Yes, some do. It's enlightening to hear the mike noise during fade in and fade out.
"one small step for man/kind"
That phrase has been replayed many times, yet we are still not sure whether he said man or mankind.
What was that s/n ratio during that phrase?
Certainly in my head the noise was very distracting and interfered significantly with the enjoyment of the sound. But the "moment" was so significant/important that I could live with the annoyance of the noise/interference. The majority of the message got through.
Listening to that very low s/n excerpt (in quite special circumstances) is quite different from listening to low level music and ambiance in between or mixed up with other louder music and the near silence between the music samples.
Here we would not enjoy listening to ambience in the presence of avoidable random noise/interference. We expect the very low level signals that help define the ambiance/venue to be above the "noise".
Intruding noise detracts from the enjoyment of the music.
It does not matter that we have the ability to pick out some/most of a voice communicating with similar levels of noise intruding on the wanted communication.
That phrase has been replayed many times, yet we are still not sure whether he said man or mankind.
What was that s/n ratio during that phrase?
Certainly in my head the noise was very distracting and interfered significantly with the enjoyment of the sound. But the "moment" was so significant/important that I could live with the annoyance of the noise/interference. The majority of the message got through.
Listening to that very low s/n excerpt (in quite special circumstances) is quite different from listening to low level music and ambiance in between or mixed up with other louder music and the near silence between the music samples.
Here we would not enjoy listening to ambience in the presence of avoidable random noise/interference. We expect the very low level signals that help define the ambiance/venue to be above the "noise".
Intruding noise detracts from the enjoyment of the music.
It does not matter that we have the ability to pick out some/most of a voice communicating with similar levels of noise intruding on the wanted communication.
Last edited:
Of course you can encode it. Go look at the spectra. It's only supernatural if you don't understand how the system works.
Explain the experimental result of the spectra. This belief of yours is distressingly common among audiophiles and nowhere else. There's no magic, just basic DSP that any freshman engineering student is familiar with.
I told you where to find numerous experimental verifications. You just don't seem to want to look. I can also suggest experiments for you to try yourself, but I was trying to save you a few steps.
If you want textbook and literature references, start with Ken Pohlmann's excellent Principles of Digital Audio (p36 and p44 are the first places these issues are discussed). Lipshitz and Vanderkooy, JAES Volume 35 Issue 12 pp. 966-975; December 1987. Lipshitz, Vanderkooy, and Wanamaker, JAES Volume 39 Issue 11 pp. 836-852; November 1991. I can give you a few dozen more, but as I said, this is covered in any basic EE text on DSP.
If you want textbook and literature references, start with Ken Pohlmann's excellent Principles of Digital Audio (p36 and p44 are the first places these issues are discussed). Lipshitz and Vanderkooy, JAES Volume 35 Issue 12 pp. 966-975; December 1987. Lipshitz, Vanderkooy, and Wanamaker, JAES Volume 39 Issue 11 pp. 836-852; November 1991. I can give you a few dozen more, but as I said, this is covered in any basic EE text on DSP.
Well we are, because Armstrong was asked what he said and confirmed that he fluffed the line. The planned line was 'one small step for A man'.
So the fact that 3G and 4G phones pull signal from spread spectrum noise magic? Or that satellite comms works. The modern world must be a confusing place for you...
What you really said was, Take a look at low level spectra shown in Stereophile's CD and DAC reviews. You can resolve signals well below the LSB., which is what esgigt was railing against.
You are right and he is right!
You are right that a harmonically complex signal, encoded in ± 2 bits (–1, –0, +0, +1 say for convenience) can very certainly have a spectrum of harmonics that can be quite complex, and most of the energy well below ±1 LSB. Esgigt doesn't want to think about the harmonic structure of the signal. His contention is that one simply can't encode a waveform below –1, 0, +1 values. Or minimalistically 0, 1. He too is correct - about encoding the least information per sample. But time-varying information is spectral in information content too.
And that's the disconnect.
GoatGuy
Thanks for your elaboration.... but what do you mean by "I don't want to think about the harmonic structure of the signal"? Please explain...
No. What all of the literature, all of the experiments, and all of the applications in mission-critical areas outside of audio (I first learned about this stuff when I was doing high resolution infrared spectroscopy back in the 1970s) show is that this is true for ANY signal, "complex" or single frequency. Noise decorrelates the quantization, whether it is naturally present noise or deliberately added noise. Period.
Great way to spread knowledge.... NOT
You expect others to share the same knowledge and experience as you have and fail to see there are other walks of life possible. So, start explaining instead of yelling at people as if they're idiots, who do not understand what you understand... I tried to explain my vision on the subject and you don't even take the time to explain yours .....
we don't listen to sinewaves within narrow bands.
in my ( non audio)_ experience with DACs/ ADCs noise is rarely a determining factor in linear performance esp. complex waveforms in multi octave bandwidths. Look into 'spur free dynamic range' & 'spectral regrowth' for ideas.
in my ( non audio)_ experience with DACs/ ADCs noise is rarely a determining factor in linear performance esp. complex waveforms in multi octave bandwidths. Look into 'spur free dynamic range' & 'spectral regrowth' for ideas.
