Originally Posted by joeyjojo
Lots of good posts and well thought out views, thanks all. This thread took a bit of a dip in the middle (although I learnt a lot about nature, that slug video is amazing) but looks like we're back.
I agree (I don't think anybody disagrees) that you've got to listen to gear as well as look at the specs and make measurents. I might be inclined to say though that in recent years, with technology reaching a state where the theory and the production capabilities are so mature, the differences are getting smaller and smaller, and I might even go as far as to say that eventually those differences will become imperceptible. This applies to something like a DAC or an amp, and not really to speakers, as the former are meant to be as neutral as possible (IMO).
An interesting case though I think is the beta22, which I write about a lot at the moment as I built one this year. This from the History page:
Italics added by me. Many people contributed to the beta22 of course, and Ti and others have obviously built and heard previous versions, but I find it interesting that until the first prototype PCB was printed up and populated Ti hadn't actually listened to it, but rather analysed in a simulation the output. I add emphasis to "quite" as we don't know just how reliable the simulations were or how drastic the changes after hearing the prototype, but I suspect they were very small and the topology changed little, if any.
As technically probably the best amp in the world I find it worthy of note that the beta22 was designed "earless" and only listened to later, and then found (fortunately!) to sound damn good.
But there are quite a few people out there that do not believe you have to listen to a DAC before you buy it: "if the specs say it is perfect, then it must be". Caveat emptor.
I have worked on design teams where the simulations we performed DID NOT not reveal all the problems we actually had when we actually built and populated the PCB.
Amplifiers would oscillate, microprocessors would not boot up, the equipment would conduct and radiate excess noise, power supplies would oscillate, user interfaces would be awkward, etc.
There is absolutely no substitute for building a working prototype (not a breadboard or whatever, a real working prototype) and testing the hell out of it.
A manufacturer would have to be insane to simulate, then go straight into production.
What changes are made when going from simulation to prototype to production? The true real world answer is: IT DEPENDS.
As Currawong pointed out, techincally the Benchmark DAC1 is a technical tour de force, but some folks just don't like the sound of it.
Check out some of the comments that Bryston equipment gets on the internet, technically they make great stuff, but some folks just find it to be too analytical, i.e they don't like listening to it.