My humble apologies, somehow I misread or misunderstood your post, thinking you meant get a second mike even if a cheap one. How I concluded that from what you wrote I have no idea as it diesn’t Suggest that at all. But the point is as well established.
And yes, a boom stand - and indeed that need be nothing fancy
No worries at all!
I finally got round to doing the A-B between Unitiserve and Roon (test track was Even Less by Porcupine Tree hi res rip). First time my son and I thought The Userve was better (Roon played first followed by US). The second time I did the blind test on son and wife and they marginally preferred the second play, which was Roon 
The test involved no DSP or up sampling
And remember you need to do for each sample rate… and Roon will then load the correct tap profile into memory for the right sample rate that you have created in REW… however it will attempt to use what you give it… but if not the correct sample rate you will be impacting SQ…
Also it’s worth editing the measured response to smooth out excessive dips or peaks… you don’t want these in a filter response if you are to avoid side effects…
I don’t need to do anything HAF created my filters they work for every sample rate.
If making them in REW then yes after corrections have been applied you export each as a.wav file for each sample rate, zip them up and use in convolution it will choose the correct one when it’s needed.
Correct… just ensure you put the filter tap files for each sample rate into the same directory… or from memory you can create a zip file containing the tap kernels for each sample rate . (Check Roon documentation)
You will see the higher sample rates have more tap values and so larger files… (Room correction uses FIR filters)… if nothing else going through this process one might start to get an appreciation of what actually happens with DSP FIR implementations…
BTW any multiplication of a bit stream by a matrix involves convolution… so digitally filtering a bit stream always uses convolution… so typically we just say digitally filtering… as opposed to the mathematical process undertaken on the discrete number series. So in this case the appropriate correction filter kernel is applied for each sample rate.
I am using the Audeze LCD-4 setting. Can’t say I hear a difference.
This topic was automatically closed 60 days after the last reply. New replies are no longer allowed.