part in the sustain character. This technique was demonstrated in a rather extensive emailing to some of the Forum folk that requested the previous FSA experiment mailing. This emailing showed the approach used in designing and analyzing the BEAST.

Richard & Bob H…open strings or Not open strings…as I seem to recall, B.E. has been rumored to strum the unplugged PSG and determine something from the sound…I did not hear him say so, so I will class it as a rumor for now. He is also rumored to have spent all day just picking one string in a variety of ways and locations to see what sound variations could be coaxed out of it. I do know that he plays without being “hooked up” for various reasons because we had a phone conversation re the whys and wherefores of the Zirconia bars differences from the metal bar above the 12th fret.

In the test outline post above, the comment re “strums and total neck response” (at the end) was meant to convey that the sum of output (peak held) from exciting the strings on and between all frets is part of the test sequence. The bar to be used will probably be a Zirconia one. A comparison with a BJS , or similar may be a side note. The problem here is that we are introducing the “HUMAN FACTOR” = picking, behind the bar blocking, bar pressure, bar size/material etc. when the basic question is what does the instrument/pickup combination provide. After finding that, the methodology can be extended to suit the need.

No mic’ing of strummed strings is planned at this time…will settle for the pickup(s) signals for now.

David Doggett · Posted 15 Oct 2005 10:28 am

Ed, some of the questions above might be addressed in side tests. For example, while you can't test every possible pickup in every possible guitar, you could take one Sierra or MSA and swap in a few different pickups, just to give us an idea of how much difference the pickup makes, and the different patterns of the different pickups. Likewise, if you happen to luck into two or more guitars with the same pickup, that would be an interesting comparison. Also, while you might test every guitar on the open strings, it would be interesting to test at least one guitar on both open and fretted strings, just to show us how much difference there is.

Earnest Bovine · From: Los Angeles CA USA

I'm more interested in the sound of barred strings than of open strings.

Bob Kagy · From: Lafayette, CO USA

imo open string response is not nearly as telling as fretted string response, but I'd like to hear opinions on this - I could be missing something big here.

Also, where the string is picked relative to the bar position, or open string position seems like a huge factor.

My 2 cents worth.

BK

ed packard · From: Show Low AZ

DD…right on…the operative term here is comparison…I suspect lots (endless actually) “tests”. That is the way it worked out when using the FSA process on the BEAST design. The open string tests tend to define the instrument with pickup sound…barred adds the difference between players and bars into the mix. Here we get into the repeatability issue. I tend to draw the line when the test cannot be repeated within about 2 dB. (2 dB is a voltage ratio of about 1.26).

EB…expect some Hughey Land comparisons as well as string noise on the wounds.

BK…again, fretted string response is largely player dependent (to the extent that the “tone is in the hands”)…as your comment re where to pick indicates. The open string response tends to define the structural signature of the instrument more. That said, with my hands, and my bar, the response will be measured all the way up to the 36th fret.

Re different pickups in the same instrument...that will be done on the SIERRAs; Jim had several MSAs but I don't know how many different pickups for them.
My SIERRA pickups are all "tapped" so the differences re settings can be seen.

[This message was edited by ed packard on 17 October 2005 at 10:29 AM.]

Forum Index > Electronics	Goto page 1, 2 Next	Next oldest topic :: Next newest topic