Hi everyone,
I have read some epic misunderstandings here. I won't brag about the science involved in Digital signal processing and the different methods used to digitalize analogue signals, but :
it's a complete non sense to bluntly compare the sampling frequencies of DSD64, and PCM (whatever encapsulation or form is used), those are two different ways to digitalize music whith each pro and cons. it's like to compare FM and AM for radio broadcast, or CCD and CMOS for sensors, two very different technologies to achieve the same goal.
For those interested, you can learn of to measure the absolute quantity of information that can be stored using a method (entropy).
having 1 bit quantification or 16,24,32 bit quantification is very different, but telling "more is better", it is not true. keep in mind that when there is conversion, there is noise (always) induced by the very nature of quantification or sampling. When you digitalize, you have sampling noise (cuting the analogue signal in very thin segments), but quantification/quantization noise as well (measuring the value of the segment)
Keep in mind that it's just to get the idea of it.
in the late 90'DSD was designed to reduce as much as possible the drawback of the previous methods, that were limiting by design the absolute performance of the media. SACD was design to achieve better dynamic, more bandwidth (the true meaning of bandwidth is the width of the frequency band that can be reproduced). The SACD was thought to be better than the CD by design.
DSD use a different approach than PCM, but that doesnt mean that PCM is bad.
DSD files displayed in time can be seen as a signal of 0(flat) and 1 (diracs, or peaks). simple. but its the density of the peaks that does it all. In fact it's very close to digital amplification structure (class D and derived class). DSD was designed to be directly ingested by digital amp with a simple sum filter.
This approach is very clever, maybe too clever to be marketed to mainstream audience, and damn, was the marketing very poor on SACD!
if you create DSD from PCM, it's very often no use at all (except for very high frequency professional file format). But if you create a DSD master from the original analog tapes, then it's interesting. DSD conversion from analogue signal can be virtually free of quantification noise until 88-90khz, so completely free of conversion noise on the entire frequency range that is considered Hifi (20Khz max)
again, keep in mind that the format was designed nearly 18 years ago to compete with CD, AND embed mutlichannel version on the same disc (quite a feat). When you are lucky to find a SACD version of an album, you have 99.9% of chance to have the best version available on earth of this album. Then there is no point of keeping inferior version like 16/44, or even 24/96 PCM (because it's difficult to be sure where it comes from).
But, and this is important, if a remaster of an old album, would be done in PCM with the latest technologies available in 2015, the very best ADC, and directly from analogue master tapes, the result would be very good and quite similar to a DSD64 master (exception for DSD128 or DSD256 which are even better). But sadly, most high res files you have today comes from the master created years ago for the CD version (hence the high number of 88.8khz files / 96khz), in times where ADC where not as good as today.
I'm not sure, but I thought that it was compulsory for a SACD to be issued to be mastered from the original Analogue tape (that might be wrong though) => more reliable as a source
Following this, I have nearly 40% of my lossless collection in SACD /DSD. It's typically the case for 60', 70', and 80' albums. At this time, most of the masters were analogue, and digital conversion was at its prehistoric era as far as music is concerned. A SACD Version of an album of these years is a complete rejuvenation (special DSD master created directly from the original analogue master tapes...)
So yes, being able to natively play it is a major issue for me. I want to be able to play my music, and I don't want to do anything special to play it on a DAP. I want it to be natively supported, and don't want to spend hours to do file conversion.
And yes when DSD is natively supported by the DAC, it's better sound overall, because it means less conversion, so less noise (remember : conversion = noise). You can hear it comparing X3ii and X5. I will compare X5 with X5ii in this regard.
Bottom line : DSD exists, and is very often the best version avalaible for a music track. the first to natively handle all SACD files, including DST, will gain a new client : me. I am willing to pay 800-1000$ for this.