Hello everyone,
I have been tackling some issues in the SPL curve of my room using a 2-step room correction (after proper acoustic treatment).
Would anyone be knowledgeable enough to explain me what's happening:
My strategy has been: using MSO to handle the lower part of the spectrum and then using REW EQ filters for the rest.
For the low end, I optimize with MSO until 300 Hz and I implement the result biquads in my miniDSP 2x4HD. It works great.
For the rest of the spectrum, I calculate the EQ in REW. I then put the filter impulse response into a convolution device to flatten the curve above 200 - 300 Hz. It works fine.
The audio goes first thru the convolution in my DAW then out to the miniDSP then to the speakers and sub.
Individually each technique does the job but when I do the whole thing I get a nasty dip from 200 Hz to 2 kHz...
This happens only on one speaker (the room being not perfectly symmetrical).
Would that be some sort of phase issue?
I can provide more info and graphs (but when I did on another forum no one dared to answer 🤔).
Many thanks.
Cheers,
I have been tackling some issues in the SPL curve of my room using a 2-step room correction (after proper acoustic treatment).
Would anyone be knowledgeable enough to explain me what's happening:
My strategy has been: using MSO to handle the lower part of the spectrum and then using REW EQ filters for the rest.
For the low end, I optimize with MSO until 300 Hz and I implement the result biquads in my miniDSP 2x4HD. It works great.
For the rest of the spectrum, I calculate the EQ in REW. I then put the filter impulse response into a convolution device to flatten the curve above 200 - 300 Hz. It works fine.
The audio goes first thru the convolution in my DAW then out to the miniDSP then to the speakers and sub.
Individually each technique does the job but when I do the whole thing I get a nasty dip from 200 Hz to 2 kHz...
This happens only on one speaker (the room being not perfectly symmetrical).
Would that be some sort of phase issue?
I can provide more info and graphs (but when I did on another forum no one dared to answer 🤔).
Many thanks.
Cheers,
A first glance I would wonder why you are breaking up the task between MSO, Convolution and the miniDSP? Have you tried just using one method to verify?
You mention a phase issue - could it also be a timing issue between the corrections that leads to the dip?
You mention a phase issue - could it also be a timing issue between the corrections that leads to the dip?
If you want to post your graphs it would be nice to see with and without subs, with and without DSP.
Hi Pano,
Thanks for your replies.
This is because MSO has only a limited range of optimization (max 300 Hz) and I want to optimize the full spectrum up to 20 kHz. The convolution is less precise than the MSO correction because it is based on the optimization on one spectrum (be it an average or not) whereas MSO calculates the best parameter sets fitting several listening positions.
Below you can see how each work nicely, on their own:
And below you can see the two corrections at the same time (red curve). There is something wrong from 200 Hz and it recovers slowly around 2 kHz... *
There is a peak disappearing at 550 Hz for some reason but this thin peak was nothing related to the large dip appearing with the 2 corrections at the same time.
The bulges and bumps in the lower end are due to the Xover in my DAW: (but I don't mind so much if the SPL is nice enough):
Thanks.
* there is a few dB offset between the correction of the low-end and that of the high end. I thought fixing afterwards, easily with some negative gain on the high end.
Thanks for your replies.
This is because MSO has only a limited range of optimization (max 300 Hz) and I want to optimize the full spectrum up to 20 kHz. The convolution is less precise than the MSO correction because it is based on the optimization on one spectrum (be it an average or not) whereas MSO calculates the best parameter sets fitting several listening positions.
Yes, on each part of the spectrum: MSO (calculated biquads implemented in miniDSP) for the low end, and the convolution for the higher part.
Below you can see how each work nicely, on their own:
Well, actually this is a little bit the same as far as I understand. But I did try to alter the delay between the low-end signal and the higher-end signal without any significant improvement. (for instance I tried to use the delay equivalent to 180 deg at 400 Hz, but I also tried plenty other random delays too, under 1 ms and up to 10 ms.)
I am going to search for the spectra without the sub. But here is what I thought was maybe worth looking at:: the (excess) group delay, sub on and sub off.
There is a peak disappearing at 550 Hz for some reason but this thin peak was nothing related to the large dip appearing with the 2 corrections at the same time.
The bulges and bumps in the lower end are due to the Xover in my DAW: (but I don't mind so much if the SPL is nice enough):
Thanks.
* there is a few dB offset between the correction of the low-end and that of the high end. I thought fixing afterwards, easily with some negative gain on the high end.