Sunday, November 10, 2013

Alternate explanation of Denon al32 interpolation provides useful insights

Denon's Advanced AL32 Processor expands audio data to 32 bits and uses a proprietary algorithm to interpolate the data and perform up-conversion and sampling, achieving a playback sound that is close to the original source.

Since high-performance devices capable of large-capacity processing read data samples across a wide spectrum and process them in a single stage, they interpolate signals with greater precision compared with multi-stage digital filters and other such devices. In addition, the use of algorithms ideal for frequency characteristics outside the audible range to filter sudden bursts of musical data or continuous sound at high frequencies protects sound quality from the adverse effects of aliasing noise or drops in high-range response.

The Advanced AL32 Processor reproduces the delicate nuances of music, as well as spatial information such as the position of the artist and the breadth, height, and depth of the stage, in a more natural manner. High-precision 32-bit, 192-kHz D/A converters have been used to bring out the maximum performance of the Advanced AL32 Processor.
from Denon Asia DCD-2020AE page

I've taken a circuitous route in an attempt to understand the quality of sound I heard from the CD layer of the BIS Records' hybrid disc of Beethoven's 9th symphony. I assume that the CD player was a Denon, since Denon is a favorite of broadcasters. Since the Denon explanation didn't satisfy me, I researched advances in CD playback in general, and hit a wall when trying to perform research into Cambridge's ATF2 technology, which seems to imply that it corrects for errors in the INPUT data by performing a curve-fitting operation, which has a tinge of hype to it.

So, I went back to researching the Denon system, and found a better explanation on the Denon Asia website, which I excerpted above. It seems to indicate that the audible superiority of al32 is due to the use of a DSP to implement a one-pass digital filter with high precision and with a characteristic optimized for its transient response, which is very similar to, if not exactly the same as, the approach used by Ayer Acoustics (whose products are highly regarded by the likes of Stereophile) as explained here. The benefit provided by using a DSP chip therefore boils down to greater precision, and perhaps the ability to implement a more complex and accurate algorithm, and not, as I was assuming, the difference between linear (or no) interpolation, and interpolation based on digital filtering.