I recently learned of a company called Cerwinski Labs. I didn't make the immediate connection, but soon learned it was started by Gene Czerwinski, founder of Cerwin Vega. Googling around to see what info I could find about the company, I came across this paper published by Czerwinski and a few others. I started to read it but it's all a bit over my head. I'm not looking for help understanding anything, I just figured I'd share it. Maybe some of the more apt minds on the forum might find it interesting
Thank you! That was helpful. I glanced through and read the conclusions.
1. The authors suggest that THD is less useful in correlating performance with audibility. Multitone testing is better. Even better would be if we used multitone testing and if we could appropriately weight higher order distortion products and if we could separate IMD products from harmonic products.
2. Higher order distortion products are much more audible and can be offensive even in small quantities.
3. Lastly, whether a new weighted multitone signal and distortion measurement results in better prediction of performance in terms of perceived quality remains to be seen because it would require significant psychoacoustic testing.
This is good stuff. So much we don’t know but we’re making progress.
1. The authors suggest that THD is less useful in correlating performance with audibility. Multitone testing is better. Even better would be if we used multitone testing and if we could appropriately weight higher order distortion products and if we could separate IMD products from harmonic products.
2. Higher order distortion products are much more audible and can be offensive even in small quantities.
3. Lastly, whether a new weighted multitone signal and distortion measurement results in better prediction of performance in terms of perceived quality remains to be seen because it would require significant psychoacoustic testing.
This is good stuff. So much we don’t know but we’re making progress.
That paper was published 18 years ago, with final words:
Setting any boundaries relating objective information and nonlinear distortion audibility requires extensive computer simulation and involved psychoacoustical tests. Without such information about the relationship between objective and subjective parameters, the measurement data will only be able to tell us that one loudspeaker has more or less nonlinear distortion. The question of how critical this difference is from the standpoint of distortion audibility will remain unanswered.
And where are those studies about IMD audibility levels and standards? Very easy to run measurements today with REW etc.
https://sound-au.com/articles/intermodulation2.htm
http://www.gedlee.com/Papers/The Perception of Distortion.pdf
https://audioxpress.com/article/testing-loudspeakers-which-measurements-matter-part-2
http://aplaudio.com/conc2/products/tda-im
Setting any boundaries relating objective information and nonlinear distortion audibility requires extensive computer simulation and involved psychoacoustical tests. Without such information about the relationship between objective and subjective parameters, the measurement data will only be able to tell us that one loudspeaker has more or less nonlinear distortion. The question of how critical this difference is from the standpoint of distortion audibility will remain unanswered.
And where are those studies about IMD audibility levels and standards? Very easy to run measurements today with REW etc.
https://sound-au.com/articles/intermodulation2.htm
http://www.gedlee.com/Papers/The Perception of Distortion.pdf
https://audioxpress.com/article/testing-loudspeakers-which-measurements-matter-part-2
http://aplaudio.com/conc2/products/tda-im
Easy to run all kinds of measurements Juhazi, but not easy to correlate with perception. Geddes papers also say the same thing: higher orders bad, lower orders good. We can extend this to amplifiers too, where tube amps with low order products but high thd are doing quite well in a market where there are plenty of amps with unmeasurable levels of THD. I have stopped believing that so many listeners are fools and they don’t know what they are listening too. In my opinion, they do know what they like and they are actively choosing it.
Agree with Juha and RA7.
I was out of DIY for roughly 8 years, and just returned to see waveguides and active speakers trending (finally!).
THD is a poor correlator yes. Earl Geddes did great work with the GedLee metric. Has this been tested in study with a large enough number of participants (powered) to determine statistical significance? Is it gaining traction in industry?
When can I buy speakers or headphones with a good/great/GedLee/other metric?
I think behind the doors, Klippel and Comsol and Purifi are doing things and leading cone transducers to lowest levels of non-linear distortion. I wonder if Fraunhofer IIS has published anything on the issue, since they were one of the leaders in perpetual modelling in the late 20C.
I was out of DIY for roughly 8 years, and just returned to see waveguides and active speakers trending (finally!).
THD is a poor correlator yes. Earl Geddes did great work with the GedLee metric. Has this been tested in study with a large enough number of participants (powered) to determine statistical significance? Is it gaining traction in industry?
When can I buy speakers or headphones with a good/great/GedLee/other metric?
I think behind the doors, Klippel and Comsol and Purifi are doing things and leading cone transducers to lowest levels of non-linear distortion. I wonder if Fraunhofer IIS has published anything on the issue, since they were one of the leaders in perpetual modelling in the late 20C.