Well I haven't had much time for audio lately but I was standing in line at the store and a thought came to me. If it has been shown that even order harmonics are more benign to our ears, and odd order harmonics are more offensive, why don't we have a Total Odd Order Harmonic Distortion measurement? This could be in lieu of, or in addition to, the traditional THD measurement. I believe that such a measurement could potentially help us better explain what we are experiencing when trying to convey an amplifiers sound to one another on some level.
Thoughts?
Thoughts?
There are many-many "improved" indexes.
A very rational one is to weight the harmonics by order. 9th is much worse than 3rd. Multiply each harmonic by its order. 0.1% of 9th would count as 0.9%.
Some of these schemes have been around for 70 years. Do not hold your breath.
A very rational one is to weight the harmonics by order. 9th is much worse than 3rd. Multiply each harmonic by its order. 0.1% of 9th would count as 0.9%.
Some of these schemes have been around for 70 years. Do not hold your breath.
It's not even remotely that simple. You can search on some of earl geddes work, he's in the speaker forums as gedlee. There are some good but very old threads that suggest looking at some standards that really spell things out, but these are mostly dated and concerned with broadcasting, so not perfectly relevant.
my suggestion was to see if a good perceptual code like AAC uses bits to encode the distortion - then it should be more likely to be audible
About 70 years ago a BBC engineer (Shorter?) suggested weighting harmonics by (n^2)/4. It never caught on.
THD is used because it is simple and has some meaning, and it is fairly easy to get good THD figures. Attaching too much importance to THD is just as silly as ignoring THD.
THD is used because it is simple and has some meaning, and it is fairly easy to get good THD figures. Attaching too much importance to THD is just as silly as ignoring THD.
We can't even present phase of frequency response and distortion properly (Not that we can't, but we can't.), we're not ready for anything more complex.
Why do we need to present phase of frequency response? In most cases it is fixed by the amplitude frequency response, so contains no new information.
Sound engeneers often have hadaches when they want to capture properly instruments sounds with all their harmonics on a resticted frequency range, it seems to me very tricky to capture all the sound oddities of a loudspeakers with the same technology on the whole spectrum.
Last edited:
- Status
- Not open for further replies.