Room Correction EQ: How it Affects Loudspeaker Performance 

kevin weithers thanks. Hopefully I’ll be around a while longer.

Home Theater Gurus has an excellent video tutorial about multisubs. ru-vid.com/video/%D0%B2%D0%B8%D0%B4%D0%B5%D0%BE-_A6gPCczhuU.html

If you watch both videos you will see that I specifically note that the average consumer cannot measure the listening window of a speaker accurately on their own. The purpose of this video was to highlight a problem with auto room correction that most would not realize are true. I also stated in this second video that my suggestion is to buy good speakers with pretty Spin-o-Rama charts, optimize the setup, and the only use auto eq as icing on the cake. As for specific eq suggestions, I noted that I personally like DIRAC when done correctly. But to make sure the target curve doesn’t attempt to extend the bandwidth of the speaker but instead follows the natural response of the speaker. Specific reviews of eq or how-to’s was something we suggested as a future project. It was not meant to be a part of this. I wanted to dispel the myths that auto-eq fixes cheap speakers or bad setup.

Matthew Poes couldn’t you take the speaker outside and measure them there, like the method you guys do? Sadly right now I can’t afford speakers with pretty Spin-o-rama charts but I’m planning to get something like DIYSG HTM-12 it has great on and off axis measurements and it fits my budget.

azzy mazzy you can but you need to know how to do it correctly. You need a quality microphone and proper rig. I use a piece of 6.5’ aluminum DJ rigging with a baseplate and top turntable. You could also use a ladder. You need to measure it precisely at 5 degree intervals. Then process the data. The HTM-12 has pretty Spin-o-Rama charts. I’ve heard them many times. Matt grant is a friend. I am a fan of those speakers.

Matthew Poes is there an article on how to do it correctly and will cross spectrums lap’s Umik-1 will good enough for it? I’m not going to review speakers just trying to do some EQ. Do you think there is a better speaker than HTM-12 for around the same price +/- $50

What you DO is what you DON'T do.

If I do this project, and I want to, it won’t go that far. I simply couldn’t test in multiple rooms. This has been on my wish list for a while. I previously reached out and secured promises of product for the experiments but didn’t have time. Still don’t. Hoping to recruit some help and move it along this summer.

@@PoesAcoustics makes sense, it's a huge undertaking. Hopefully the more focused scope allows you to do it well. Good luck!

Do you use Audyssey Reference or Flat, and Dynamic EQ on or off?

I have xt32 and Kef speakers too. What receiver/processor are you using, and how do you limit the curve to below 500Hz?

@@SorikuXIII reference, dinamic eq on

@@rosstee I have Marantz 7013. You can only do this with the newer receivers and you must download/pay the app on your smartphone

@@audioreviewchannel3365 Ok thanks, I have the Denon 4500 which is still less than 2 years old so perhaps it can do the same.

I too go from on and off, I like what it does to the sub, but not the rest of the speakers. I find that just leaving it off and maybe a tweak to the manual eq, gives happy results.

Absolutely: much more technology can be brought to bear on this "problem". I put that in quotes because, like a lot of the latest and greatest features on new products it's hard to discern whether or not the feature addresses an actual problem or provides an incentive for the consumer to upgrade. #TeamGene ;-)

Mike Weinstein I’m more of an acoustic treatment fan myself for addressing this “problem” which I do feel is a problem. But likewise the treatment is sometimes worse than the disease. Acoustic treatments often overpromise and underdeliver, (seems to be a common thread in the audio industry) allot of it is also the consumer being overly optimistic making it easy to be misled by spokesmen selling false hope, but I would love even more an advancement in acoustic technologies than in corrective software. It may be time for the the construction industry to step up to the plate, so that rooms are easier to work with from the outset.

Best wishes to Spain. I hope the Coronavirus fades quickly.

Great to see the Trinnov system, ARC, Room Perfect and ASPEQT evaluated

I think it’s useful or not equally in any sized room. Large rooms with longer RT’s can’t really be corrected with EQ. Reflections can’t be corrected using current SIMO type EQ (DIRAC and the like). Only advanced MIMO eq can actively change the rooms acoustics and that isn’t a viable market product right now.

Paul Roscelli roomperfec isn’t really any more advanced or more expensive than DIRAC or Audyssey. I’ve used it. I like it. I like DIRAC better.

DEQ unbalances surround and main channels, so off it goes.

@@ericshutter5305 THIS. I liked playing around with DEQ for my mains, but when playing games or watching movies, my surrounds were too distracting. So i just leave DEQ off now and got used to my setup in my -30 to -15 listening range. I played around with a room curve on minidsp for subs but even that i just toned down to practically flat so that when listening to -15level it's fine. I went from tinkering constantly, to just finding a happy medium (or compromise if you will) and enjoying the system.

@@martheunen yeh I have a love hate relationship with DEQ, but I agree that just turning it off and forgetting about it, you will be happy. Turning it on gives you the initial wow this is better, but later trying to watch something, movies, sports etc you find yourself tinkering with it, cos it seems unnatural or makes center channel harder to hear, although it does enhance faint background noises in the scene. But ultimately off is more enjoyable, day to day, even the wife says I can hear what they are saying more clearly with it off. What I’m saying is you can enjoy all the time with off, you can also enjoy with on, but not all the time, and not without tinkering with the offset level for everything u watch.