Man, this thread is too funny. My troll-o-meter(R) (patent pending) almost jumped off the desk...
If you don't buy that "noise decorrelates the quantization", you can try it out yourself and listen to the results.
A digital transmission channel that is sufficiently dithered in one way or another effectively turns into a band-limited analog one, limited in dynamics by maximum amplitude on top and noise at the bottom. It is considered transparent. You could add, say, a sine of arbitrarily low amplitude, and you'll always be able to dig it back out again with a sufficiently narrow band filter. You will not be able to modulate it arbitrarily fast if you want to transmit actual information with it, of course, Mr. Shannon would have something against that. It's like trying to communicate with space probes.
As I wrote on the above page, "the basics discussed here stem from like three or four different lectures", and no-one's going to cram all of that stuff into a forum post. Other people have spent considerable time and effort (more or less) trying to understand it, there's no shortcut that I'm aware of which would allow doing the same within in hour or so.
All of this assumes perfect ADCs and DACs, of course, but modern audio converters do a pretty good job approximating that. Assuming the implementation isn't plagued by correlated jitter or other such fun stuff, but bad implementations do not imply that there's anything wrong with theory.
If you don't buy that "noise decorrelates the quantization", you can try it out yourself and listen to the results.
A digital transmission channel that is sufficiently dithered in one way or another effectively turns into a band-limited analog one, limited in dynamics by maximum amplitude on top and noise at the bottom. It is considered transparent. You could add, say, a sine of arbitrarily low amplitude, and you'll always be able to dig it back out again with a sufficiently narrow band filter. You will not be able to modulate it arbitrarily fast if you want to transmit actual information with it, of course, Mr. Shannon would have something against that. It's like trying to communicate with space probes.
As I wrote on the above page, "the basics discussed here stem from like three or four different lectures", and no-one's going to cram all of that stuff into a forum post. Other people have spent considerable time and effort (more or less) trying to understand it, there's no shortcut that I'm aware of which would allow doing the same within in hour or so.
All of this assumes perfect ADCs and DACs, of course, but modern audio converters do a pretty good job approximating that. Assuming the implementation isn't plagued by correlated jitter or other such fun stuff, but bad implementations do not imply that there's anything wrong with theory.
Last edited:
I gave you several references and pointed you toward routine experimental data demonstrating your error. Have you looked at any of that yet?
Well I don't know which ha the bigest math but from my mortal point of view :
So if the whole system have a low S/N, we are able to hear in each case the highest volume of the peaks (at worst at 0 dB = no attenuation after the dac) but the close lowest level of the disc will be distorsed if the dynamic gap is higher than the S/N of the dac chip; e.g. 95 dB in the case of the TDA1541A ?
Do I translate well the bit depth in relation to the noise/floor of the chip (assuming all the rest of the system has a better S/N and speaker can output with less noise... just for the basic understanding of the story) ?
So if the whole system have a low S/N, we are able to hear in each case the highest volume of the peaks (at worst at 0 dB = no attenuation after the dac) but the close lowest level of the disc will be distorsed if the dynamic gap is higher than the S/N of the dac chip; e.g. 95 dB in the case of the TDA1541A ?
Do I translate well the bit depth in relation to the noise/floor of the chip (assuming all the rest of the system has a better S/N and speaker can output with less noise... just for the basic understanding of the story) ?
Am I wrong again here ?
Could I say the attenuation is not playing an important role in relation to the S/N of the dac chip if the highest dB peaks are played (so without noise) because they mask the lowest music (so the distorsed one because in the noise floor) in the short moment they are played ? (the added noise will not mask the highest décibels played ????). Or do I mix the concept again ?
ESGIGT answer seems wize and clear to me : better higher bit depth, but would like understand what trade offs a 16 bit dac chip is able to in real life (I can't follow you guys on the theoric point of view because the math !) !
I'm sure many are interrested (but if I'm totaly wrong, of course !🙁) to know how it is transcribed in a real hifi experience : did you notice this noise floor yourselves, especially whom have a high sensivity speakers with very low noise floor amp/pre) ?
Do I simply put my money in two PCM 1704 24/96 chips instead a TDA1541A (I love this one for its sound presentation in a very good layout 🙂 )
Could I say the attenuation is not playing an important role in relation to the S/N of the dac chip if the highest dB peaks are played (so without noise) because they mask the lowest music (so the distorsed one because in the noise floor) in the short moment they are played ? (the added noise will not mask the highest décibels played ????). Or do I mix the concept again ?
ESGIGT answer seems wize and clear to me : better higher bit depth, but would like understand what trade offs a 16 bit dac chip is able to in real life (I can't follow you guys on the theoric point of view because the math !) !
I'm sure many are interrested (but if I'm totaly wrong, of course !🙁) to know how it is transcribed in a real hifi experience : did you notice this noise floor yourselves, especially whom have a high sensivity speakers with very low noise floor amp/pre) ?
Do I simply put my money in two PCM 1704 24/96 chips instead a TDA1541A (I love this one for its sound presentation in a very good layout 🙂 )
- Status
- Not open for further replies.