Forums

Reverse of SACD?

Started by Curious June 24, 2004
SACD uses 1-bit with a super-high sampling rate.

Could something similar be done using 1 Hz sampling with a super-wide
bit resolution?
Curious wrote:

> SACD uses 1-bit with a super-high sampling rate. > > Could something similar be done using 1 Hz sampling with a super-wide > bit resolution?
You can encode the entire contents of the Library of Congress on a stick less than two inches long. Simply encode it all as 32-bit unicode, concatenate all the codes. and place a binary point after the first bit. That represents a number that is at most 1.111... with a zero certainly somewhere. Cut a stick to that length, and there you have it, to carry around, decode whenever you like, and have a wealth of information in your pocket. What's wrong with that scenario? The short answer is that like all other schemes based on super-wide binary numbers, the theoretical resolution exceeds what can be implemented in practice. Jerry -- Engineering is the art of making what you want from things you can get. �����������������������������������������������������������������������
On Fri, 25 Jun 2004 00:02:27 -0400, Jerry Avins <jya@ieee.org> wrote:

>Curious wrote: > >> SACD uses 1-bit with a super-high sampling rate. >> >> Could something similar be done using 1 Hz sampling with a super-wide >> bit resolution?
The 1Hz sample rate causes problems with audio frequencies above 1/2 Hz. Allan.

Curious wrote:

> SACD uses 1-bit with a super-high sampling rate. > > Could something similar be done using 1 Hz sampling with a super-wide > bit resolution?
Yes, if you had a large enough table to contain all 1 second sequences that meet a given SNR and bandwidth requirement you could index into it and select the proper one with that code you give it every second. That's kinda big though, as tables go, and your bit resolution would need to be log(2) of the number of entries in it. Bob -- "Things should be described as simply as possible, but no simpler." A. Einstein
"Curious" <curious11112001@yahoo.com> wrote:
> SACD uses 1-bit with a super-high sampling rate. > > Could something similar be done using 1 Hz sampling with a super-wide > bit resolution?
I think you'll also find that SACD is a little more complex than just converting a signal into a single bit digital signal. I imagine (without much knowledge of it) that it uses a comparator and a sawtooth/triangular wavefor to generate a PWM signal suited to the input waveform. The PWM signal must be at least twice the input waveforms frequency (nyquist theorem) and must actually be much greater to achieve some additional processing gain from increased sample rate (engineering common sense). This "additional processing gain" allows the PWM output to be recognisably similar to the analog input. You could implement a system with a very high bit resolution but very low sampling rate, however all that allows you to do is very precisely represent a very slowly changing signal. Bevan
Bevan Weiss wrote:

