Why Do ~You~ Still Play CDs?

@no_regrets I agree.  When I drive I listen to local KUSC which often has new music and most of the music played are newer recordings from the last decade or so.  Thereby, I am being exposed (as much as I want) to new music and new performances.  Sometimes I get excited by a performance and purchase the CD.  So, streaming has a place after I get through about 10,000+ LPs and CDs that I haven't heard in my collection.   I just purchased a collection of about 700 Japanese pressed Jazz LPs (50/50 mono/stereo) and 200 45 rpm Jazz remasters (Kevin Grey).  So, I have much to hear in the next several years that I never encountered (although my Jazz listening best friends have heard them).  

@richardbrand 

Thank you for the kind words! I intentionally left out DSD as I did not want to over-complicate my response. I am a huge supporter of DSD and it certainly deserves a mention as a high-resolution format for its unique approach to audio reproduction. As you know, it is currently only available as downloads.

DSD’s ability to capture micro-dynamics and its “analog-like” character made it a worthwhile endeavor for me. While not without challenges (e.g., noise shaping at higher frequencies), DSD’s purity and simplicity remain compelling for me to keep buying …800 plus albums so far.

@lalitk You mentioned that DSD "is currently only available as downloads".

I guess that was a reference to streaming!  There are over 6,000 SACDs listed at Presto Music if you like classical and jazz.

I find it interesting that PCM bit-depths are multiples of 8, as in 16, 24 and 32-bits. This is more to do with computing than audio.  Around about 1970, computer designs started to coalesce on 8-bit bytes.  An eight bit byte can take 256 values, sufficient to hold a character from upper- and lower-case Latin alphabets, numbers and many other symbols.  

Fast forward 20 years and the need to incorporate Asian and other language symbols forced an expansion from 8-bits - amazingly it was lifted to 16-bits!  This immediately doubled the storage, processing and bandwidth required for text, but meant 256 times as many symbols could be represented. 

Similarly, each eight bits added to the PCM bit depth in theory gives 256 times the number of distinct sound pressure levels that can be encoded, and conversely need to be decoded.  In my opinion, many DACs struggle with 16-bits, in particular ensuring that each recorded increase in sound level results in a monotonic increase on playback.  This is particularly difficult when several bits switch off and a higher bit switches on.

One solution is the delta-sigma approach which in effect locally converts PCM into DSD.

“I guess that was a reference to streaming!”
@richardbrand 

Yes! Correction to my previous post…DSD is available as downloads or physical media (SACD). My go to reference for downloads is NativeDSD and there are few others that are very transparent on provenance of file. If I am paying for music, I don’t buy until I am able to verify the file source. 

I do agree, delta-sigma conversion combines the strengths of both PCM and DSD, offering a practical solution to the limitations of traditional PCM DACs. However, to my ears, their sound is not as natural as DAC’s that uses Sabre ESS Pro Chips. 

@lalitk

"as natural as DAC’s that uses Sabre ESS Pro Chips"

I would add that some AKM DACs also handle 32-bit PCM and DSD natively.  To my ears, Marantz uses these chips beautifully.

I always suggest looking at the data sheet for any embedded DAC to see what it is designed to do, If it doesn’t mention DSD, it is London to a brick that DSD is not natively supported!

DSD can be converted to PCM, and only loses high frequency timing,  Going the other way requires interpolation (guesswork).

The only noise shaping required for DSD output is a gentle low-pass filter in the MHz range, unless you particularly want to fiddle with what the sound engineers recorded! Contrast this with the many filters on offer for PCM.

Happy listening ...