Welcome, Guest. Please login or register.
Did you miss your activation email?

Login with username, password and session length
Forceaudio


 
Pages: [1]
  Print  
Author Topic: Thoughts on the work of GedLee?  (Read 1124 times)
DevilDriver
Guest
« on: October 23, 2006, 08:11:23 pm »

I'm sure that Dan is very familiar with the work of Dr. Earl Geddes and Dr. Lidia Lee.  Has anyone else read any of their work?  I have yet to purchase their AES papers, but from what I have read, their work has been quite impressive.  For example, in late 2004, they did a double blind test where they found that differences in frequency response were very audible while measurable differences in distortion were not nearly as audible to the subjects.

I think this type of work contradicts some of the growing trends in the audio community.  For example, there is great weight in Dr. Klippel's Loudspeaker R&D System; I think a lot of engineers agree with a concern for maximum linearity (ie. BL, Cms, Le, etc) and it seems as if there is an interest in raw THD and IMD measurements.  Does the work of GedLee change an engineer's opinion on the correct design goals?  Is it maybe more important now to understand the perception of hearing than it is the cause and effect relationship of driver design?

BTW, if this sort of conversation interests you, I highly suggest you do two things:
1. Buy "An Introduction to the Psychology of Hearing" by Brian Moore
2. Read some of the .pdf's published on http://www.gedlee.com

What are your thoughts?

Neil
Logged
Forceaudio
« on: October 23, 2006, 08:11:23 pm »

 Logged
DanWiggins
Guest
« Reply #1 on: October 24, 2006, 06:28:12 am »

I've had the pleasure of many dinners with Drs. Geddes and Lee.  Wonderful people!  

Personally, I find the GedLee metric VERY useful.  It shows that IMD is a better indicator - as well as the ratio of the harmonics - of subjective preference.  THD has little correlation with preference, actually (this is assuming you have a decent frequency response).

Personally, I find IMD and BL compression to be the big detractors from sonics.  THD is usually not as bad an issue; for some reason, 1% IMD is a lot worse than 1% THD!  Likewise with BL compression; a bit leaves the music flat, and a lot makes it completely lifeless.

So I tend to design drivers with a goal of flat frequency response, and low IMD and low BL compression.  Optimizing for those two sometimes leaves a bit higher THD, but subjectively the driver sounds better and more accurate.  So that's what I go for...

Dan Wiggins
Adire Audio®
Logged
95Honda
Guest
« Reply #2 on: October 24, 2006, 07:13:31 am »

Dan-

Do you think IMD becomes even more of a concern with very wide bandwidth drivers that cover multiple octaves (like your FRs for instance), more so even than flat BL?
Logged
Forceaudio
« Reply #2 on: October 24, 2006, 07:13:31 am »

 Logged
DevilDriver
Guest
« Reply #3 on: October 24, 2006, 09:09:09 am »

Quote from: 95Honda;1695
Dan-

Do you think IMD becomes even more of a concern with very wide bandwidth drivers that cover multiple octaves (like your FRs for instance), more so even than flat BL?


IMD would increase, but it would be doppler IMD, which is not overly offensive to the ear (Doppler IMD is entirely even order).  Over a wide bandwidth, flat BL is still extremely important because if it were not there, overlayed frequencies would be playing with different parameters.

Imagine a midrange is capable of 10mm of xmax (aka 70% of xmag) and has a Qts of 0.5.  At 60hz, the midrange reaches 10mm of xmax.  At 6khz, the midrange is only a 1mm xmax.  Note that 70% of xmag will double your Qes (and thus your Qts).  When the midrange is playing the 60hz and 6khz tones at the same time, it will be playing with two different parameter sets.  For the 60hz tone, the driver will have an effective Qts of 1.0, but for the 6khz tone, Qts can still be close to rest (this is without considering thermal power compression, which has a similar effect on Qts/Qes).  This is where distortion can come from wide range or full range drivers, so it is even more important to have ruler flat BL.  That's my understanding of it anyways...

I guess I'm going off on a tangent, but I hope I helped with your question.

Neil
« Last Edit: October 24, 2006, 11:35:14 am by DevilDriver » Logged
Forceaudio
   

 Logged
Pages: [1]
  Print  
 
Jump to: