Mythbusters: CDs Optical Properties Affect The Sound?
I came across the following claim in an aging post about very expensive “$2000 glass CDs”. The author, a well-known audiophile writer, insists that the optical properties of a compact disc can affect the sound of the resulting playback. “I’ve long been fascinated by the idea that a CD’s optical properties can affect the sound in an analog-like manner even though the datastream remains unchanged,” he wrote. Regular readers will know that I’m of the camp that says that the same digital bits output from a glass disc or a conventional CD will result in exactly the same audio fidelity…provided the clocks are completely reconstructed as they are on high-end equipment and the rest of the signal chain is the same. This viewpoint is the basis of digital audio and everything else stored as digital information. Unlike vinyl LPs or analog tapes, the medium has no impact on the delivery of the message.
Is it possible that audible changes can result by using glass instead of polycarbonate? Can comparisons such as the following really be true? “I hate to rely on that old cliché in describing improved digital sound, but the glass CD sounded more ‘analog-like.’ The glass CD was smoother, more spacious, more open, deeper, and had greater ease. By comparison, the polycarbonate CD was flatter and had less air between images; instrumental textures were less natural, sounding slightly synthetic by comparison. The polycarbonate CD by contrast overlaid timbres with a patina of glare.” Who writes like this except audiophile reviewers?
This whole debate requires a firm understanding of how a CD or any optical disc works for the matter. I think most of us have a basic understanding of how a compact disc is made but I think it’s worth explaining the process once again. The digital information from a digital master tape (Sony 1630) or digital audio file (DDP) is used by an LBR (laser beam recorder) at the mastering room of the replication facility to create a glass master. The LBR burns away very small pits in the surface of a pristine glass disc, which is then used as the master in a sequence of steps that results in a stamper. A metal stamper is good for about 15,000 disc pressings. However, contrary to the author’s claim that, “Digital data are stored on the CD in ‘pits’ (indentations in the disc) and ‘lands’ (the flat disc surface),” a series of manufacturing steps actually results in exactly the opposite physical arrangement on the surface of the discs that we play. They are bumps NOT “pits”.
A stamper is used in the pressing lines to replicate discs. Polycarbonate pellets are heated and fed into the pressing machine where the stamper is slammed up against the material. The hot plastic forms a disc with millions of tiny “pits” in it. At this point, the new disc is completely clear. The disc is “sputtered” with aluminum to make it reflective and then a thin layer of additional clear polycarbonate is layered on top by dropping a small bead of liquid plastic and spinning the disc rapidly. The disc is then turned over and the disc artwork is silkscreened onto the surface. The disc when read by the optical pickup in your player shoots the laser through the clear layer and hits the “bumps” or the “lands” between the bumps and is interpreted as the two states required for digital information. There is a complex process of conversion to the actual digital stream but it not affected by the surface of the discs.
To be continued…
14 thoughts on “Mythbusters: CDs Optical Properties Affect The Sound?”
You wrote, “provided the clocks are completely reconstructed…” Various jitter reduction schemes have been in use for decades, but having the timing of the data at the DAC completely independent of the transport is only recently becoming common, and, sadly, is still not universal – as I pointed out in a previous post. Did the reviewer specify on what equipment he heard this difference?
Reclocking has been around for a very long time. My point is that reclocking or rebuilding the clock separate from the incoming stream removes any chance for jitter or timing errors to cause the difference…and I don’t believe there is. The author didn’t mention the equipment.
Re-clocking is a vague term that encompasses a variety of methods. A well-designed PLL or asynchronous sample rate converter can significantly reduce jitter, and using a buffer can eliminated entirely. If, however, the DAC in question doesn’t have good jitter immunity, or it’s a single box CD player with inadequate electrical isolation of the transport from the DAC and analog sections, there is a clear path for changes in the operation of the transport to affect the sound quality.
The reviewer may, indeed, be blowing smoke. But, without knowing the equipment used in the test, his observations are not helpful, and your grounds for disregarding those observations rest on a lot of assumptions.
Seems to me I remember back in the day some super ears were cryogenically freezing discs and claiming all those flowery improvements to them too. Something about microscopic alignment of the pits making for a smoother more analog sound. 🙂
Not directly related to today’s post, but significant nevertheless: from Cookie’s site:
Provenance: Pacific Premieres was originally recorded to 9632 PCM (9632 is shorthand for 96kHz, 32-bit (float) sampling). The 9632 WAV files are the original digital file generation received from the artist or label and are not available for sale.
