What are the really good opamps for audio, that don't cost stupid money and are available in DIPs not just SMD?

I'm interested in what people think,especially if you've used them in your own circuits.

I think the TL071 is really really good, but it sucks quite a bit of juice.

All of the Burr Brown stuff (now Texas) is meant to be great, but it's pricey.

The idea in asking this, is that I've realised that in many audio applications, because of the way opamps are used, you can improve the fine details of performance for your whole circuit just by using a different type (I mean overall current drain, output offset, noise etc).

 

An OPA134 @ $2.45 is not expensive. Not when you consider a distortion figure of .007% and an input impedance of 10^13th. You can probably get LM741's cheaper, but not quite the same performance. Run off an AC supply and forget worrying about power consumption. I lose more in heat off the regulators than my preamp consumes (kind of old-fashioned that way).

 

If you want low power then use a horrible LM358 dual or LM324 quad.
They are low power (the LM358 uses 40% of the current of a dual TL072) but have severe crossover distortion and have a response to only 2kHz (the TL072 has very low distortion and goes to 100kHz).

The TL07x opamps have a problem called "phase inversion" where the output suddenly inverts when an input goes lower than the allowed common-mode voltage range. I have not seen the problem occur in many audio circuits.

 

I'm coming to the conclusion that higher power consumption is the tradeoff for top audio quality. I'm probably going to stick with a TL074 (quad TL071) and some other single op amps.

That phase inversion problem is interesting, I guess that wouldn't show up in SPICE.

I tried out an LF444 I think in SPICE and it was terrible, the power consumption was tiny and the output offset for unacceptable. Everthing I've tried that starts LM or LF hasn't been that good in simulation.

I guess I'll decide after a breadboard check and getting out my USB scope. I'm gonna have to make a handy signal generator for testing, probably a variable frequency wienbridge will be enough.