Running interference

Join Our Community Subscribe to Paul's Posts

I have long been fascinated that reviewer Michael Fremer finds PCM copies of vinyl listenable yet not CDs. Both are employing the same digital technology, yet one is listenable while the other not. How can that be?

And here’s another question. When mastering engineer Gus Skinas makes a transfer from analog tape to digital it is always DSD first, PCM second. And his transfers always outperform everyone else.

The best sounding versions of reproduced music have first passed through the analog lens of either DSD or vinyl.

What intrigues me is the why behind these observations. What are the two aforementioned lenses—DSD and vinyl—doing (or not doing) that PCM alone fails at?

One possibility is high frequency interference. We know that PCM is quite unhappy with incoming frequencies outside its abilities. Steep filtering is required so aliasing doesn’t occur. In the case of CD quality PCM that filter is very steep making sure nothing above 22kHz gets through to cause problems. In the case of DSD that filtering is quite a bit higher and less aggressive—50kHz if memory serves—and far less important to eliminate for proper performance. In other words, DSD seems less bothered by high frequency interference. Vinyl, of course, could care less because it can’t record it anyway.

Could it be that the gentle prefiltering of ultrasonics offered by both vinyl and DSD playback perform a service to a PCM recorder we hadn’t considered? Do they somehow shape the sound in such a way that PCM can then perform at its best?

I want to do some more research on the subject and will report my findings in future posts.

For the time being, I will suggest that the lens of both DSD and vinyl are enough to eradicate what PCM does not like—which is why DSD and vinyl recordings have a type of “sound” we prefer over PCM recordings that have not been first captured and played back through them.

There is much more to this story to come. I hope to continue my thoughts and research.