DSD Confusion
Subscribe to Ask Paul Ask a QuestionDSD is the best sounding musical format there is, yet much in the way of confusion surrounds its operation. How does DSD work and how does it compare to PCM
COMMENTS
Leave a Reply
You must be logged in to post a comment.
Sound engineers tell me that mixing and mastering always has to be performed in PCM even if the digital recording was made in PDM/DSD. Thus why should reconverting the PCM mix to DSD improve the sound quality? The only effect I could imagine that the reconstruction filters for PCM are less audiophile than the simple filter design for DSD. But then all cutting edge DACs should convert PCM data to DSD. And what about DSP designs for digital equalizers and digital crossovers? Are there concepts for DSD here?
Paul, Very much appreciate your videos. So in this one on DSD you seemed to explain “loudness” (varying density). What about tone (musical frequencies)? How is a 16Hz organ pedal note conveyed compared to the high frequency sound of a tinkling triangle? This is a mystery to me. I am also confused about DA specs. Using your own Direct Stream DAC as an example. What does “raw DSD mean”? What is DoP (I believe that is DSD over PCM)? Is DoP true DSD? Some DACs seem to only accept DoP? Is that good or does that represent a limitation to the design and especially the ultimate sound? Finally, why does an expensive DAC like the Direct Stream not accept DSD 64, 128, 256 and 512 but only 64 and 128? Thanks.
It’s probably a little hard to visualize. But we understand that as more bits are added, we get louder: greater bit density offers high volume. The speed at which we add those bits is how we get frequency. If the pulse density increases 16 times per second, the result will be a 16Hz organ note (to put it simply).
Yes, DoP is DSD over PCM. It’s simply a way to fool a computer that cannot handle DSD into doing so. To do this a DSD stream is broken into many strings of DSD data (kind of like the way a freight train uses many cars to move a single big load). Each of the cars contains DSD data. To fool the computer, we label each freight car with a set of delivery instructions common to PCM words. When your DAC gets these instructions, it knows to throw away the added identifiers and link all the individual freight cars of DSD data together to get back to where we started.
Our DACs are limited by their hardware. When we launched DirecStream double rate DSD was all there was. It wasn’t until much later people got interested in higher rates. To be honest, there isn’t a lot of sonic advantage to them,.
Doesn’t DSD require the most accurate clock compared to PCM in order to get the jitter effects as low as in PCM?
DSD requires very low jitter, yes.
Does the old MASH 1 bit systems have any advantages over the 16, 18, and 20 bit systems as far as more analog sounding?
MASH was “1 bit” but it relied upon PWM rather than PDM. With Pulse Width Modulation you’re limited in dynamic range to about 16 bits.
The old 14 bit CD players achieved 16 bits through oversampling and some people loved the sound more then the first 16 bit players. Was the thought to design a 1 bit and achieve 16 bits through oversampling believing if 14 bits can sound really good through oversampling then 1 bit through much more over-sampled to achieve 16 bits should sound even better? Or was that design meant as a lower cost way to achieve 16 bit with less time on the bench to fine tune the DA converter then a 16 bit or higher DA converter thus saving labor costs? The real good high end 16 bit DA converters were expensive due to the additional time required on the bench. I have an old Conrad Johnson CD player which was a modified Magnavox/Philips 1 bit and said to be very warm sounding in the audio magazine reviews.
The Philips TDA1541A S2 double crown d/a converter comes to mind still sought after by audiophiles, and vintage machines that have it demand high prices in the used market.
https://www.dutchaudioclassics.nl/philips_tda1541-s2/