Take it from someone who has designed op amps into a lot of commercial products plus in-house test equipment for our company. The ideal op amp is easy to define but impossible to make. There are thousands of different types on the market, each with its own compromise of characteristics which are fairly well documented in its 10- or 20-page data sheet. You can find loads of these on, for example, Linear Technology's or National Semiconductor's websites. You would have to understand the circuit and the new op amp's characteristics to be able to make a good match. Simply reading reviews, if there were any, probably would not do you any good when you're trying to come up with drop-in replacement upgrades. Just because someone says "I'm using op amp X and it sounds great!" does not necessarily mean it will work in your circuit if your equipment is a different model.
Op amps are not simply a matter of good-better-best, or an easy tradeoff between quality and price. There are many factors that can even make the choice seem impossible when you consider a lot of different ones and see virtually all the ones you pinned your hopes on eliminated for various reasons, like that they can't handle the voltage range you needed, inputs can't go all the way to ground in a single-supply configuration or the output can't go low enough, there's no compensated version that's unity-gain-stable for low-gain applications, there's no quad version available to drop into the spot that's on your circuit board and meets the requirements for all four sections of the circuit that quad takes care of, the circuit is different from what is required for stable undistorted operation of the particular op amp you're considering, etc..
The tricks to good operation won't necessarily all be in the data sheets either, meaning that once you've gotten experience with a particular op amp and learned by trial and error how to make it do its best under a particular set of conditions, you won't be so quick to switch to a new type. You also need to understand input noise voltage versus input noise current, consider biasing needs (not just input, but sometimes output bias affects things like crossover distortion), possible latch-up situations you could run into from input inversion if the input voltage goes outside the permissible common-mode range in a circuit that was designed for a different op amp, and of course the more obvious things like gain-bandwidth product, slew rate, output drive capabilities, and even whether the power supply can feed a slew of higher-performing op amps.
There are other things besides op amps anyway that can add to the noise and distortion in a circuit. Even something as simple as ceramic chip capacitors which we assume conform pretty well to the simple capacitor law volts=coulombs/farads turn out to be poorly behaved if their voltage rating is not much more than the application calls for. X7R dielectric is not nearly as great an offender as Y5V, but can still produce objectionable distortion if the designers tried to cut costs or reduce the size too much. The April 12, 2007 issue of Electronic Design News (EDN) had an article starting on page 77 about the nonlinearities of ceramic chip capacitors.
If I haven't convinced you yet to leave it alone, one op amp I like for audio applications where I can supply at least 12V is the LT1124/1125/1126/1127 family. But even if the price were $.25 each instead of $10 each, it still wouldn't be right for every application.
, but I do appreciate the response. You obviously know alot more about op-amps than I could ever wish to.