The DSF, WAV and FLAC files are considered second generation and are conversions made using our proprietary methods.
Thanks Alan. There is no such thing as “second generation” digital. Generation implies loss at each transfer or conversion. Digital doesn’t experience loss when sticking with the same specifications. If Cookie’s technique is converting a file from PCM to DSF, then there will be some change in the data stream but it’s not the same thing as “generation” loss. I haven’t looked at the Blue Coast Records site in a while but would like to know more about originally recorded at 96 kHz/32-bits. I’m not aware of very many ADCs that were running 32-bits back in the day.
Don’t forget about the tweak that involved marking up the inner and outer edges of a CD with a green marker….It made SUCH a difference! 😉
Was that description of CD manufacturing yours, or was it from the post you were quoting? It contains so many errors about the process that I hardly know where to start…
I’m interested in what you would argue with…
I found the article about the glass CD. It’s an interesting concept. I haven’t yet found any details on how they physically produced the pits on the disc surface – engraved, or pressed into a thin plastic layer on the glass, or “burnt” into a photosensitive layer like a CD-R? Also, the glass they are using would be a special order. Most optical glasses have a lower refractive index than polycarbonate, so the glass would need to be thicker to properly focus the laser on the pits. A thicker disc could cause problems with the clamping mechanism.
Now to the arguments… 🙂
“The LBR burns away very small pits in the surface of a pristine glass disc…”
It’s more complex than that. It is a photographic process. The laser “exposes” what will become the “pits” onto a photosensitive layer coated on the disc. “Developing” the layer and etching causes the areas that will become the pits to be etched away. There are a couple of diferent processes in use, but the end result is physical “pits”. I’ll grant that you may have oversimplified rather than be wrong. 🙂
“However, contrary to the author’s claim that, “Digital data are stored on the CD in ‘pits’ (indentations in the disc) and ‘lands’ (the flat disc surface),” a series of manufacturing steps actually results in exactly the opposite physical arrangement on the surface of the discs that we play. They are bumps NOT “pits”.”
In all cases, the physical arrangement on the surface of the final disc is a series of pits.
From a reading point of view it doesn’t matter if the result is “pits” or “bumps”. It’s just that when moulding the final disc, the plastic flows better around the “bumps” on the stamper than it would fill “pits” in the stamper.
“… the pressing machine where the stamper is slammed up against the material. ”
The stamper forms one side of a disc-shaped mould, which stands still while liquid plastic is pumped in.
(Unlike pressing vinyl, where a piece of solid vinyl is squeezed between two hot stampers.)
“… a thin layer of additional clear polycarbonate is layered on top by dropping a small bead of liquid plastic and spinning the disc rapidly. The disc is then turned over and the disc artwork is silkscreened onto the surface.”
A thin layer of UV-curing lacquer is used, not polycarbonate.The lacquer is applied to the plated side, which is the same side that is printed, so the disc is not turned over.
Regarding the way two discs can sound different, it’s pretty well understood that if you use a DAC that is not affected by jitter, the differences disappear. There’s no excuse nowadays for a DAC (separate, or built into a player) that is audibly affected by jitter at any level below that bad enough to cause actual bit errors.
Thanks Don, I feel better.
I found a description of the glass CD pressing process.
A metal stamper is made with the same process as used for polycarbonate CDs. A glass disc is coated with a soft plastic resin and pressed against the stamper. The resin is hardened by UV light. The disc is then plated, coated and printed in the same process as a polycarbonate CD.
It’s hard to see how this could produce a CD with the better pit geometry etc claimed. Most of the geometry is determined by the quality of the steps prior to pressing.
– Glass master exposed and etched (pits).
– Electro-deposit nickel on the glass master and peel to create a metal “father” (bumps).
– Electro-deposit on the father and peel to create a metal “mother” (pits).
– Electro-deposit on the mother to create a “son” or stamper (bumps). (Repeat if more stampers needed.)
– Mould into plastic to create the CD (pits).
Also worth noting, neither the pits nor lands are ones and zeroes. It is the transition between them that is read as a one. Whether that transition is pit to land or land to pit. You get zeroes always until a transition occurs.