BTW I'm curious how many people enjoy the FIR-2x/4x modes, and does FIR-NOR only use two of the 8 dacs?
on a personal note: I work in the wider field related to acoustics and hearing, but I am not a sound engineer, nor have I experience with designing and implementing filters. I work on the "hardware", especially on the implementation of medicinal drugs for restoring hearing pathologies from a pharmacological point of view... so take my impressions of FIR filters with a grain of salt.
:
I "don't want to apply" heavy FIR filters (or in the DX260, let's say I leave it at "normal" which likely might be the minimum tap length of n=23 for an FIR filter to work as a standard low-pass filter according to Fred Harris rule of thumb?? *). ...from a pure practical point of view, and not mathematical, here's why:
one of the most relevant things FIR filers do is making a frequency response curve smoother. E.g. smoothening the low, and high FQs. The higher the filter value, the more "smoothness" you get from a digital signal (where time and frequency are principally linked). Another by-product is to filter out a potentially present digital noise-floor.... all at the cost of quite significant processing power, resulting in higher temperatures and battery power consumption.
*EDIT: iBasso does not disclose the FIR filter length (taps) it just states "normal", "2x" and "4x"... so anyones guess is just fishing in milky waters.
>>
@Paul - iBasso would your engineers be possible to shine some light ont the iBasso FIR implementation please?
Anyway, what I think it means for us audiophiles with good gear?
When you don't need FIR:
a) if you have a well developed DAC implementation in your source gear, there's likely no audible noise floor by the source itself > so, no 2x or 4x FIR needed
b) if you're happy with your FQ-response curve of your headphones (or source gear as well), why would you like to smooth things out with 2x or 4x?
When you still would benefit from FIR:
a) if you have crappy recordings, or some not so good live recordings, a FIR filter could do some magic there
b) if any gear in your chain produces noise, applying FIR might give better results.
c) if you have a somewhat analytical headphone where you get psychoacoustically annoyed by fatiguing highs, you can try FIR.... but
...d) be aware that also the bass response might get more "muddy" (in terms of reduced frequency separation... which is now smoothened out by the FIR, but you wanted it^^).
But again, from a psychoacoustic point of view, you get the impression that there might be a tad bit more bass and more rolled off highs and your acoustic impression turned into something all so slightly warmer. In general, don't expect FIR filtering to have a massive impact, IMHO it is for sure audible, but still relatively subtle. It's just to average out some FQ response peaks in the time domain, that's all.
