DSD and PCM - Horses for Courses

AR-DSD2a.jpg

Lately audiophiles have been hearing a lot about DSD, both pro and con. Some record labels, such as AIX Records, who've been recording in high-definition PCM for years, see DSD as a competitor to PCM and have posted a articles stressing DSD's shortcomings. I think this is a giant waste of time and effort.

No recording medium is without intrinsic issues that limit fidelity. Whether it's analog tape's tape scrap flutter, or vinyl record's off-center pressings, record warp, limited S/N, PCM digital's timing, sampling and quantization errors, or DSD's spurious noise at extreme upper frequencies and limited editing capabilities, no media is perfect. The recording engineer's ultimate goal isn't to enumerate each medium's problems, it's to work around those problems to achieve the best sonic results.

PCM and DSD should and can co-exist in the modern world. Each has certain advantages over the other, primarily based on what they do best. DSD is the best way to archive analog recordings if you need maximum flexibility to release in different digital formats. DSD can be converted easily, without decimation errors into any PCM sample and bit rate. PCM is not so flexible.

In the early days of digital recording the only commercial release format was redbook CDs at 44.1/16. Whatever recording medium an engineer used, it would eventually have to be converted to CD. While 88.2 kHz easily converts to 44.1 kHz, 96 kHz does not decimate cleanly, generating rounding errors that effect fidelity. Some engineers, such as Stereophile's John Atkinson, recorded at 88.1 instead of 96 kHz to avoid that issue. Nowadays digital files can be recorded and released at any PCM format from lowly 128 BPS MP3 all the way up to 192/24 (yes, 384/32 is a possibility, but not yet commercially realized).

AR-dsd3.jpg

It has gotten to the point that an engineer can choose the sample-rate and bit-rate for a PCM project based on workflow and fidelity rather than final format considerations because any recording can be released in its "native" or original format. If a recording is made at 96/24, it can be released at 96/24. If a recording is made originally in DSD, in can be released in DSD. There's no reason to convert a high resolution PCM file into DSD or convert a DSD file to PCM for final release. Native rate rules.

If a project requires multi-tracking, dubbing, and EQ corrections, PCM is a far better recording choice than DSD. When I record my bluegrass band we use PCM, click-tracks, and multi-tracking because, frankly, we're not good enough to do an entire song in one take perfectly. Many bands are in a similar position. Other reasons for using PCM include recording situations where instruments can't be recorded all in one take such as when a drum track is needed behind an acoustic guitar. Drums are so much louder that they bleed into the guitar's microphones, which requires either isolating the drums (sometimes in another room in the studio) or forcing the drummer to play at unnaturally low levels.

DSD is best for live recordings of acoustic concerts. If you only need two tracks, plan to mix down (if needed) real-time and don't need to use any overdubs or additional tracks, DSD is the way to go, in my humble opinion. I've been using DSD for live classical concert recordings since November 2007. With DSD I can easily generate a 44.1/16 version as well as 96/24 and 192/24 files. With even the simple the editing program AudioGate I can do edits, fade-ins and fade-outs, and all my PCM conversions (if I need to).

AR-dsd1.jpg

PCM and DSD are different and I believe complementary recording methodologies. Smart audio engineers and producers must consider both before choosing one or the other for a recording project.

Perpetrating the idea that either DSD or PCM are technically "better" than one another is a waste of intellectual horsepower. The debate should be how to use each format to achieve the best sonic results. If less time was spent elucidating format issues and more time finding solutions that best utilize current recording methods, the results would be better recordings and less time wasted championing things that don't need champions.

comments powered by Disqus

Audiophile Review Sponsors