Different body and different legs but same 517 date code. Sloppy logo on both of them. Cue in the Sesame Street "One of these things is not like the other" song.
I don't have any ST 741s but here is a real TL082 with date code 0322.
Hard to tell full date code of the 741s as ST only marks three digits on the parts, there is no decade indication.
You have "517" -- this can be 0517 or 1517.
e3 logo as added in the mid 2000s as part of their RoHS initiative.
Here is a close up of real TL082IN. Note the pin 1 indicator (dot) on these.
This pin 1 indicator may or may not be present and depends on the lid lot that they had at the time.

TL082 (resized).jpg

U8 absolutely, positively must be the unbuffered type.
They do not use the 4049 as an inverter -- they only use one FET transistor within the IC as part of their mixer. If the part is buffered then the FET transistor that they use is isolated and won't work.

CD4049BE is buffered. CD4049UBE is unbuffered. Someday, if you want a real head scratcher, take a look at the U8 connections and compare to pinout of 4049. Redraw internals to match the pinout and you will see they are using a FET transistor as their mixer. As to why they didn't just use an off the shelf FET transistor to do this - we will never know.

Only a few manufacturers made the buffered version and nobody makes the buffered version any longer.
Note that the old "A" revision parts didn't have this issue. The buffering added by the "B" revision caused the issue which forced manufacturers to remove the buffering on the outputs of these. If I remember right - only GE/RCA (now TI) and Motorola came out with the buffered parts at first. Designs had issues (especially oscillators) and GE & Moto quickly came out with an unbuffered version. GE quickly stopped making the buffered version, Moto continued to sell two versions: 4049UB and 4049B. ... or was that vice-versa where Moto stopped buffered parts and GE sold both? Those that started selling the B series later (e.g. Fairchild, ST) completely skipped over the buffered version.

As Quench mentioned - did you check C8? And did you check your sustain circuit to make sure the gate input to the mixer isn't dead (pin 3 of U8)?