Difference? The name of IC’s are punched or printed or stickered on top of it

Your signal chain performance is limited by the weakest link, and for audio that's generally the DAC and ADC. You need to be careful that the analog portions don't degrade the DAC/ADC, but that's eminently do-able.

The digital/analog conversion process produces artifacts that manifest themselves as spurious signals -- 'spurs'. It's spurs that contribute to your SNR (more serious parts will specify their SFDR -- spurious-free dynamic range). If you were dealing with pure tones, it would in theory be possible to map and ignore or mask spurs, but it's unrealistic in practice, and well-nigh impossible to do for an arbitrary audio waveform.

Bottom line is it's very hard to improve your DAC/ADC from the outside. (If it was easy, everyone would be doing it.) You can add dither to your waveform, which has the effect of reducing spurs, but it comes at a cost of raising the noise floor -- it's a tradeoff that needs to be checked on the bench. If your DAC supports higher sample rates, upsampling can provide better performance, but it really depends on the part -- analog performance vs frequency tends to peter out at some point.

The part you have chosen looks like it's aimed for consumer applications where nobody is going to look very closely at the technicals (it's only specced at 48 ksps). It just needs to sound okay. So if you want better performance, you need a different choice of parts.

"Bypass" in radio circuits generally refers to a decoupling circuit, usually a low impedance capacitor to ground, to allow the high frequency signals to "bypass" the rest of the circuit. It could be a local power supply for an amplifier, or a reference point where you don't want stray RF signals. Or a measure of the signal strength, fed back into the RF stages, to control their gain (AGC) in which case you don't even want audio noise on it - hence the massive C17.

And in this circuit you can see pin 1 is connected to GND via C2, pin 7 via C7, and pin 16 via C16 and C17. The next page of the circuit lists all of these as "bypass" capacitors.

As Peter's answer suggests, some of them may have an auxiliary function : for example, with the waveband switch set to "FM", the "RF Bypass" pin is shorted to ground, which can reasonably be assumed to kill the RF amplifier.

As Pin 2 is the "FM IF input" we can assume the RF and mixer stages for FM are handled by a separate circuit, not this RF stage. (When these chips were made, IC processes were limited in frequency range : they couldn't handle 100MHz signals very well) - in fact the circuits on pages 6 and 8 confirm that.