11,289 MHz one-bit recording?

Posted: Thu Oct 16, 2008 7:59 am

Hi,

I've heard that there should a device out there that can record 11,289 MHz one bit signals. Anybody knows of this?

Best regards,

Jesper

Posted: Thu Oct 16, 2008 9:35 am

11+ megahertz? No. 2.83MHz YES!

So you want your frequency response to include light? Wow! What the heck for? The rest of the equipment can't.

But if you want DSD? It's available at your local music store for less than $750 and manufactured by KORG. Just don't think about adding any plug-ins or other processing as it's not really available in a convenient manner. It does down convert well to the bit depth & sample rate of your choice. So you can make really good sounding recordings that will still sound like CDs in the end. And from those recordings you can also make SACD's. If anybody could play them back which most can't. But if you wanted to buy everybody a $1000 player, each, my address is P.O. Box 4181, Falls Church, Virginia, 22044-0181. Thanks!

I can hardly wait!
Ms. Remy Ann David

Posted: Thu Oct 16, 2008 11:53 pm

11.2896MHz is the rate you get from a 128x DSD output when both L+R channels are taken together at a PCM equivalent rate of 44.1KHz. It is just one of the standards associated with DSD. The SACD standard uses half this rate.

You can buy A-D converter devices that will produce signals at this frequency, e.g. the TI PCM4222, but I don't know of any commercial product that records DSD at that rate.

Posted: Fri Oct 17, 2008 12:13 am

gentlevoice wrote:

I've heard that there should a device out there that can record 11,289 MHz one bit signals. Anybody knows of this?

It's called a logic analyzer. They go into the pico second range (giga hertz)

Tektronix makes a nice one for lab use. not quite 11GHz, but 8GHz. Not so great for product verification even though they say it is.

How would this be applicable to audio?

Posted: Fri Oct 17, 2008 1:57 am

The Korg MR-1000 can do 5.6448 MHz @ 1-bit.

Posted: Fri Oct 17, 2008 2:30 am

Ohh...
http://en.wikipedia.org/wiki/Delta-sigma_modulation
http://en.wikipedia.org/wiki/Direct_Stream_Digital

I never heard of that before. In theory it makes sense. With an extremely high sample rate you essentially eliminate interpolation error when down sampling. But don't you loose bit depth? Is the idea that like the move from IDE (parallel data) to SATA (serial data) You can have a higher data rate because with less wires synchronization error goes down?

Posted: Sun Oct 19, 2008 10:11 pm

GeckoMusic wrote:

Ohh...
http://en.wikipedia.org/wiki/Delta-sigma_modulation
http://en.wikipedia.org/wiki/Direct_Stream_Digital

I never heard of that before. In theory it makes sense. With an extremely high sample rate you essentially eliminate interpolation error when down sampling. But don't you loose bit depth? Is the idea that like the move from IDE (parallel data) to SATA (serial data) You can have a higher data rate because with less wires synchronization error goes down?

That's not really the right way to think about it. Conventionally-sampled data consists of independent parallel words at whatever sampling rate is used (e.g. 24 bit words at 48KHz). Given a perfect clock (no jitter), the accuracy is in the precision of the 24-bit samples, both in the original ADC sampling process and in the DAC reconstruction. What DSD and other reduced-bit representations do is to shift the precision from wordlength into time, with the ultimate being single-bit decisions made at (say) 256x the original sampling rate. The mean value of these is a representation of the original waveform.

One advantage of the technique is that it avoids the need to make major-bit decisions for waveforms that are close to zero, since in a DSD stream, no output value is any different from any other value. This improves linearity and is particularly noticeable in quiet passages, so perhaps a greater improvement is gained for recordings of classical music rather than heavy rock.

In terms of noise, by using various techniques in the single-bit decision process, it's possible to push some of the quantisation noise into a part of the sampled spectrum that will get filtered out by the waveform reconstruction process. This technique (noise shaping) is not unique to DSD, but does give a lower noise floor in the resulting data.

Posted: Mon Oct 20, 2008 8:02 am

Thank you for your replies Very Happy

I am aware that the Korg MR-1000 does DSD at 5,6 MHz, however, I'm hoping to go further up in sampling frequency. And I have heard from what I consider a reliable source that an 11,289 MHz sampling "device" should be available...?

Nobody heard of it?

Best,

Jesper

This topic sponsored by: