I have posted before about problems that I encountered when using the LM4562 low distortion op amps, including most recently in the LM358 article in the Lab Notebook.
Just today, I stumbled across what may be the answer to the issue with that chip. It seems that the high value input resistor is not supplying enough bias current at startup and the chip latches. I have not tested this by lowering the value of the resistor, but I plan on putting the direct box pedal back on the bench to see what I can find.