Sorry, tried to add after 30 minutes..
Very nice. I've no idea how to interpret the different filters w/r to the time symmetry, others here are far more advanced in that aspect.
When two sines are mixed, the resultant wave will phase shift and amplitude modulate at the beat frequency. Our examples of say, equal amplitude 16 and 20 sines mixed, when you look at the inner carrier of the lobes, the frequency is (16 + 20)/2, or 18, modulated by (20-16)/2. I showed this on the old thread by overlaying the average freq on a sum waveform.
This phase walking depends on the relative amplitudes of both sines, reduce one and the carrier freq shifts within the lobes towards the dominant frequency, and the amplitude envelope stops being 100%. In the limit, zero envelope modulation and at the frequency of the remaining signal.
I believe an FFT does not see that center frequency because it flips phase every lobe, others here can elaborate on that.
So when we mix two sines, an FFT cannot see the mix, but instead will report only the two frequencies that were summed. IOW, the FFT cannot determine if the waveform was generated by addition or multiplication, the trig identity.
Jn
Last edited:
Then it seems to me an audibility test of the idea could be devised by shifting everything down in frequency?
Hi-tek redneck at yer service! Your welcome.......I think?😛
My vote is for moving up to 40k..... is there a downside to doing so?
My big concern was the water is not antiseptic. I worry about sepsis, blood poisoning, whatever the terms are.
Jn
Only when it comes to standing waves.
For loudspeaker alignment, it makes no difference as I stated. Vas is the relevant metric here, as you may learn upon further study, and it is not impacted by the molecular weight of the gas.
Case closed.
Last edited:
Figure out how many gallons and add the appropriate amount (doesn’t take much) of iodine to it? Also, Wouldn’t be a bad idea for when/if they had to emptied......might smell a lot better at the least.
Last edited:
Yes. The sell.
Look at the veritable bloodbath of a technical discussion we had here, among some very heavy hitters intellectually (I speak of others with that statement..me, just a cabbage) ..😉
Imagine trying to sell that technical argument to the layman.
As to tech..JC pointed out microphones that made it to 40k.. Somebody would have to sell studio owners on the need, there I could see it being too expensive for the incremental improvement. The media is already set that high, many here are capable of electronics design flat to 40k.
I suspect all that is needed is to get it to the amp output terminals. I do not think speakers need to go there, we're just trying to not filter out stuff that causes audible content.
Jn
Edit: am I remembering correctly that ESL panels already make it up there?
Last edited:
The main reason for injected oil to cause sepsis is that there is no mechanism for the body to move/eliminate the fluid, that causes the tissue to die in situ, water whilst not good as it damages some tissue by moving it about can at least be eliminated by the body and then the white blood cells do the rest.
so it is much less harmful and best avoided as an open wound is something to avoid at work.
Made an error on post #241. I said the modulation would no longer be 100% when one sine is reduced..that is incorrect. It will no longer be 200% modulation. 100% doesn't flip carrier, cos modulation does.
Jn
Jn
Audio signal below Fs is not altered when sidebands above Fs are removed.
No, lower sidebands below Fs remain when their origin above Fs has been removed.
See spectrum cymbal file below, before and after Brick Wall filtering @20Khz.
Below 20Khz, spectrum is still exactly and 100% the same, and that's what we hear.
We don't listen to time domain envelopes but to frequencies.
If you postulate something, please come with a prove, not just words that are confusing.
Hans
P.S. In case someone wants the .wav files, here they are
Dropbox - Cymbal - Simplify your life
Last edited:
Actually, you already provided proof that it does.
Look up localization.
I am keeping it as simple as possible. If it is confusing to you at this level, I fear I cannot help you.
Jn
Getting back to the 20k * 5k-raised-cosine blip, I made the 20x5 signal at 384k sample rate (blue, and blue in time domain), and a linear phase filter, Linkwitz-Riley @ 22kHz, 96th order (green, and dark red), and the filtered blip (red, and green).
Spectra are as expected. Time domain shows a "stretching" of the zero-crossings at the pulse center, a lower apparant frequency than the "carrier", but as well a ringing in its tails (to either side) that has the "frequency" of the natural ringing of the steep and kinky filter (after adding a proper time offset) and which is close to the original "carrier" but effectively arbitrary by choice of filter. By these two effects, the zero-crossing distances vary and must do so. They should not be taken literally to infer some kind of "instantanuous frequency".
Spectra are as expected. Time domain shows a "stretching" of the zero-crossings at the pulse center, a lower apparant frequency than the "carrier", but as well a ringing in its tails (to either side) that has the "frequency" of the natural ringing of the steep and kinky filter (after adding a proper time offset) and which is close to the original "carrier" but effectively arbitrary by choice of filter. By these two effects, the zero-crossing distances vary and must do so. They should not be taken literally to infer some kind of "instantanuous frequency".
Attachments
Thanks, I was thinking of the top 2 scenarios in your post #109.
20k modulated with and without the upper sideband...
But as continuous tones - so they can be examined more easily...
Just a thought...
Again with this raised cosine thing..
That is not a signal found in music. If you want to prove an entity incorrect, you need to actually test that entity.
Try a sin(20)cos(5)
Or, sin(20)cos(4), lose the upper sideband, then compare to unfiltered sin(18)cos(2)
No more squirrels please.
Jn
Edit: your last sentence...clearly you've not done any phase locked loop work. Nor, have you considered the trig identity I spoke of. Do the correct modulation, it will clear it up a bit.
Last edited:
A continuous tone is not what this is about though, that's why I suggested shifting ALL the frequencies downwards so there is a chance of hearing the effect. I presumed you wanted to listen to the files?
Isn’t the point whether or not the brick wall in the 20k area effects the timing/frequencies below it?
You mean like one period of the direct multiplication (not raised)? I'm pretty sure that won't make much of a systematic difference, slightly different blip, slightly different response, that's it. I might give it a try.
In terms of frequencies below cutoff it doesn't, in the the time domain it does and must do so, by first principles.