Room EQ typically address the interaction between the loudspeaker and the room from the listening position(s). It can also address defects from the loudspeaker itself in the process.
The cleaner the setup (ie symmetrical placement, acoustically treated room, bigger room, etc.) the less room EQ you should need.
Speaker EQ is more intended at addressing the defects of the loudspeaker itself, regardless of the room it will be installed in. This is for example what JBL/Harman does with their spinorma approach.
This type of EQ can be done in room, but special care has to be taken to isolate what comes from the room (at the measurement position) and what comes from the speaker.
You can probably do both in the same time, but I think it is always better to start with speaker EQ and add room EQ if/where needed (ideally only for LF if the setup is OK and the speaker has a well behaved off axis response)
1sec of FIR (eg 44.1k taps at 44.1kHz and 192k taps at 192kHz) is generally more than enough for any kind of real world correction.
(By the way what convolution software are you using? You need to resample everything to the sampling frequency of the FIR, or have specific FIRs for each possible sampling frequency)
If CPU is not the limiting factor then you have to consider what delay you can accept.
1sec FIR with middle centering will impose 0.5s of delay.
Of course if you mostly use minimum-phase correction (which is often the case), then you can reduce that delay by lowering the centering (eg "1%" when doing exclusively minimum-phase correction, which result in 1ms of delay for a 1sec FIR).
Also, in you situation you can avoid using the "optimization" feature: it is very useful when you are tap-limited, but can generate pre ringing.
I use EqAPO. It requires 192khz sampling rate impluse file (I think because the Asus PCI soundcard is set to 192).
I don't mind the delay, I can adjust in the video player the sound also (Kodi).
Middle centering setting also has impact on pre ringing?
I will use 'no optimization', thanks for the tip.
I don't mind the delay, I can adjust in the video player the sound also (Kodi).
Middle centering setting also has impact on pre ringing?
I will use 'no optimization', thanks for the tip.
Nope, no impact on preringing per se, but less truncation, which is a good thing.
Please note that the 1sec FIR suggestion is on the high side, to be able to cope with all situations: you probably don't need that much most of the time in practice.
My advice would be to use just enough taps for the task, ie the number of taps for which you get satisfying result curves (ie matching the target ones close enough) without having to resort to optimization.
Also you should use the same number of taps for all your channels (ie the numbers of taps required for the lower channels, because LF typically requires more taps) to avoid delay problems (FFT engines can be tricky, and simple delay compensation with different tap counts might not give predictable results)
Oh, and a small typo caught my eye in my last post (too late to edit)
Please note that the 1sec FIR suggestion is on the high side, to be able to cope with all situations: you probably don't need that much most of the time in practice.
My advice would be to use just enough taps for the task, ie the number of taps for which you get satisfying result curves (ie matching the target ones close enough) without having to resort to optimization.
Also you should use the same number of taps for all your channels (ie the numbers of taps required for the lower channels, because LF typically requires more taps) to avoid delay problems (FFT engines can be tricky, and simple delay compensation with different tap counts might not give predictable results)
Oh, and a small typo caught my eye in my last post (too late to edit)
Last edited:
Aha, thanks! (I need to learn more before I understand!)
I am thinking for next release,
is it possible that each time we generate an impulse file *.wav, an *.rephase with the same settings is saved?
This can be useful, I forget to save last night my Paragraphic phase EQ and the rest, and if i just want to change the xo point I need to redo all other settings.
I am thinking for next release,
is it possible that each time we generate an impulse file *.wav, an *.rephase with the same settings is saved? This can be useful, I forget to save last night my Paragraphic phase EQ and the rest, and if i just want to change the xo point I need to redo all other settings.
I speculate if it is usefull in cases where you have reflections back to the driver inside a box or standing waves inside the box. Or maybe diffraction on the front plate.
DRC is used with success outside the box (in the room). Just speculating if it can be as good with refelections and standing waves inside the box.
Further flattening of phase and amplitude and xover filtering can then be done by rePhase FIRs summed with the DRC FIR.
DRC is used with success outside the box (in the room). Just speculating if it can be as good with refelections and standing waves inside the box.
Further flattening of phase and amplitude and xover filtering can then be done by rePhase FIRs summed with the DRC FIR.
Thinking of Denis Sbragion's DRC:
DRC: Digital Room Correction
Has been around for a while
And Denis says it works best at close reflections.
The "stolen idea" is to correct speaker elements well into the xover region to try to make the two elements be acoustically in phase in the overlapping region. And thus be in accordance with linkwitz-reilly requirement of 0 phase difference between drivers through xover region (3 octaves). Puh. Maybe over complicating things
DRC: Digital Room Correction
Has been around for a while
And Denis says it works best at close reflections.
The "stolen idea" is to correct speaker elements well into the xover region to try to make the two elements be acoustically in phase in the overlapping region. And thus be in accordance with linkwitz-reilly requirement of 0 phase difference between drivers through xover region (3 octaves). Puh. Maybe over complicating things
I see, interesting, I presumed you meant outside the box, in the room, for obvious reasons
Under what circumstances in the room has/is it used successfully? I understand it is still in its infancy and wonder how effective it is and how it differs from other attempts at digital room EQ
I would not consider DRC to be in its infancy. However to do correction out in the room does bring up a few things to keep in mind.
Among many others I successfully use DRC. With it's many options to figure out it isn't an automated process. I have spend lots of time to create my own templates for my specific needs.
I use it more like speaker correction than room correction as the biggest culprits have been solved with old fashioned acoustic measures (damping panels).
There hasn't been any new DRC-FIR development for quite a while but it's a fully operational package that can create very satisfying results.
As with anything, the results depend largely on how you use this versatile tool. Just as with RePhase.
