AVS Forum banner

1 - 20 of 856 Posts

·
Registered
Joined
·
615 Posts
Discussion Starter · #1 ·
Hi All,


Beta testing has been going well so I've made the latest version of my Room EQ Wizard software available for download from http://homepage.ntlworld.com/john.mu...meq/index.html


Room EQ Wizard is a Java application for measuring and correcting room resonances. It includes tools for generating test signals, measuring SPL and frequency responses and automatically adjusting the settings of parametric equalisers to cancel the effects of room modes. It was initially written to help with the setup of the parametric TMREQ filters on TAG McLaren Audio AV32RDP and AV192R, but the latest version also supports the BFD Pro parametric filters (with correct modelling of the effect of the bandwidth control, which is quite different to the description in the BFD manual
).


Measurement is stepped sine, with local loopback for soundcard response compensation and DFT to isolate the measurement frequency. Log swept sine measurement and impulse response extraction are on the dev list but some way off. The Wizard can also import measurements in the ETF export formats or from text files in a basic comma-delimited format - details are in the help files, which can be browsed online at the site and are included in the program.


The sig gen provides sine waves to 0.1Hz precision, sine sweeps (linear and logarithmic), square waves and various pink noise signals including "full" range (pink spectrum down to just below 10Hz), speaker cal, sub cal and custom filtered to suit your needs.


Note that the app requires V5.0 or later of Sun's Java Runtime Environment (JRE) to be installed, available from http://java.sun.com/j2se/downloads.html


Hope it proves useful.


Best regards,
 

·
Premium Member
Joined
·
24,664 Posts
Very cool! Too bad I just spent too much for an analog parametric, which I was going to use with TrueRTA. At least I haven't bought the latter yet



So what's the learning curve and user friendliness EQ Wizard? Those were the reasons for my initial choices.


How long would it take to do the correction for a sub? Does the swept sine measurement mean it takes awhile?


Thanks
 

·
Registered
Joined
·
615 Posts
Discussion Starter · #4 ·
Hi Noah,


Measurement is stepped rather than swept, and it does indeed take a while as the software waits for the reading to stabilise (SPL display turns green when it is stable) before making a measurement and moving on the the next frequency. How long it takes to settle depends on the room, the frequency and the background noise, but probably averages about 1-2s per measurement so around 2-3 minutes to measure 20-120Hz in 1Hz steps. A bit tedious I know, but fairly reliable



You still need an equaliser, this app just helps work out how to set up the filters, so your money certainly wasn't wasted. It's fairly easy to use, and you get immediate visual feedback of the effect your filter tweaks will have on the response so its a quick way to manually tweak if you don't like the app's efforts at choosing filter settings. Easiest is to play around with it and see how you get on - just looking at how the individual filter gain and Q adjustments affect the overall filter frequency response can be useful.


Happy to answer questions as they arise.
 

·
Premium Member
Joined
·
24,664 Posts
Hi John,


"It's fairly easy to use, and you get immediate visual feedback of the effect your filter tweaks will have on the response "


Immediate meaning the 2-3 minutes (which isn't too bad), or do you mean dialing down a peak once they're located? But then I don't know what the BW needs to be.


Or am I not getting something?


Thanks
 

·
Registered
Joined
·
615 Posts
Discussion Starter · #6 ·
I mean you get to see on the graph display the frequency response of the gain and bandwidth you are setting and how that will affect the measured response after you apply those corrections in your equaliser, you can easily see what bandwidth you need to set by adjusting the bandwidth setting for the relevant filter until the filter's response mirrors the measured response around that peaks's frequency - to make this easier the Wizard has a check box to draw the filter response inverted, so that it overlays the measurement. Have a look at this help page for more: http://homepage.ntlworld.com/john.mu...djustment.html


Regards,
 

·
Registered
Joined
·
91 Posts
did you write the curve fitting piece of the app that originally was on the TAG site or are you layering your work on top of it?


either way is great, but i always thought the underlying curve fitting piece was super slick.
 

·
Registered
Joined
·
615 Posts
Discussion Starter · #9 ·

Quote:
Originally posted by qxlxp
did you write the curve fitting piece of the app that originally was on the TAG site or are you layering your work on top of it?


either way is great, but i always thought the underlying curve fitting piece was super slick.

Yes, I wrote the original app as an interesting way to learn Java. Glad you liked it
 

·
Registered
Joined
·
15,247 Posts
Very nice. Out of curiosity, have you compared the results to other programs?
 

·
Registered
Joined
·
615 Posts
Discussion Starter · #11 ·

Quote:
Originally posted by Chu Gai
Very nice. Out of curiosity, have you compared the results to other programs?

Do you mean the measurement results? Measurements have been compared with data acquired by ETF, you should get a similar result to the t=0ms slice in an ETF low frequency response if you use a gate time of 300 - 600ms (in smaller rooms 600ms may be too long, the ETF measurement starts to look noisy and the level shifts when gate time is set too long). Since the software is measuring values you can (more or less) read off your SPL meter that also provides a handy check - differences are that the wizard is applying a DFT to exclude frequencies other than the current measurement frequency (noise, harmonic distortion) and if you have told it you have a C-weighted SPL meter (the default setting) it applies the inverse of the C weighting curve to get the true SPL.
 

·
Registered
Joined
·
615 Posts
Discussion Starter · #13 ·
Few updates made, details in the revision history on the web site, app now at V3.17. If anyone uses Behringer's DSP1100 software to look at filter responses, be warned: the bandwidths it uses for its response curve are sqrt(2) narrower than they should be and it appears to be modelled on analog prototype responses rather than actual digital biquads so it gets progressively more wrong compared to the actual unit as the frequency increases.
 

·
Registered
Joined
·
15,247 Posts
Well done JohnPM. Have one on me


 

·
Registered
Joined
·
1,055 Posts
I can confirm there is no problem importing files from ETF. Just make sure you have the graph that you want to export from ETF visible in the ETF screen; the whole file is not exported in one shot.


John's progam is outstanding to use in conjunction with ETF. I would almost consider it mandatory. I wouldn't doubt that the program is excellent on its own but I haven't tried it yet.


This link takes you to a wonderful guide on setting up subs and speakers with ETF:

http://www.avforums.com/forums/misc...EQCaseStudy.pdf


John, in your detailed pdf you make no mention of phase. For those that don't know ETF can produce phase graphs for your sub and speaker and phase can be adjusted by using time delay in your processor or on your sub. I had read that it is important for the phases of the sub and speaker to be fairly close at the crossover point.


What are your views on this? What would be the procedure for aligning phases using ETF?


Many thanks,

George
 

·
Registered
Joined
·
615 Posts
Discussion Starter · #18 ·
Hi George,


I haven't used phase responses in setting up the system, if there is a problem with relative phase between sub and mains it will show up in frequency response particuarly in the crossover region between the two, just make some measurements of each main speaker with bass redirection active (assuming the mains are bass limited, otherwise a bit moot) so the sub is contributing and look at the overall measured response. A couple of measurements with the sub phase normal and reversed will give a quick indication which is working best, could then tweak the phase from there. It can be pretty tricky to arrive at a setting that works for all of L, C and R even if your sub is symmetrically placed (which it usually isn't). Have to pick the best compromise.
 

·
Premium Member
Joined
·
24,664 Posts
"John's progam is outstanding to use in conjunction with ETF. I would almost consider it mandatory."


John's program, or ETF?


Thanks
 

·
Registered
Joined
·
1,055 Posts
If you have ETF I would consider it mandatory to use John's program in conjunction with ETF.

All the best,

George
 
1 - 20 of 856 Posts
Top