technical issues

MP3 Encoding
Preparing for the roll-out of this site we spent days researching - talking to musicians, engineers and wading through reams of Opinionated Yak on the internet. It is shocking how much seemingly purposeful misinformation and bad aesthetic judgment there is Out There. Finally, Paul Hamann and David Thomas engaged in some empirical testing at Suma. We can't stress this point enough: not one of the decisions they made was a close call. The conclusion to be drawn from every single A/B test they made was apparent within seconds! One can only assume there are thousands of cloth-eared and/or cynical "experts" out there.

Stereo Imaging.
We tested "Joint Stereo" versus Interleaved (or "Normal") stereo files. Interleaving yields dramatically superior results for studio recordings. For live cassette or faux-stereo recordings Joint Stereo can sometimes be adequate.

Sampling Frequency.
Didn't test it. It's possible that we could cheat down to 22khz for cassette recordings. Why bother?

Encoding Quality Setting.
Didn't test it. Again it's possible we could cheat down the scale from High but on a practical level that would necessitate hours of frustrating A/B testing for every release.

Kilobyte Per Second (KBPS).
We started at what is considered to be the minimum quality setting, 190kbps. Yielded grainy results and too much distortion. Went to 240kbps. Better but not adequate unless we work hard to deceive ourselves. Top level of 320kbps was adequate. The throughput and bit rate scheme are intertwined factors so the testing process was not a simple straight-line affair. Again, it's possible that an endless and impractical tweaking process between all the variables involved might, in the end, yield a slightly smaller file size at a comparable quality but the advantage would be relatively insignificant in the overall scheme of things compared to the amount of work involved and, more importantly, the opportunity to make bad decisions.

Constant Bit Rate (CBR) versus Variable Bit Rate (VBR).
We tested all the varieties of VBR along with variants of KBPS. At every level CBR was superior, most obviously in the depth of field. At every combination VBR yielded flatter audio reproduction. Since the meaning of sound is encoded primarily or significantly in its spatiality this sacrifice seemed to us to be madness itself. That said, a high quality VBR setting can be used for live cassettes and still yield adequate results.

Conclusions.
It's clear that the only MP3 encoding method that can yield adequate results is an interleaved stereo file encoded at a Constant Bit Rate of 320kbps. This yields the largest possible file sizes but also allows for audio reproduction comparable to a vinyl pressing.

Now, be clear on this point, "vinyl quality" is, in Ubu-speak, a derisive expression. "Gee, that sounds as good as vinyl" is another way of saying, "Gee, that sounds like a dog's dinner." The putative "warmth" of vinyl is another one of those mass-hysteria hoaxes that the human race is prone to. "Vinyl warmth" is not some semi-mystical, undefinable phenomenon. There is actually a technical term that audio engineers have for what you are hearing - it is called distortion. The bottom end is distorting. Now, distortion is a valuable audio tool, and an Ubu favorite, but only when the distortion is distortion we choose. You may like the phenomenon but it is NOT what we wanted and it is NOT what we heard in the studio.

It is possible to cut 12-inch vinyl with music that's been produced by post-1970 recording techniques as long as each side isn't much more than ten minutes in length and as long as you play by the rules. It is, however, easier to produce far superior compact disks using up-to-date techniques of analog tracking and high quality digital sampling - without the gratuitous distortion, overwhelming surface noise, and oppressive mix restrictions that vinyl imposes. Follow this link for more on this issue.

Note that we have had long experience of the Vinyl Regime. When we were producing vinyl our records were mastered and cut by the best engineers in the business. Every single one came back a bitter and humiliating disappointment.

Every medium is a compromise. The current state of digital technology is nowhere near the end of the road but it, and the compact disk as a delivery system, marks a progression, an advancement, and has opened dramatic and exciting narrative possibilities for the language of sound. Download audio is a step sideways, at best, but it is on a par with vinyl reproduction and therefore an adequate medium for release. Every medium is a compromise.

A brief note concerning AAC encoding.
We prefer AAC encoding but it is not enough of a standard at this point and our lives are too short to go about offering multiple download formats. Yes, we know about FLAC, and every other encoder you want to tell us about. For the record, FLAC files are a pain in the butt because of the portability issue. We don't want to waste away our lives telling people how to get them to play. MP3 is the standard. Like it or not.
Some sort of LossLess compression - FLAC or Apple LossLess - will be explored for special releases. Frank has placed a bet that people given the choice will still choose MP3. We'll see.
garage icon
An Irony-Free™ Site