- Joined
- May 21, 2019
- Messages
- 4,036
- Likes
- 6,827
In a recent thread I asked whether or not the Klippel NFS can create polar plots of distortion measurements. @amirm clarified that it's not set up to do so and that in any case distortion would show no spatial dependency unless the source was acoustic, i.e., the distortion was caused by the sorts of air perturbations found in horn-loaded speakers.
My thought was that investigating differences in radiation between the fundamental and distortion would help establish the latter's audibility. Perhaps distortion would overpower or simply come from a different direction.
I did a little research and found this open-source AES paper with measurements and details about the method (which incidentally involves certain Klippel modules): https://secure.aes.org/forum/pubs/conventions/?elib=16815.
Some highlights:
The main finding was established early on:
So my initial idea was wrong. Regardless, the study is well done and digging into the details shows some difference in radiation, though nothing of the sort that I expected when I first came to the idea.
Citation #1 above is behind a paywall, unfortunately: https://secure.aes.org/forum/pubs/conventions/?elib=7448
Below is a measurement example of a soft cone direct-radiating loudspeaker. The graphs are the FR of the fundamental, H2, H3 and MTND ("Multi-Tone Nonlinear Distortion"). The signal for the latter is not specified but is said to be similar to a music. Citations #2–#5 may have the details since they are referenced immediately after the term is introduced.
Note that the black & white shading reflects different scaling in each plot. Bright white does not mean they are at the same level.
One other interesting thing is a proof of why mechanical resonances are so important to eliminate:
In this metal cone-based loudspeaker the resonances (particularly H3 at 12kHz) are very strong across a much larger angle than I would have thought.
My thought was that investigating differences in radiation between the fundamental and distortion would help establish the latter's audibility. Perhaps distortion would overpower or simply come from a different direction.
I did a little research and found this open-source AES paper with measurements and details about the method (which incidentally involves certain Klippel modules): https://secure.aes.org/forum/pubs/conventions/?elib=16815.
Some highlights:
The main finding was established early on:
So my initial idea was wrong. Regardless, the study is well done and digging into the details shows some difference in radiation, though nothing of the sort that I expected when I first came to the idea.
Citation #1 above is behind a paywall, unfortunately: https://secure.aes.org/forum/pubs/conventions/?elib=7448
Below is a measurement example of a soft cone direct-radiating loudspeaker. The graphs are the FR of the fundamental, H2, H3 and MTND ("Multi-Tone Nonlinear Distortion"). The signal for the latter is not specified but is said to be similar to a music. Citations #2–#5 may have the details since they are referenced immediately after the term is introduced.
Note that the black & white shading reflects different scaling in each plot. Bright white does not mean they are at the same level.
One other interesting thing is a proof of why mechanical resonances are so important to eliminate:
In this metal cone-based loudspeaker the resonances (particularly H3 at 12kHz) are very strong across a much larger angle than I would have thought.
Last edited:
