Generating Simple Frequency Response Correction Filters using Room EQ Wizard

I decided to start making some simpler videos, starting with this one. Here we cover the essential steps needed to create basic magnitude (frequency) response correction filters for convolution applications. Filters made using this method are compatible with just about any convolution reverb processor available that can load an impulse response in WAV format. I will cover implementation of the filters in a later video.

[@richardpena6201]

Thank you!! With your assistance I've created my first FIR files for use with the convolution filter on Roon. Now I can just rest on my laurels or try and tackle phase correction. :D

[@dantwomey]

IMHO your tutorials are the best articulated on this topic. Well done!

[@davidbrancato]

Thanks. I really think that this is easy to do but the knowledge isn't common to many people who would be willing and able to take advantage of it. I'll keep making content in hopes that it reaches more people.

[@dantwomey]

@@davidbrancato Your approach and explanations are very well done. I think anyone who watches your videos will benefit.

[@newmachine00]

Thank you so much! i searched all over internet for a explaination of where was the button "Calculate" Or "Match Response to Target". The best video.

[@davidbrancato]

Cool I am glad this was helpful.

[@akshaykhanna1396]

That’s for a great video. Could you please do a bit on the convolution process and how you use it?

[@davidbrancato]

I have a video here that shows how I am currently using correction filters for two channel stereo with JRiver Media Center. I have not yet tried doing anything beyond 2.1 stereo, so for the time being it's just for any two channel setup. No bass management process yet.

[@yungspoiler]

2 questions, are you doing the psychoacoustic filter instead of lets say 6 or 12 db smoothing for anything specific reason? like is it better to use psychoacoustic over the other choices? Also are you able to put this IR back into REW to then run it through the room sweep measurement to see IRL if the freq curve did infact work as predicted? side question do you, or does anyone else here, know if its possible to put this IR on your main computer outputs so its just on all the time? for example so you can listen to spotify through the IR aswell?

[@vstrec116]

Posso caricare il risultato in qualche eq free? Quale consigli?

[@chillbutdeep]

Great video. I exported the filter and now want to use it for my Mac OSX and Cubase. How do I do that? Thanks buddy

[@mikelastname9583]

For some reason, I can't have the target curve change slope on the graph even though I change the numerical amount in the Target Settings. I also don't have the PK settings appear under the thumbnail graph on the left hand side. It is blank.

[@JamesMaycomposer]

Thanks for this! I'm running into an issue at export: I've followed your steps exactly, go to "Export Filters Impulse Response as WAV" and select the correct measurements (the original ones, where I ran the EQ)-then got "The selected filter set has no filters that have an effect, nothing to export." Any tips?

[@kunglaux]

Quick question - what was the point of creating the EQ sets if you just exported the original measurements to make the WAV file? If you convolve the original measurements, do you also need to run the EQ settings at the same time? Thanks!

[@davidbrancato]

Ah but I didn't export the original measurements to a wav file. Watch it again closely...

[@billwolf8036]

RME TotalMix user - I followed this and get a WAV file of my filter frequencies - I also limited it to 9 bands for TotalMix. But this outputs a WAV file - how can I get txt or just hand enter the correction settings into TotalMix? TIA!

[@digitalstudiosystem9556]

Come inserisco la correzione in Qrange? Si puo esportare il risultato come Txt? Per copiare i valori in un altro equalizzatore? Lo chiedo perché non ho convolver... Grazie

[@robintanj]

this method can use in professional audio installation?

[@ts6640]

It just occurred to me that I guess you are right…….you really should set the same target level (I.e. in your example 70db) for the left/right speakers. I never thought it would matter because the target level in REW is there solely for purposes of equalization (obviously….haha) but I guess it does make sense that the two should be the same?

[@davidbrancato]

I have always made them the same. It never made sense to me to make them different!

[@ts6640]

Curious, have you tried limiting the corrections via ‘Frequency Dependent Windowing’ in REW for purposes of avoiding the correction of reflections?

[@davidbrancato]

I have. As it turns out, none of the methods I came up with that used FDW produced filters that sounded as good as these filters do. I think that simply using the psychoacoustic smoothing model on the measurements evens things out well enough.

[@ts6640]

@@davidbrancato Got it. Many thanks!

[@l.s.1709]

​@@davidbrancatodid you also try other levels of smoothing?

[@davidbrancato]

@@l.s.1709 Oh yes. The smoothing levels I'm using are the ones I've found to work best for me.

[@ts6640]

Probably a noob question but why not merely export the filters as a text file and then import them as PEQs into your equalizer? Is the use of a convolver a necessity?

[@davidbrancato]

The use of a convolver is not necessary, no. However, if you want the best quality you are not going to get it without a hi-res impulse response in a high quality convolver. I used to use Equalizer APO when I got started. It has a built-in equalizer that does import the settings from text files. When I was new to this I thought that was great. I didn't even know that we could use impulse responses to get the job done. However, once I learned how to implement convolution I heard the difference immediately. It was like night and day. There was no comparison. Impulse responses fed into top-tier convolvers deliver pristine sound quality that those text-fed equalizers can't match. Additionally, with convolution you can correct the time domain as well as the frequency domain. That isn't even possible if you are only working with text data for eq bands.

[@ts6640]

@@davidbrancato yea, makes total sense. I am merely a home theater enthusiast trying to calibrate my system using a Denon 8500HA receiver/MULTEQ-X/MiniDSP 2x4HD; nonetheless, I think your view is worth exploring. I will let you know how it goes Thanks

[@l.s.1709]

​@@davidbrancatobut EAPO can also do convolution where you import wav files instead of text files

[@davidbrancato]

@@l.s.1709 It probably can, but I have always found Equalizer APO to be a clumsy and inefficient way of implementing DSP. The platform doesn't allow us to see what's happening in real time, only what's shown in "captures" of your settings.

[@l.s.1709]

@@davidbrancato which (free) tools do you suggest in that case? I'm still pretty new to DSP and installed EAPO on my PC and HTPC.

[@junkyoutube9060]

thank you for taking the time to make these REW videos. Also have you seen Obsessive Compulsive Audiophile and his way of making files with an Inversion method ? kzbin.info/www/bejne/a4rGeWqgZ5J_Z7s

[@davidbrancato]

You're welcome. Yes I have seen what he's doing. I do this a little differently, though. I am still inverting the signal, just not the same way. I prefer to generate correction based on a smoothed response, either around 1/6 or psychoacoustic. My filters all end up as impulse responses, so the fact that they are generated with REW's bi-quad filters isn't relevant.

[@RoboBeaver6]

Last step, "All you need to do is..." What the hell is a convolver

[@davidbrancato]

A convolver is a processor that takes an input signal and translates it through an impulse response. The process is called convolution. Usually this is used as a studio reverb effect to place a sound source into another acoustic space without actually having to go to the physical location. Take a look at Altiverb and you'll get the idea. The convolved IR is blended with the dry signal as a reverb effect. What we are doing with room correction is essentially the same. We take our input signal and process it through an impulse response of the correction needed to fix the issues our speakers exhibit in the room. The only difference between this and the reverb effect is that we do not want any of the dry signal. All we want is the corrected "wet" signal. That's basically what's going on.

[@radicalturkey]

I am trying to use room eq wizard to create a headphone frequency measuring setup. I have got it working fine but I want to calibrate my mic using filters. I found a reference curve for my headphones and imported it and normalised the levels. But now I need to create an eq filter to adjust my measurement curves from my measured curve to the reference. I can't work it out. Is there a variant of your method that would work please? Thanks.