Paul McGowan writes:
In our post on Software Jitter I got several emails from folks telling me “that’s an obvious distortion path”. True enough but it sure wasn’t before we found it.
Fact is we’ve been struggling with trying to figure out why small changes in the code that runs our products have an impact on the sound quality – why folks love one version of software and dislike another. There are many great mysteries concerning sound quality and they only become obvious once you figure them out.
I remember when we first learned why most IC op amps sounded worse than their discrete counterparts – something still true today.
On the face of it, an integrated circuit amplifier should be better than the same thing built from individual parts spread out over a circuit board. The IC has many advantages: one piece of silicon, perfect temperature tracking, precise values of related components, no physical leads, short connections between components and so on.
Discrete op amps should (and do) have many flaws: multiple solder joints, physical leads and connections, PC board traces, parts variability, temperature based differences between components, etc.
Yet consistently the simpler discrete amplifier circuits sonically trounce most IC amplifiers that have much better measured specs.
Then along came Mattie Otala who started talking about new forms of distortion like TIM and SID – terms we’d never heard or even thought of. Without delving too deep it turns out that the ability to control how well an amplifier operates without feedback, something easily handled in a discrete amplifier and generally impossible in the available IC’s of the day, has everything to do with how it plays back music.
Once we understood the phenomena we could control it. If we wanted to take advantage of an IC amplifier we knew what to look for – or we could just stick with what we were using already, discrete op amps – but now we knew why they sounded better. This knowledge helped raise the sound quality of everything solid state in high-end audio from that point on. In fact, once high-end audio designers unravelled all of this, we turned the corner and started closing the gap between tubes and transistors. It was a pivotal piece of information.
Heck, we should have known because it was obvious!