> "Curious" <curious11112001@yahoo.com> wrote: > >>SACD uses 1-bit with a super-high sampling rate. >> >>Could something similar be done using 1 Hz sampling with a super-wide >>bit resolution? > > > I think you'll also find that SACD is a little more complex than just > converting a signal into a single bit digital signal. I imagine (without > much knowledge of it) that it uses a comparator and a sawtooth/triangular > wavefor to generate a PWM signal suited to the input waveform. > The PWM signal must be at least twice the input waveforms frequency (nyquist > theorem) and must actually be much greater to achieve some additional > processing gain from increased sample rate (engineering common sense). This > "additional processing gain" allows the PWM output to be recognisably > similar to the analog input. > > You could implement a system with a very high bit resolution but very low > sampling rate, however all that allows you to do is very precisely represent > a very slowly changing signal. > > > Bevan
At one time or another, we lose sight of Curious's obsession with coding, as opposed to representing, signals. Suppose there were a huge data base of digitized music available. Every performance of every piece of music ever recorded and now in digital form. Suppose further that each has a unique number. To play music, one does not need a number per second, but a number per piece. Such a scheme doesn't violate the Nyquist limit, bit it does give new meaning to the phrase, "Play that number again." Backing away from such a scheme, consider how a vocoder generates speech. In fact, Curious should consider how a vocoder generates speech. Then he might understand the true meanings of his questions. A little learning is a dangerous thing; drink deep, or taste not the Pierian spring: there shallow draughts intoxicate the brain, and drinking largely sobers us again. .. Alexander Pope Curious: Sober up! Jerry -- Engineering is the art of making what you want from things you can get. &#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;
"Bevan Weiss" <kaizen__@NOSPAMhotmail.com> wrote in message news:<iUWCc.3931$LT3.154181@news.xtra.co.nz>...
> "Curious" <curious11112001@yahoo.com> wrote: > > SACD uses 1-bit with a super-high sampling rate. > > > > Could something similar be done using 1 Hz sampling with a super-wide > > bit resolution? > > I think you'll also find that SACD is a little more complex than just > converting a signal into a single bit digital signal. I imagine (without > much knowledge of it) that it uses a comparator and a sawtooth/triangular > wavefor to generate a PWM signal suited to the input waveform. > The PWM signal must be at least twice the input waveforms frequency (nyquist > theorem) and must actually be much greater to achieve some additional > processing gain from increased sample rate (engineering common sense). This > "additional processing gain" allows the PWM output to be recognisably > similar to the analog input. > You could implement a system with a very high bit resolution but very low > sampling rate, however all that allows you to do is very precisely represent > a very slowly changing signal. > > > Bevan
Pulse Width Modulation is used for SACD because there is only 1-bit of represenatation of the amplitude. Could the "reverse" use Pulse Height Modulation to acheive similar results?
Curious wrote:
>>"Curious" <curious11112001@yahoo.com> wrote: >> >>>SACD uses 1-bit with a super-high sampling rate. >>> >>>Could something similar be done using 1 Hz sampling with a super-wide >>>bit resolution? > > Pulse Width Modulation is used for SACD because there is only 1-bit of > represenatation of the amplitude. Could the "reverse" use Pulse Height > Modulation to acheive similar results?
SACD doesn't use pulse-width modulation. It uses noise-shaped sigma-delta modulation. Totally different, more subtle animal. To answer your question: no. Pulse Height Modulation is commonly referred to as "sampling", the precursor to PCM quantization. As Allan has pointed out, 1Hz sampling limits you to only being able to reproduce frequencies lower than 0.5Hz. Good for geological surveys, perhaps, but not audio. On the other hand, if you meant a 1Hz encoded symbol rate, independent of the signal sampling rate, then yes, you can do that, but all that you would gain is inconvenience and memory consumption. It's meaningless. Think of an audio CD as a (1/3600)Hz (or so) data-rate representation of a stereo album. (With about five gigabits/baud...) -- Andrew
Curious wrote:

> "Bevan Weiss" <kaizen__@NOSPAMhotmail.com> wrote in message news:<iUWCc.3931$LT3.154181@news.xtra.co.nz>... > >>"Curious" <curious11112001@yahoo.com> wrote: >> >>>SACD uses 1-bit with a super-high sampling rate. >>> >>>Could something similar be done using 1 Hz sampling with a super-wide >>>bit resolution? >> >>I think you'll also find that SACD is a little more complex than just >>converting a signal into a single bit digital signal. I imagine (without >>much knowledge of it) that it uses a comparator and a sawtooth/triangular >>wavefor to generate a PWM signal suited to the input waveform. >>The PWM signal must be at least twice the input waveforms frequency (nyquist >>theorem) and must actually be much greater to achieve some additional >>processing gain from increased sample rate (engineering common sense). This >>"additional processing gain" allows the PWM output to be recognisably >>similar to the analog input. >>You could implement a system with a very high bit resolution but very low >>sampling rate, however all that allows you to do is very precisely represent >>a very slowly changing signal. >> >> >>Bevan > > > Pulse Width Modulation is used for SACD because there is only 1-bit of > represenatation of the amplitude. Could the "reverse" use Pulse Height > Modulation to acheive similar results?
Sure. That's analog, but harder to implement than other analog strategies. Jerry -- Engineering is the art of making what you want from things you can get. &#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;&#2013266095;