Recording for Analysis

  • Thread starter Thread starter classicalsax
  • Start date Start date
C

classicalsax

New member
I am beginning to work on a doctoral dissertation comparing different tones of saxophone players - primarily focused on classical players, but players with other setups will be compared as well. I will be doing spectral analysis and some sinusoidal modeling techniques to compare the core 'tone' of one player to another.

This needs to be a traveling setup so it'll be a laptop with some sort of usb/firewire box with mic preamps and such. I will be recording in "concert halls" mostly.

What microphone would be best for this endeavor? Would it be wise in getting a matched pair or spending a little more and get a better single mic? I would like not to spend more than $500 (I'm a grad student you know) but want the most accurate samples I can get without having to sell a kidney to get a tube mic or something. I would also like for this setup to record recitals - acoustic instruments only.

I am aware of all the variables of mic placement, room size, color of paint on the walls... and all the compromises that have to be made. I just don't know enough about mic’s to begin to look at them. I’ve read through the FAQ and have some ideas, but would like some opinions also. Thanks

Keith
 
classicalsax said:
I am beginning to work on a doctoral dissertation comparing different tones of saxophone players - primarily focused on classical players, but players with other setups will be compared as well. I will be doing spectral analysis and some sinusoidal modeling techniques to compare the core 'tone' of one player to another.

This needs to be a traveling setup so it'll be a laptop with some sort of usb/firewire box with mic preamps and such. I will be recording in "concert halls" mostly.

What microphone would be best for this endeavor? Would it be wise in getting a matched pair or spending a little more and get a better single mic? I would like not to spend more than $500 (I'm a grad student you know) but want the most accurate samples I can get without having to sell a kidney to get a tube mic or something. I would also like for this setup to record recitals - acoustic instruments only.

I am aware of all the variables of mic placement, room size, color of paint on the walls... and all the compromises that have to be made. I just don't know enough about mic’s to begin to look at them. I’ve read through the FAQ and have some ideas, but would like some opinions also. Thanks

Keith


Not to rain on your parade, but if you are going for a phD, the highest degree awarded by any college, your dissertation has to be based in fact. So, $500 in equipment will not get you anywhere near the quality you seek. You cannot reference one tone to another tone without referencing one tone to *itself*. You really need a totally professional studio with well designed rooms to capture the *real* tone difference between two sources. Only when you can hear the real truth can you select an accurate mic. The mic and soundcard interface, once verified in a proper environment, can then be used for recording the various sources and the differences in the "tone" can be quantified against a known reference "tone".
 
Re: Re: Recording for Analysis

acorec said:
Not to rain on your parade, but if you are going for a phD, the highest degree awarded by any college, your dissertation has to be based in fact. So, $500 in equipment will not get you anywhere near the quality you seek. You cannot reference one tone to another tone without referencing one tone to *itself*. You really need a totally professional studio with well designed rooms to capture the *real* tone difference between two sources. Only when you can hear the real truth can you select an accurate mic. The mic and soundcard interface, once verified in a proper environment, can then be used for recording the various sources and the differences in the "tone" can be quantified against a known reference "tone".

No rain here... only snow.

As I stated in my original post, I am quite aware of the compromises that must be made for this study. This is not a physics doctoral study - it is a music doctoral study; perception is what is important. This is also an entry point into a lifelong field of study that will likely not be completed in my lifetime. Rarely, if ever, is there a published work by any singular author that covers any given subject without a bias, compromise or some uncontrolled variables. There is no real truth in music or sound. Untrained saxophone players often say Kenny Gorlick (Kenny G.) has a great sound. Trained saxophone players would say that it is not very good. What is the “truth” here? What I am interested in is the difference between the sounds and how they are perceived given a list of generalities.

Acoustics and psychoacoustics are different topics that require different research techniques. For example: Sound waves from player A, all things being equal, will react differently in Carnegie Hall than in Boston Symphony Hall, in a church down the street or my living room. But we can identify this as the same player. Although a physicist could tell us all the details of why there are slight variations in the sound from venue to venue based on dimensions, wall coverings, flooring material, etc., this is not applicable to musicians because all musicians will not perform in the same venue - and never in an acoustically perfect room - because one does not exist because of people's perception of the sound is different. A panel of professional, internationally known saxophone players, could compare player A’s sound to their own ideal sound and come up with answers such as: sound is too bright, too dull, to fuzzy, not resonant enough, too dark, centered, not centered, right on, awesome, radical… well probably not the last few, but you can see my point. They could probably agree on more general items as “good tone” or “bad tone" and list reasons why.

A "clean" room will accomplish nothing if from one recording to the next is at a slightly different pitch, different decibel level, difference in the position of the reed, a reed change, a difference in air pressure, tongue position, throat position, embouchure setting... you soon run up a list of variables that cannot practically be controlled. And then try to get all of these variables worked out and add another player, let along a large sample of players - it is too difficult to even list. All variables cannot be controlled. This is the bias of the study. It will be fully documented, with surveys, comparisons and a myriad of other research and study items that has nothing to do with recording the actual sound.

While I agree that the same recording venue or lack thereof will change the results dramatically – it also significantly narrows the scope of the study. This is the biggest compromise I must make. With a big enough sample, trends can be seen – but no “truth” will ever be found. And at the end of it all, there may not be a great conclusion – that’s academia. The process and what is learned is often more important than conclusion. My only hope is that it might aid in teaching and give more of a quantifiable reason for saying a particular player does not sound as good as another. This is not earth-shattering research – there is little of that in music. This is a work that will further research for years to come for myself and many other people who become interested in this niche. And maybe after several years of being a professor I can get a grant to get high dollar equipment to continue the study.

Sorry for the rant – I just got the impression that just since it says “newbie” on my post that it means I am naïve. I assure you, I am not or my advisors would have said something.

Keith
 
With the parameters you laid out above, exactly how "flat" of a response in your signal chain do you need to achieve the goals of your analysis? And what are the minimum parameters for overall frequency response?

There are small-diameter mics available that can do ruler-flat from about 10Hz-15KHz. They are designed for measuring room response.

Likewise there are preamps that can do ruler-flat from 2Hz-200KHz. But those are extreme big-boy toys.

So what kind of leeway do you have in the coloration of the mic and pre combined, and do you have access to sw or other means to "subtract" the coloration of the signal chain from the sample. Not literally, but say graphically or mathematically account for say a 3dB rise in response @ 8KHz, for example?
 
Shoot! I just spent quater an hour typing and all gone....

OK, basically I wanted to say:

Classicalsax,

I agree with folks here--you need a SD omni meausrement mic, as it is very accurate, flat, and doesn't exhibit proximity effect. As Crazydoc has mentioned, the Behringer ECM8000 would be a nice option, and for your purposes you will probably need only one. However, I would suggest you to modify it (I have a thread on that). From the first glance at waves on Protools 6.0, I can see that modified has more 'detailed' ones. If you want I can post wave's pics here.
 
Last edited:
Back
Top