The main sources of noise in this type of devices:As you see, we get noise reduction from 80 to 400 Hz but above that, the headphone electronics is injecting its own noise. Shame that we are stuck between having good frequency response and noise.
- NC microphones have a finite SNR from 63dB (every other maker) to 73dB (a select few like Infineon and Knowles) so they generate their own noise...
- Portable electronics also have a finite SNR, Bose usually uses Qualcomm BT IC (QCC51xx) and ADI DSP (ADAU 1777, 1787 and newer)
- NC function in particular the Feedback (FB) loop (as opposed to the feedforward) where you need to amplify the signal from the FB mic by 15-25dB
- The FB loop is not perfect: example of the QC35.2 anything above 0dB will actually amplify the noise!
BTW QC35.x are better than the 45 and 700 in terms of NC.
So yes, for now, you have to trade SNR for accuracy to get portable devices offering NC with decent battery life/weight.
Accuracy is NOT an issue as BT ICs and any DSP will have more than enough EQ capabilities to match pretty much any target.
Therefore no EQ for those.
Score no EQ 55.3
Score with 75.5
Regarding the L/R imbalance that is too much i.e. it barely meets the industry low standard (within 6dB) and way too much for a large maker.
It should be within 3/4dB max up to 5kHz I reckon, above both diaphragm break-up and cavities modes make it trickier.
In the old days prior to iPhone, Nokia bundled earbuds that were within 2dB up to 8kHz and would still be industry leading today.
I would not be surprised at all that someone could pick that up, QA apparatus can 100%, and trained personnel definitively would...
So it is what is it because it fine by their standards and might be related to built-in L/R issues of the design and consistent across production.