Post by Hitters November 18, 2014 (51 of 93)
|
|
Thank you all for the expert information you are sharing with neophytes like me. Very interesting data.
|
|
|
Post by Chris November 18, 2014 (52 of 93)
|
|
Iain said:
Will these even play with currently available technology, such as the Sony BDP-S7200?
That machine will play 64 fs, but not sure of >128 fs.
I wouldn't know regarding the Sony you mention.
But one thing is sure,there are already many DACs out there that can cope with both rates and also pcm up to DXD.
2L already has titles at DXD and the Dacapo Nielsen two first titles can now also be downloaded in their native DXD rate from at least one site.
Strangely not from Dacapo's own site though.
They sell 24/192 and 24/96.
Other sites have the Nielsen as 24/88.2. Ripped from SACD?
Regarding 128 and 256 I am pretty sure there will be more titles from several labels sooner rather than later.
Eudora Records which released the first guitar album at both 128 and 256,have a new album coming with piano music by Mozart and Chopin. The first 9 minutes of the Mozart sonata can be downloaded for free at 128 and to me sounds very good.
|
|
|
Post by Astral November 18, 2014 (53 of 93)
|
|
Chris said:
I wouldn't know regarding the Sony you mention.
But one thing is sure,there are already many DACs out there that can cope with both rates and also pcm up to DXD.
2L already has titles at DXD and the Dacapo Nielsen two first titles can now also be downloaded in their native DXD rate from at least one site.
Strangely not from Dacapo's own site though.
They sell 24/192 and 24/96.
Other sites have the Nielsen as 24/88.2. Ripped from SACD?
Regarding 128 and 256 I am pretty sure there will be more titles from several labels sooner rather than later.
Eudora Records which released the first guitar album at both 128 and 256,have a new album coming with piano music by Mozart and Chopin. The first 9 minutes of the Mozart sonata can be downloaded for free at 128 and to me sounds very good.
Both Nielsen/Gilbert titles are available direct from DaCapo in exceptional 24/192 (not the native DXD which was the recording format, not the 'master' format) and the 24/88 files available elsewhere are from the DSD conversion supplied by DaCapo.
|
|
|
|
|
|
AmonRa said:
If things were that simple… straight 1 = raising voltage and 0 = diminishing voltage would result in equal sample strings for a fast transient and slow low frequency wavefront. For this reason there is a comparator circuit which checks if the voltage has actually risen enough compared to the previous sample to warrant another 1. If not, one step back is taken.
Thus maximum slope would look like 1111101 (or 0010000) etc, where there can be maximum of 24 ones or zeros in any string of 28 bits (+3.10 dBDSD). A slowly rising wave would be represented by something like 110110110110 etc, even though there are no dips at all in the analog signal.
Even this is not the final truth, as in the DSD stream the densest parts (most ones) are not the slopes, but amplitude maximums, and most zeros are at amplitude minimums, not downslopes. How this is achieved needs to be explained by somebody else.
Ok. The truth, as you suggest, may indeed be more complex than I thought and was led to believe by oversimplified articles on the subject. I have done some more digging and I discovered this:
http://www.canadapromedia.com/Articles/sacd/sacd.htm
There in the 4th paragraph under Direct Stream Digital (DSD) Encoding is a more detailed explanation of exactly how the bit values are determined in DSD. This explanation addresses your point that the 1 values are clustered at the signal peak, not on the rise. The 0 values are the opposite counterpoint to that. The key question appears to be exactly what is compared to what in order to generate a 1 or 0 value for a particular sample.
So, my point that the value of a sample is logically determinate is true. Therefore, the density of samples is meaningful and critical, as tailspn has described. But, it is not as simple as I believed. My apologies for compounding any error by repeating oversimplified arguments. And, thanks, AmonRa, for identifying my error in understanding the complexities.
|
|
|
Post by tailspn November 19, 2014 (55 of 93)
|
|
As I've posted before, this (my) mind numbing exercise of tracing the table of clock events, provided an excellent insight into the digitizing process of a simple PDM first order modulator:
http://www.embedded.com/design/debug-and-optimization/4406844/The-basics-of-sigma-delta-analog-to-digital-converters-
As I'm want to do, I point out the Digital Filter post the dashed line modulator, which is the PCM conversion device within the vast majority of available A/D converters today, is the element where the first sound quality deterioration of the modulator's (DSD) output occurs.
|
|
|
|
|
|
Tom is right. I've seen a higher influx of DSD projects coming in for mastering over the past few years from local studios. I even got a DSD128fs project from the Tri-cities here in WA! All the Alesis Masterlinks are breaking down and studios are buying Korg MR2000s and Tascam DA-3000 machines.
I'm now getting label requests for quad-DSD tape transfers. Everyone wants bigger and better. I haven't used the Grimm this year! Everything is DSD128/256. The way we do DSD mastering is take it out of the box, go through an analog chain and back into the A/D. We do not do any digital conversions.
|
|
|
Post by M Bishop November 21, 2014 (57 of 93)
|
|
I'm not surprised at the increase in DSD mastering projects. The proliferation of DSD-capable DACs at reasonable pricing has brought about a renewed interest in DSD recordings for sure.
Having recorded, mixed, and mastered almost exclusively in DSD since 1998, DSD is all we see on an everyday basis. Obviously it's our preferred recording format even though DSD tools have been a bit difficult to come by until recently. The trend has definitely been toward 128x DSD with only a glimmer at 256x DSD, probably because of the lack of 128x-capable DACs. We made the world's first 256x recording in partnership with Merging Technologies, but that format is seeing a slow start still. 2.8MHz DSD may be losing ground because the flavor-of-the-day samplerate light doesn't light up, but many great-sounding recordings made in 64x DSD still sound great.
|
|
|
|
|
|
Well Quad-DSD doesn't seem to be getting a slow start here. Someone seemed to open up the flood gates. I get at least one or two calls a week asking about it. Seems to be the new buzz-word now.
Chad even seemed concerned that his recordings were not being done in the highest DSD sample rate, because that's what his customers wanted. I'm pretty format agnostic as long as care was taken during the recording/mixing. I've never gotten into the recording/mixing aspect anyway. I'll let the experts do that. We only specialize in mastering.
|
|
|
Post by RWetmore November 22, 2014 (59 of 93)
|
|
DSD_Mastering said:
Tom is right. I've seen a higher influx of DSD projects coming in for mastering over the past few years from local studios. I even got a DSD128fs project from the Tri-cities here in WA! All the Alesis Masterlinks are breaking down and studios are buying Korg MR2000s and Tascam DA-3000 machines.
I'm now getting label requests for quad-DSD tape transfers. Everyone wants bigger and better. I haven't used the Grimm this year! Everything is DSD128/256. The way we do DSD mastering is take it out of the box, go through an analog chain and back into the A/D. We do not do any digital conversions.
It's a shame Grimm won't be making a DSD 128x A/D converter, especially since Eelco say they could make one with equally low distortion as the AD1.
|
|
|
|
|
|
RWetmore said:
It's a shame Grimm won't be making a DSD 128x A/D converter, especially since Eelco say they could make one with equally low distortion as the AD1.
I'd be the first in line if they did.....
|
|
