Sure, and it is a done deal, probably have been used by recording industry since the 90s. Seems to me everybody here agrees except for JN who seems to think that higher sampling rate solution as a stopgap approach.... If you have non- continuous signal and transients instead, then a higher sampling rate is needed for accurate capture...
... I do not consider raising the rate because people think "something is amiss" is a real engineering approach. Yes, it will certainly raise the margin such that modulation sidebands are no longer an issue, but it is not an elegant approach, just a band aid style fix...
Is it the reconstruction filter that makes the envelope created sidebands jn is referring to?
Yes, the shape that looks like modulated. No sidebands! There is just one frequency, 21kHz.
If you have non- continuous signal and transients instead, then a higher sampling rate is needed for accurate capture.
Richard,
In the context of talking about sampling, the term 'continuous' doesn't exclude musical transients or other musical events at all. Continuous is meant in the mathematical sense, maybe more like 'unbroken over time and without infinitely fast discontinuities.'
Thanks, I couldn't understand where this so called AM was happening 🙂Yes, the shape that looks like modulated. No sidebands! There is just one frequency, 21kHz.
Between those who want our listening experiences to bend theory and those who want the opposite, we will not get away with it.
Personally, I don't really care about the theory. Only the result counts. IF the very slight difference heard on the attacks of a percussion instrument is an intermodulation distortion, and this distortion of such short duration that it is not perceived as such but gives the impression of a more percussive attack , ie closer to reality, it is won.
This is how many sound engineers used the flaws of magnetic tapes to create the sound they wanted. Many recorded on the multitrack the kick drums at a very high level. The reduction in the treble level, combined with the increase in the level of distortion, gave the fairly natural impression of hearing the skin impact a little better. The goal being to separate a little more the kick from the bass.
Who cares if 16/44.1 is better or 24/96. We produce records with what is available, and chose what we prefer on our listening preferences. That explain why, you will find some records better with red book and other with 24/96. And that depends too of you hearing abilities, qualities of your speakers etc.
Some producers like to separate instruments as much as possible, Phil Spector did the contrary, with its attempt to create the well known "wall of sound".
It is all about music, guys, not rocket science. Recorded music is produced by using both the qualities and flaws of the equipments, like the photographers use the lack of depth of field and the bokeh of the lenses to *create* an artistic emotion. Sorry, dear objectivists, there is no measuring instruments for our musical emotions.
Musical instrument makers do not complain about the distortions (harmonics) that produce their instruments, they use-it ;-)
Personally, I don't really care about the theory. Only the result counts. IF the very slight difference heard on the attacks of a percussion instrument is an intermodulation distortion, and this distortion of such short duration that it is not perceived as such but gives the impression of a more percussive attack , ie closer to reality, it is won.
This is how many sound engineers used the flaws of magnetic tapes to create the sound they wanted. Many recorded on the multitrack the kick drums at a very high level. The reduction in the treble level, combined with the increase in the level of distortion, gave the fairly natural impression of hearing the skin impact a little better. The goal being to separate a little more the kick from the bass.
Who cares if 16/44.1 is better or 24/96. We produce records with what is available, and chose what we prefer on our listening preferences. That explain why, you will find some records better with red book and other with 24/96. And that depends too of you hearing abilities, qualities of your speakers etc.
Some producers like to separate instruments as much as possible, Phil Spector did the contrary, with its attempt to create the well known "wall of sound".
It is all about music, guys, not rocket science. Recorded music is produced by using both the qualities and flaws of the equipments, like the photographers use the lack of depth of field and the bokeh of the lenses to *create* an artistic emotion. Sorry, dear objectivists, there is no measuring instruments for our musical emotions.
Musical instrument makers do not complain about the distortions (harmonics) that produce their instruments, they use-it ;-)
Last edited:
No, I said increasing it without understanding why it makes a difference is not elegant engineering.Sure, and it is a done deal, probably have been used by recording industry since the 90s. Seems to me everybody here agrees except for JN who seems to think that higher sampling rate solution as a stopgap approach.
Read Richards post on sampling.
A filter is required in front of the sampler if the analog signal exceeds the nyquist rate of the sampler.
If a brickwall digital filter is used in front of the sampler, it has it's own sampling rate.
jn
Is it the reconstruction filter that makes the envelope created sidebands jn is referring to?
No, it is the amplitude modulation. A time varying amplitude creates sidebands. If the sidebands are wide enough, problems occur with nyquist.
jn
Yes, the shape that looks like modulated. No sidebands! There is just one frequency, 21kHz.
You posted a picture of a two frequency beat pattern that you claim proves there is no amplitude modulation, without understanding the trig identity which defines equivalence between two sines summed and a sine that is modulated by a cosine.
Someone who is as good and smart as you (in my estimation) posting such erroneous statements, is disturbing...you did not let facts get in the way of your bravado.
You are much better than that.
jn
Last edited:
I asked you when the sidebands were created?
The sum of two sines in beat is exactly equivalent to amplitude modulation of the average frequency of the two sines multiplied by the cosine of the difference between the two sines.
The sum of a 17.5khz sine and a 22.5 khz sine is exactly equivalent to the product of a 20 khz sine modulated by the cosine of 2.5 khz.
