The difference between 'real world' and 'chipset' performance
In deciding between two similar interfaces, the often hard-to-decipher performance specifications play an integral role for the audio engineer. And here’s the problem – you would be fooled if you thought that these performance figures were reliable. Because companies do not adhere to the same measurement standards, their performance figures often misrepresent the true performance of a product. Digital dynamic range figures are a classic example of this. Worse than this, some companies in the pro-recording industry quote figures from the data-sheets they receive from the manufacturers of the silicon devices they’ve selected for their product. These figures are extremely misleading, since once placed in the environment of a product, with essential circuitry around the chip, performance figures change for the worse.
Herein lies the real talent of good product design – to get the most out the chips used. Focusrite R & D have found that the difference between real-world and chipset performance can be as much as 12dB dynamic range. In other words, a chip's D-A conversion performance might be specified at 114dB by the manufacturer, but when used within an audio interface could perform at only 102dB. That difference of 12dB is equivalent to a reduction from 19bits to 17bits – essentially, the studio is being short-changed by two bits!
The AES17 standard provides a set of guidelines on how to accurately measure the 'real-world' performance of audio equipment. Focusrite measure to this standard, but many interface companies do not. So when you’re told that a product does ‘exactly what it says on the tin’, treat these words with caution – appearances can be deceiving.