Abby Normal

While in Arizona giving a presentation I was asked what I meant by voicing, when speaking about DirectStream. The simple answer is voicing refers to the work we do achieving a specific sound in our products. I will go into some depth on this subject, but I should mention the same person asking the question remarked that when he posed that question to a very famous and well respected manufacturer, what he received in reply was a blank stare. Not everyone voices their products. In fact, most electronics manufacturers do not. Or, put another way, many designers of products build to a specific design goal and prove they've reached that goal through their test equipment. If the measurements bear out their design goals, a quick listen should be all that's needed to make sure it performs perfectly and into production it goes. This is the norm when it comes to designing audio electronics. As Marty Feldman playing the role of Igor said, "Abby Normal" is perhaps the way a few high-end designers of electronics operate, by voicing their products. The term voicing no doubt stems from loudspeaker designers who certainly had to listen to their prototypes before releasing them for sale. This because the measured performance of a loudspeaker is pretty abysmal relative to the measured performance of an audio amplifier or DAC. Even the best measuring loudspeakers are several dB off of flat, thus forcing designers to choose which version of not-flat sounds more like music to them. Thus, the sound of a loudspeaker is invariably a combination of the skill of the designer to produce a product that measures "pretty close" to right and whose "voice" has been tweaked and adjusted to sound how they wish. Voicing in this case involves a series of obvious decisions such as crossover points, types of drivers and cabinet parameters, coupled with "if I am willing accept a 3dB error in frequency response, what frequencies should be up or down in level within that window?" But how does this relate to audio electronics where the accepted measured norm is flat to within 1/10th of a dB? Let's find out tomorrow.