DRC can be whatever you need it to be. You can control each and every variable.
The way you use DRC will determine it's outcome.
For instance, the summing of multiple measurements within REW can be used as a base for DRC too, just like it can be used in RePhase.
No idea what you are getting at with that second sentence.
The way one uses these tools determine the outcome. Some people actually like to sit with a "head in vice" like stereo reproduction. Others prefer a more wider sweet spot.
DRC can be used to manipulate phase, but it's not a strict requirement. It's up to the user to determine what it is you want from it. It can do minimum phase as well as linear phase correction.
What you put into it will determine the outcome, just like with RePhase. Don't try to correct the phase of a 500 ms measurement window, measured in a room.
The way you use DRC will determine it's outcome.
For instance, the summing of multiple measurements within REW can be used as a base for DRC too, just like it can be used in RePhase.
No idea what you are getting at with that second sentence.
The way one uses these tools determine the outcome. Some people actually like to sit with a "head in vice" like stereo reproduction. Others prefer a more wider sweet spot.
DRC can be used to manipulate phase, but it's not a strict requirement. It's up to the user to determine what it is you want from it. It can do minimum phase as well as linear phase correction.
What you put into it will determine the outcome, just like with RePhase. Don't try to correct the phase of a 500 ms measurement window, measured in a room
.
The second sentence was my attempt to condense this from Wikipedia's description of DRC, did I misunderstand it?
"DRC systems are not normally used to create a perfect inversion of the room's response because a perfect correction would only be valid at the location where it was measured: a few millimeters away the arrival times from various reflections will differ and the inversion will be imperfect. The imperfectly corrected signal may end up sounding worse than the uncorrected signal because the acausal filters used in digital room correction may cause pre-echo. Room correction filter calculation systems instead favor a robust approach, and employ sophisticated processing to attempt to produce an inverse filter which will work over a usably large volume, and which avoid producing bad-sounding artifacts outside of that volume, at the expense of peak accuracy at the measurement location."
"DRC systems are not normally used to create a perfect inversion of the room's response because a perfect correction would only be valid at the location where it was measured: a few millimeters away the arrival times from various reflections will differ and the inversion will be imperfect. The imperfectly corrected signal may end up sounding worse than the uncorrected signal because the acausal filters used in digital room correction may cause pre-echo. Room correction filter calculation systems instead favor a robust approach, and employ sophisticated processing to attempt to produce an inverse filter which will work over a usably large volume, and which avoid producing bad-sounding artifacts outside of that volume, at the expense of peak accuracy at the measurement location."
It may sound odd, coming from me, I don't consider Digital Room Correction (by itself) a smart solution. I've often said: choose the speaker design that can work together with the room you're in. These tools can be remarkably helpful to optimise those results. Personally I use DRC-FIR, some RePhase (which also is a very handy simulation tool for me) REW and JRiver to get what I want. Knowing what you want is part of that deal.
Most of my time was spend learning to be able to relate the measurements to what I hear and what influence the room had within those measurements and my perception.
Most of my time was spend learning to be able to relate the measurements to what I hear and what influence the room had within those measurements and my perception.
That's not a wrong statement. However, as a user you can actually control all variables that determine what happens within DRC-FIR. I purposely deviate from any standard template to get my desired results. For me it is a manipulation tool I've turned inside out to know what part does what.
That's the amount of freedom it offers you.
There are several steps within DRC, each of them can be turned on or off at will. Learning how it was meant to be used is mandatory. After that you can start to manipulate it, should you want to do that.
Same goes for a tool like RePhase. It can be used in various ways, but user be aware if that way is a good one. I have tried to do it all with RePhase, I could not match the performance I get from DRC-FIR, with some minor manipulation done afterwards.
Similar tools to DRC-FIR are Acourate and Audiolense. However those tools are more streamlined packages that make (some) decisions for the user.
I like the fact that I can control every part of the tool I use.
I'm a nutcase and don't expect just everyone to do as I do. You don't want to know how much time I've spend on these tools. And more important: learn what it is I want from them. I do have ears to please too
.
It isn't a measurement tool. It requires an input IR and will optimise that IR, based on the template and target that is used. Within the template are a lot of choices made for a global use. The actual result will be largely determined by the room + speaker.
Let's move this discussion to this thread... should you want to learn more about DRC.
Different tools meant to do different jobs... some overlap does exist. This thread is about RePhase though.
Half of the job in any case would be: what do you want it to do? Easier said than answered.
Let's move this discussion to this thread... should you want to learn more about DRC.
Different tools meant to do different jobs... some overlap does exist. This thread is about RePhase though.
Half of the job in any case would be: what do you want it to do? Easier said than answered.
No, but it could be used as a measurement tool, I thought that's what you meant about using it to see how what you hear is influenced by the room? Yes I thought RePhase was mainly about the crossover and speaker, but DRC was mentioned here so wondered how phase manipulation could be used for room correction and over what frequency range it could be helpful used in that way
Just to make it clear, for all measurements I have used REW.
Manipulation was done with DRC-FIR. Measurements before and after FIR correction were done with REW.
As I correct line arrays I use it from 20 to 20000 Hz (frequency response), however my phase correction is limited to frequencies between 20-500 Hz.
Even there one should be aware what the room/speaker combination does within the listening space. The average of measurements along my "listening couch" is a spitting image of a measurement at the sweet spot. But that did require planning and panels.
Manipulation was done with DRC-FIR. Measurements before and after FIR correction were done with REW.
As I correct line arrays I use it from 20 to 20000 Hz (frequency response), however my phase correction is limited to frequencies between 20-500 Hz.
Even there one should be aware what the room/speaker combination does within the listening space. The average of measurements along my "listening couch" is a spitting image of a measurement at the sweet spot. But that did require planning and panels.