Quibbles and Bits

Sigma–Delta Modulators – Part III

At this point I promised to conclude my mini-series on SDMs by touching on some of the differences between DSD and PCM formats.  Much has been written about this, and it can tend to confuse and obfuscate.  On one hand, with a PCM data stream, the specific purpose of every single bit in the context of the encoded signal is clear and unambiguous.  Each bit is a known part of a digital word, and each word stipulates the exact magnitude of the encoded signal at a known instant in time.  The format responds to random access, by which I mean that if we want to know the exact magnitude of the encoded signal at some stipulated moment in time, we can go right in there and grab it.  Of course, when we say “exact” we understand that to be limited by the bit depth of the PCM word.

The situation with SDM bitstreams is slightly different, and I will illustrate this with the extreme example of a DSD 1–bit bitstream.  On one level, we would see the DSD bitstream as being exactly identical to what I have just described for the PCM case.  Each bit is a known part of a digital word, except that in this case the single bit comprises the entire word.  This word then represents the exact magnitude of the encoded signal at a known instant in time – but to a resolution of only 1–bit.  That is because the DSD bitstream has encoded not only the signal, but also the heavy dose of shaped Quantization noise that I have been describing in noxious detail.  That noise gets in the way of our ability to interpret an individual bit in the light of the original encoded signal.  By examining one bit in isolation we cannot determine what part of it is signal and what part is noise.  It is exactly like looking at one single ballot paper and attempting to draw conclusions regarding the outcome of the election.

If we want to extract the original signal from the DSD bitstream, we must pass the entire bitstream through a filter which will eliminate the noise.  And because we have already stipulated that the SDM is capable of encoding the original signal with a very high degree of fidelity, it stands to reason that we will require a bit depth much greater than 1-bit to store the result of doing so.  In effect, by passing the DSD bitstream through a low-pass filter, we end up converting it to PCM.  This is how DSD-to-PCM conversion is done.  You simply pass it through a low-pass filter.  The resultant PCM representation can be very close to a perfect copy of the original signal, limited only by the accuracy of the low-pass filter used.

Unlike SDMs, digital filters are very well understood.  There is virtually no significant aspect of a digital filter’s performance which has not been successfully analyzed to the Nth degree.  The filter’s amplitude and phase responses are completely known.  We can stipulate with certainty the extent to which, in any given implementation, computer rounding errors are going impact the filter’s real-world performance, and take measures to get around that if necessary.  In other words, if we know what a given filter’s input signal is, then we can determine exactly, and I mean EXACTLY, what its output signal is going to be.  SDMs, as I have attempted to describe above, are not like that.

So, for some, we finally come to the $64,000 question – what does all that mean for the DSD-vs-PCM argument?

I cannot offer you a simple gift-wrapped answer to that.  I still have a lingering preference for the sound of DSD over PCM, although the technical arguments offer no sound basis upon which to stake an absolutist position.  Also, there are some absolutely stunning 24/352.8 PCM recordings out there from Northstar Recordings that may just be the finest I have ever heard.  Whatever….

In the meantime, I offer the following talking points:

DSD is primarily listened to by audio enthusiasts.  The market for DSD comprises people who like music but still desire to hear it well recorded.  It is still a small market, and it is served almost exclusively by specialist providers who are happy to put in the time, expense, and inconvenience required to generate a quality product for their customers.  People like Cookie Marenco at Blue Coast Records, Jared Sacks at Channel Classics, Morten Lindberg at 2L, Todd Garfinkel at MA Recordings, Gus Skinas at Super Audio Centre and many others focus on delivering to consumers truly exceptional recordings of uncompromised quality.  All this despite the fact that DSD imposes some very severe limitations on what a record producer can do in his or her studio.  Surprisingly, maybe, there are even signs that it is becoming the norm these days for high end classical music to be recorded in DSD – and as often as not in DSD128 (and even DSD256).

There are, in my view, three main factors at play.

First, tools do not exist to do even the simplest of studio work in the DSD domain.  Even fundamental operations like panning and fading require conversion to an intermediate PCM format.  Forget added reverb, pitch correction, and any number of studio tricks of the Pro-Tools ilk.  All that stuff, if done at all, has to be done in the analog domain.  Recording to DSD forces recordists to strip everything down to its basics, and capture the music in the simplest and most natural manner possible.  That alone usually results in significant improvements in the sort of qualities that appeal to audiophiles.

Second, when remastering old recordings for re-release on SACD, for digital download as DSD files, or even for archival purposes, mastering engineers will typically pay a lot more attention to many of the fine details that would normally be dismissed for a commercial CD release.  There will be no product marketing types peering over their shoulders, waving their MBAs and demanding “More compression! More compression!”  The mastering engineer will get the opportunity to dust off that old preamp he prefers to use, or those old tube amplifiers that he only brings out when the suits from the label are not prowling around.  Try listening to Dire Straits’ classic “Brothers In Arms”, which sounds a million times better when specially remastered for SACD (I particularly love the Japanese SHM-SACD remastering) than it ever did on any CD, even though the master tape was famously recorded in 16-bit PCM.  Go figure.

Third, unless you have one of the few remaining ancient Sonoma DSD recording desks, if you are recording to DSD the chances are you will be using some of the latest and highest-spec studio equipment available.  That’s where the DSD options are all positioned.  You will probably be using top-of-the-line ADCs, mics, mic preamps, cables, etc.  As with most things in life, you tend to get what you pay for, and if you are using the best equipment your chances of laying down the best recording can only improve.

So I like DSD, I continue to look out for it, and it continues to sound dramatically better than the vast majority of CD audio that comes my way.  Is that due to some fundamental advantages of the DSD format, or is it that PCM offers a million new and exciting ways to shoot a recording in the foot?  I’ll leave that for others to decide.