TOPIC:

As english is not my native language I do not always have been able making my intentions clear, I will give it another try:
- my proposal for the FIR blocks to be located on the input is based on the most "natural" use of both IIR and FIR in DSP economy.
- don't ask me to provide the references, but usualy IIR is used for crossover duties and FIR for room correction.
- sure, anybody could use FIR for crossing over but this is somehow a forced take, like using heavy guns to shoot flies.
- my reproach to dev guys is that they first made possible such a peculiar use of the FIR blocks --given the limited amount of available taps-- in detriment of the most usual take for such kind of digital filter. Maybe they wanted to be the first issuing an 100% FIR crossover on the market at this price point, even if it is a very compromised one because of not enough taps and/or lack of decimation.
- I am frustrated also because I already invested some good money in miniDSP hardware which in the meantime got unsupported (miniAmp) or still cannot be used to the full potential (miniShark).
- I am open to spend more than 10 USD for a software plugin to enable use of the available HW in the most efficient way and so my posts are somehow cynical. But I am cynical on my money, so to speak.

In keeping with the miniDSP concept, the option of FIR blocks on the front end (input) does not appear to have materialized as of 2023.  What happened DEVTEAM?        As a suggestion to possibly ease the effort, based on the former HD1 plug-in and how well structured the code is (modularized), could a set of FIR blocks be added to the input path as long as the ground rules are that use is mutually exclusive, meaning one can use the input or output side but not both or a combination of both. Would love to see the 2x4 HD switchable from 96 to 48k internal sampling but another firmware to pair with another modded HD1 is likely asking a much bigger favor.  Thanks for any reconsideration this could be given.  The new Device Console, with its higher handling of multiple devices, is likely of a complexity not conducive to modification of the suggested nature, hence the former HD1 is suggested as where to most easily plant such a mod, not to mention it would maintain support for those who have a large investment in a Windows 7 dedicated digital recording studio.  Again, thanks for consideration.  The miniDSP concept presents economy in hardware, however, the multiple plug-in aspect comes with implied maintenance liability.  Please uphold the concept.

It materialized in the Flex 8.

With the advent of Device Console plug-in, miniDSP is doing a 180 on the concept of one hardware, many plug-ins.  That's a pretty significant switch.  Trying to linearize a THX LFE (IIR 120Hz) crossover from a single powered LF speaker to powered mains for 2.1 requires a linearization FIR on all three (3) outputs.  Splitting 4090 taps 3-ways won't cut the mustard.  (4090 = 4096 - 6 for the unused output then divy up 3 ways)  Flexibility of the 2x4 HD is hampered.  If FIR blocks were on the input then 2048 taps could be delegated to each of the stereo inputs without waste on unused outputs to achieve 120Hz linearization.  miniDSP might also take a lesson from Dirac and provide the flexibilty to drop down to 48kHz sampling to give 4096 taps the ability to better linearize into the low triple, high double digit crossover range.  miniDSP could find a niche in the studio market using DSP to improve performance of mixing and mastering systems on a very attractive investment for improvement gained.  And THANK YOU Thomas a.k.a Pos for a super tool, along with John M. for REW.  The sonic improvement
these tools enable is simply amazing, especially when in combination with acoustic treatment which will set you back a couple of K in a small room/studio.

Your using the wrong dsp for FIR I believe 

Any of the products that run at 96K will have low taps because of bus speed….. that’s the max for that speed. 

most ppl that are “essquues” that believe RCA cables have a sound stage and somehow believe higher sample rate is more “hi res” have been largely miss guided. 

so to make a 96K offering that at least does Hi-Rez hipster market. 


Is an array of OpenDrc boxes for your multi-way surround sound…. Or convolve through a gpu 

the 96k boxes are perfect for the price…..

use OpenDrcs…. Best hardware box ever made 

Presuming FIR blocks on stereo input and 2048 taps per channel at 96k -
96000 / 2048 = 46.875 x 3 = 140.625
Meaning, 46.875Hz is the lowest frequency the FIR will process (linearize) down to.
IIR crossover at 140.625 will enable linearization for 1-2/3 octaves, i.e. down to 46.875Hz before the FIR dead ends.
Linearization of IIR crossover at 141Hz can be achieved gracefully at 96k with 2048 taps.  Using THX LFE 120Hz crossover would be a stretch.  rePhase indicates trivial ripple with extensive optimization to linearize 120Hz 24dB plus 2000Hz 24dB LR crossover slopes, imposing a fraction over 10ms delay.
Granted OpenDRC, operating at 48k with up to 6144 taps/channel, is attrative because it could linearize the full audible bandwidth, i.e. down to 20Hz on 4096 taps/channel clocking at 48k.
Am pondering the audible benefit of linearizing down to 20Hz.  Moreover, am contemplating the benefit of front ending a 2x4 HD with an OpenDRC with the OpenDRC doing linearization only.  If anybody can speak to the audible benefit of full bandwidth linearization, love to hear it!
Tnx!