That should be interesting. Must sound pretty accurate straight thru if Kavi wants to try it. I will buy a sample of the recording and tell if it is better than what he used to use many years ago. I doubt if some of the features will be used. Just straight thru unmolested is his typical approach.
The mic setup might be interesting, too. Kavi likes an accurate sound field presentation. So, this may be better.
THx-RNMarsh
Last edited:
It has always been.
Because you were using raised sine modulation.
When it was pointed out that your resultant carrier was confounding the time domain waveform, you both ignored that aspect, and actually claimed that it was meaningless and that there was no frequency there..
Recall your time domain plots showed 18Khz modulated by 2Khz even though you started with 20k khz modulated by 4.
You put together simulations which were designed to meet your preconceived notions. On purpose, or by accident, I cannot say.
Now, you post exactly what others have already done.
As to your conclusions, you need to work on those, you clearly did not understand the point.
jn
To put the above partial quote into context:
"This situation might occur with high pitched instruments with a sharp onset (attack), such as cymbals."
"IMO that is a very big "might". How and when are the side bands created?"
I'm not disputing AM creates sidebands
Because you were and still are completely unable to tell your thesis, let alone to come with a prove.
When can we expect a brief and well formulated thesis, this is the third time I ask.
Hans
I am not referring to continuous sine wave modulation, but rather, modulation consistent with musical instruments. My first example was the exponential decay, consistent with a damped sinusoid.
Ah, book.
Continuous and Discrete Signal and System Analysis, McGillem and Cooper, 1974, Holt, Rinehart, and Winston.
Page 150 table 5-4 shows a bunch of Fourier Transforms of Energy Signals, including damped sine and exponential.
jn
I know. Are you saying the envelope of musical instruments result in the creation of sidebands?
No, but having recently come back from holiday to find nothing changed was disheartening. Not sure how much technical contribution I'm actually making and I've gotten heavily into woodworking for the moment. I should switch back to audio stuff soon, if for no reason past me having a few projects needing to be finished to a "living room acceptable" standard*. Some day it'd be nice to have a neater home.
*Those are emphasis quotes and so the reader understands they go together. Nothing scary involved. 😉
Thank you, I just wanted to be sure. I've not been able to find evidence of this, can you show me some please? It's why I asked vacuphile how and when they are created.
Last edited:
no prob.
This is page 150 of the book I cited.
Left column is function, right column is the spectra.
Near top is the exponential.
Lower down is the damped sine.
jn
Attachments
Perhaps one way to envision it is through consideration of why "windowing" is used in an FFT. The various windowings used appear to me to be a form of amplitude modulation, giving the analyzed signal a "soft start" - and a soft end. Without windowing, the FFT will show frequencies that arent there in a continuous, steady state signal being analyzed.
Attacks in real cymbal envelopes are hardly "soft", particularly considering something like Billy Cobham's small China, which he claims he can make be "the loudest thing in the room" (When that guy hits that particular piece of metal - I believe him) Like a plucked guitar string, I imagine the spectral content of such a thing is time variant, depends on where you hit it, and if there's any musical sound that's wide bandwidth noise modulated by sharp envelopes, that's got to be one of them.
A high Q filter takes time to start up - like a few cycles - even when fed a frequency the same as its tuning. It's like it imparts its own "window" on a sharp envelope signal. Generally speaking, what's the Q of a brick wall filter - high or low? If high, perhaps it's "windowing" the start of all transients (with significant content approaching the cutoff frequency) and that's what's being heard. Even if you cant hear that high a single frequency anymore, you can still hear the "windowing" effects modifying the envelope of what you can hear.
Move the sample rate and corresponding filter out well beyond "22.5" and it sounds more real, because transients are now "windowed" differently by the same or perhaps even lesser Q filter tuned to the appropriate frequency for the higher sample rate.
The above may be all bullsh*t, but how to account for people (who cant hear to 20k...) claiming they can hear artifacts imparted by filters operating almost an octave beyond their tonal perception capability?
Attacks in real cymbal envelopes are hardly "soft", particularly considering something like Billy Cobham's small China, which he claims he can make be "the loudest thing in the room" (When that guy hits that particular piece of metal - I believe him) Like a plucked guitar string, I imagine the spectral content of such a thing is time variant, depends on where you hit it, and if there's any musical sound that's wide bandwidth noise modulated by sharp envelopes, that's got to be one of them.
A high Q filter takes time to start up - like a few cycles - even when fed a frequency the same as its tuning. It's like it imparts its own "window" on a sharp envelope signal. Generally speaking, what's the Q of a brick wall filter - high or low? If high, perhaps it's "windowing" the start of all transients (with significant content approaching the cutoff frequency) and that's what's being heard. Even if you cant hear that high a single frequency anymore, you can still hear the "windowing" effects modifying the envelope of what you can hear.
Move the sample rate and corresponding filter out well beyond "22.5" and it sounds more real, because transients are now "windowed" differently by the same or perhaps even lesser Q filter tuned to the appropriate frequency for the higher sample rate.
The above may be all bullsh*t, but how to account for people (who cant hear to 20k...) claiming they can hear artifacts imparted by filters operating almost an octave beyond their tonal perception capability?
Last edited:
Not a characterization but a description of how you had it set up.
6th December 2019: "OK quickly... short version ---> accuracy test. One using no electronic test equip:
Find a person who you know very well. Someone you talk/listen to every day. A male and a female would be nice.
Have that person stand between your speakers and recite something. While they are reciting, record them also. You sit in your usual listening place.
Now play back their recorded voice. Adjust level to be the same as when they talked. Does the reproduced voice sound exactly the same as hearing it "live"?
If it does, you have an accurate system.
You are comparing live voice sound you know well in the same room and space and location acoustics as your music reproduction system. At least for the critical midrange freqs."
Have you ever done live vs. recorded sound accuracy comparison of cello?