No FIR can emulate an IIR exactly as FIR filters intrinsically have no poles, only zeroes (or put another way IIR's have infinite impulse responses, and infinity is not finite!).
However the key word is "exactly" - in reality you want a good enough response, you don't need infinite responses (even for a heavy reverb).
You can take the response of an IIR, taper it off to zero smoothly and derive a finite response that's very very similar. However you do get way more coefficients typically, 100's or 1000's where the IIR might only have had 5...
However the key word is "exactly" - in reality you want a good enough response, you don't need infinite responses (even for a heavy reverb).
You can take the response of an IIR, taper it off to zero smoothly and derive a finite response that's very very similar. However you do get way more coefficients typically, 100's or 1000's where the IIR might only have had 5...
Hmm i found a really easy way for linear phase:
one really easy way to implement linear phase would be to actually not create a convolution file with REW but import the filter settings into the graphical eq of easyeffects (linux eq), it allows for FIR/FFT/SPM processing and also you can set "APO" (purely digital as far i unterstood it) for each filter instead of the approximation algorithms to resistors
while you cant correct for your speakers themself this way, with this approach you pretty much dont alter phase at all, which for me is honestly enough
i also use the easyeffects EQ as normal EQ for a kind of house curve... the difference between IIR and SPM processing is astonishing imo ... you really hear how IIR messes timing/rythm up and you can actually hear funky phase stuff going on (even with mild EQ)
i plan to get a proper mic stand soon and repeat the measurements i have done already and correct again the speakers, and i will probably try this easyeffects approach, in theory it should work quite good even without rephase (unless your speakers have funky phase to begin with)
i can also just recommend easyeffects from what i heard... the best EQ with the right settings i tried so far
one really easy way to implement linear phase would be to actually not create a convolution file with REW but import the filter settings into the graphical eq of easyeffects (linux eq), it allows for FIR/FFT/SPM processing and also you can set "APO" (purely digital as far i unterstood it) for each filter instead of the approximation algorithms to resistors
while you cant correct for your speakers themself this way, with this approach you pretty much dont alter phase at all, which for me is honestly enough
i also use the easyeffects EQ as normal EQ for a kind of house curve... the difference between IIR and SPM processing is astonishing imo ... you really hear how IIR messes timing/rythm up and you can actually hear funky phase stuff going on (even with mild EQ)
i plan to get a proper mic stand soon and repeat the measurements i have done already and correct again the speakers, and i will probably try this easyeffects approach, in theory it should work quite good even without rephase (unless your speakers have funky phase to begin with)
i can also just recommend easyeffects from what i heard... the best EQ with the right settings i tried so far
@Ghoostknight :
Can you tell some more about SPM processing? Never heard of it and Google is of no help.....
Can you tell some more about SPM processing? Never heard of it and Google is of no help.....
i had honestly the same problem... but SPM sounds the best imo, FFT is second best (audible improvement was pretty much in line with the same order as below)
you can atleast read a little here: https://lsp-plug.in/?page=manuals§ion=para_equalizer_x32_stereo , this is the plugin easyeffects uses
- IIR - Infinite Impulse Response filters, nonlinear minimal phase. In most cases does not add noticeable latency to output signal.
- FIR - Finite Impulse Response filters with linear phase, finite approximation of equalizer's impulse response. Adds noticeable latency to output signal.
- FFT - Fast Fourier Transform approximation of the frequency chart, linear phase. Adds noticeable latency to output signal.
- SPM - Spectral Processor Mode of equalizer, equalizer transforms the magnitude of signal spectrum instead of applying impulse response to the signal
if someone knows more please let me(us) know
https://www.attackmagazine.com/technique/tutorials/what-is-spectral-processing/
if you google "spectral processing" you actually can find some stuff
if you google "spectral processing" you actually can find some stuff
Hello,
A bit off topic but looking for a vst plug-in to load the
Wav files generated by rephrase.
I use them to crate active filters for my DIY setup (3 way plus sub).
Currently using old school convolverVst but you need to register some dll and my dock won't save the config when you restart. (cantabile host and rme fireface).
For info I am using lr48 and I correct the phase of my horns.
Regards Greg
A bit off topic but looking for a vst plug-in to load the
Wav files generated by rephrase.
I use them to crate active filters for my DIY setup (3 way plus sub).
Currently using old school convolverVst but you need to register some dll and my dock won't save the config when you restart. (cantabile host and rme fireface).
For info I am using lr48 and I correct the phase of my horns.
Regards Greg