It is impossible to distinguish between signals created either way.
If I use the amplitude modulation technique, the sidebands 15.5 k and 22.5 k show up because of this equivalence.
This is the trig identity. I suspect that equation and understanding may be over a thousand years old.
Oh, a direct answer to your question could also be: get a good book on AM radio technology, it is a fundamental concept of AM radio.
jn
Last edited:
This is the trig identity. I suspect that equation and understanding may be over a thousand years old.
Perhaps that is the root of the problem 🙂
Hey, I had trouble remembering it, and that was 50 years ago (give or take).Perhaps that is the root of the problem 🙂
jn
What makes you suspect the engineers who worked on the problem in the 70's and 80's did not understand the fundamentals? I'd like to think the problem was dealt with appropriately.... I said increasing it without understanding why it makes a difference is not elegant engineering.
Kindly point to an implementation. I have never seen any, the concept seems kind of absurd, but then I am no good at this either.If a brickwall digital filter is used in front of the sampler, it has it's own sampling rate...
Richard,
In the context of talking about sampling, the term 'continuous' doesn't exclude musical transients or other musical events at all. Continuous is meant in the mathematical sense, maybe more like 'unbroken over time and without infinitely fast discontinuities.'
Thats a nice distinction.... however, a transient IS effectively broken and it has discontinuity like characteristics. .
If we are talking about transient signals like certain sounds and fast rise times. Then there is the chance the sampling rate is not high enough to capture it accurately.
The debate then seems to include What is the Bw and Tr of sounds and how many samples does it take in one pass (not CW) to capture it accurately...? Is 44.1Khz fast enough for that type of signal. Tektronix suggests a 5 times rule. I think even higher would be better. It would relax the filter requirements and move the effects of what JN talks about, also.
So what happens when the sample rate is too low on a transient, non continuous signal? Imagine a single-shot pulse. Well, we get a much reduced amplitude, for one. unless you get very lucky. Details you used to have at longer signal periods are now reduced.
Should we capture non-continuous waveforms to an equivalent of >60-70KHz BW so the leading edge and other details are captured also?
THx-RNMarsh
Last edited:
What is he talking about in your opinion?It would relax the filter requirements and move the effects of what JN talks about, also.
Okay, lets say we have two signal generators, one set to 20kHz, and the other at 2.5kHz. If we sum the outputs (not multiply them!), brickwall, then sample the summed signal, how does the summed signal ever turn into a different waveform that would require multiplying the original two frequencies?
Okay, then let's start with two signal generators set to 17.5kHz and 22.5kHz. If we sum the two signals in the analog domain them brickwall out the 22.5kHz frequency, then digitize the result consisting of only 17.5kHz, how does that turn into an amplitude modulated waveform?
Are you sure that what you are really talking about isn't something more like a single 20kHz analog signal beating with a sample rate of 44.1kHz (or with the Nyquist frequency)?
If so, a question to consider is whether or not the beat (difference) is an actual frequency or just looks like one when eyeballed? (...given we already saw in another case a beat is an interference pattern, not a modulation -- i.e. it doesn't exist as a new frequency in the frequency domain).
Okay, then let's start with two signal generators set to 17.5kHz and 22.5kHz. If we sum the two signals in the analog domain them brickwall out the 22.5kHz frequency, then digitize the result consisting of only 17.5kHz, how does that turn into an amplitude modulated waveform?
Are you sure that what you are really talking about isn't something more like a single 20kHz analog signal beating with a sample rate of 44.1kHz (or with the Nyquist frequency)?
If so, a question to consider is whether or not the beat (difference) is an actual frequency or just looks like one when eyeballed? (...given we already saw in another case a beat is an interference pattern, not a modulation -- i.e. it doesn't exist as a new frequency in the frequency domain).
Last edited:
If we are talking about transient signals like certain sounds and fast rise times. Then there is the chance the sampling rate is not high enough to capture it accurately.
But, they believe there is no transient in music and human hearing threshold is 20kHz or less. So, 44.1kHz is more than enough.
What makes you suspect the engineers who worked on the problem in the 70's and 80's did not understand the fundamentals? I'd like to think the problem was dealt with appropriately.
Kindly point to an implementation. I have never seen any, the concept seems kind of absurd, but then I am no good at this either.
Ah, forgot... The engineers worked the problem given constraints of existing media and equipment. That is not the case at this time.
As to whether they understood envelope based sideband generation, I cannot answer for them. All I see currently is the statement that they system is perfectly capable of passing steady state sines with no distortion up to our limit of hearing, 20K.
your second Query:
Anti-aliasing filter - Wikipedia
Go down to:
Audio applications
Anti-aliasing filters are commonly used at the input of digital signal processing system's analog to digital converter; similar filters are used as reconstruction filters at the output of such systems, for example in music players. In the latter case, the filter prevents imaging, the reverse process of aliasing where in-band frequencies are mirrored out of band.
or here, figure 2.
Digital Recording Techniques
jn
Last edited:
But, they believe there is no transient in music and human hearing threshold is 20kHz or less.
I didn't say that. Some people might hear something over 20kHz, probably not me given my age.
- Status
- Not open for further replies.
- Home
- Member Areas
- The Lounge
- John Curl's Blowtorch preamplifier part III