It's not easy.
Let me ask you, what dsp chip supports 384khz or 768khz. And take a step back what src chip supports it? What about DSD input? FPGA? Then how many bands do we need? What precision? Do we also need to be able to upgrade FPGA code? Via microcontroller from XMOS?
Another question what's the percentage of people knows how to use parametric eq? What if 90% of the users don't know how to use and completely jammed up our custome service channel? Do we teach them or give some online documents? What if half of them still can't get it working or don't know what to do? Maybe just ignore them? Then we are off for a big trouble in the future.
And do people really think 5/7 band PEQ is good enough? Don't they want something like 32 band and maybe auto correction from a file or something?
It's not easy. Too much thought are put into this.
All good questions. But a bigger question is, where do you go from here? Functionality is already maxed out as far as this class of product. Performance is as well. What would you do next year?
I think a simplified subset of what you list would be fine. PCM-only to 192 kHz would be fine. If you create an open interface, members could even volunteer to write the EQ front-end software to load the filter parameters into the unit. This could be a hack where one plays a PCM .wav file and it is used as the filter settings.
For precision, noise should be below 24 bit. So if you use 32 bit fixed point math, you should be OK.
On number of filters, 10 would be more than enough. And or a convolution.
Average user doesn't need to know about this feature as the device itself doesn't expose anything other than maybe a switch to turn EQ on and off.
The EQ message is getting stronger by the day. I am making sure of it.
It is key to great sound but access to it is hard. So while you should think hard about our input (as it could be bad for you/your market), I think it requires a serious look. We can provide a lot of feedback on what is needed/not needed per above.