I would go lower like 250hz

119鬼斧神工 that isn’t really what his research said. They never explicitly tested that. In fact, maybe if the JBL Synergy speakers have gorgeous spin-o-Rama data and Harman uses extensive full bandwidth eq to further improve the response. But the sentiment of what you are saying is true and is what I said in the first video. A speaker with good measurements won’t need eq much. Mostly in the bass and transition zone. The speaker design should come first.

@@PoesAcoustics Room EQ is different with speaker EQ, if a speaker measure not good at chamber, it need EQ to fix that just like some active speakers. But if you measured a speaker in a room with simple RTA signal without 2034A data, you will do something wrong if you tune the room response curve to some target curve. Because you don't know how much direct sound or reflect sound contain in the in room measurement. FFT gate will filter some reflect sound, but also reduce the measure precision...

Yamaha is one of few AVRS with manual PEQ.

Spider Man yes!

We actually show how we measure listening window response in our measurement protocol and in our loudspeaker reviews. See: www.audioholics.com/loudspeaker-design/understanding-loudspeaker-measurements

S H this is a very involved question to fully answer. A lot of people have asked about the measurements and so I think a companion video dedicated to how we measure and calculate the responses might be in order. First, as for a cursory overview of the science as to why a listening window is a better fit than a spatial average to what we hear, you need to no father than Toole. As usual he has written about this topic more than anyone in the modern era. Nobody has really synthesized the research as well as he has. www.wghwoodworking.com/audio/loudspeakers_and_rooms_for_sound_reproduction.pdf This paper is freely available and covers the topic. Olive has worked more on the modeling of the relationship between which measurements best correlate with our perception of sound quality. As to why the listening window is so critical, why the spa Yao average doesn’t do a better job, you have to go to psychoacoustics. Toole addresses it in that paper. More in his book. I cover it in my article as well. As for how I derived the listening window. The process is relatively simple but not something I expect people to go out and do. Take a speaker, amplifier, measurement microphone, laptop, REW, a turntable (lazy Susan) with 5 degree increments, a speaker stand that is about 6+ feet tall, and a microphone stand that goes up to 8-9 feet tall, with a book arm at least 3’ but preferably 6’ long. Mount the speaker in the turntable on top of the stand. Hook it up to the amplifier and connect that to the laptop. Connect the mic and place the mic in the stand. Put the mic at the acoustic center of the speaker directly on axis, 2 meters away. Using REW, run a sweep. Turn the speaker 5 degrees. Mic stays put. Run another sweep. Repeat until you have measured the speaker across a full 180 degrees to either side (360 degrees in total, yes even behind the speaker). Repeat this in the vertical direction, though the window can be smaller, say 60 degrees. Everything must be labeled carefully. Go into REW and window the measurements to eliminate the first reflections, which will be the ground. This will limit the resolution of the response below 200hz or so at these heights. Export all the processed data as text files. Combine the vertical and horizontal data such that each takes up a different row. Using this free software kimmosaunisto.net/Software/Software.html Load all the data for the speaker. Export the data for the listening window, early reflections window, sound power, etc. re-import to REW or a separate graphic software to create your graphics. If you want to calculate IIR filters to correct the listening axis specifically. Import into REW. Splice a listening window measurement with a spatially averaged room measurement at 200hz-300hz or so. Setup a proper room curve, but minimal HF tilt beyond what is naturally in the speaker as the room will do the rest in its own. Let REW calculate the EQ for you. To test the validity of the model fit for the correction, apply the EQ and measure again. Measure at multiple points throughout the listening area. Export as text into Excel. Fit a target curve to the response based on a proper in room target curve. It should generally fit through the speakers natural in room response. Calculate the root mean square error based on the difference between the measured response and the target curve. Average the RMSE for all measurement points. The lower the number, the better the fit.

S H now as for testing the corrections. That is a ton of work but man i want to do it. A lot of people have tried before but they have mostly done it wrong. Here is my plan: -Measure a speaker as above (I really want to use more than one speaker type for this) -measure and apply correction to speakers using each of the major room correction systems -measure preout of processor you obtain correction transfer function -apply correction transfer function to test material -load into MUSHRA software -run multiple people through the tests to both obtain most preferred correction and characterize subjective perception of correction -apply correction transfer function to speaker measurements as I did in this video -assess the corrections using the same RMSE method mentioned before. -write editorial and do video on results Anything short of this won’t tell us what we need to know. Blind testing is useless. Who cares if people can detect a reliable difference. We need to know what difference, why, and which is preferred. MUSHRA is what Olive used and makes the most sense to me as well. Getting all the equipment is no problem. It’s just about finding the time.

@@Audioholics I was checking out your channel to see how you doing regarding your health condition, please keep us updated, wish you a speedy recovery

@@PoesAcoustics Very much appreciate the comprehensive reply Matt. I am loosely familiar with anechoic spinorama data and the listening window plot that is derived from your described off axis/reflection measurements, my confusion is how you are generating it for and in your specific room (not quasi-anechoically outside like I believe you guys usually do). My (perhaps wrong?) understanding is that it would be effected/changed by positioning, size, shape, layout, materials, treatments, etc... of someone's room. Are you saying you performed these spin measurements in your room? Sorry for the dumb questions, it's just very counter intuitive how a spacial average can look great, but a listening window (which is itself a spacial average as I understand it) can look terrible.

Happy wife, happy life 😃

use room simulator as a guideline but it can't account for the acoustics of your room which are not 100% rigid walls. Nothing beats actually measuring with REW and adjusting accordingly.

@@Audioholics Is it better to use the test tones from the receiver or should a test disc bluray / DVD be used to balance levels of the speakers. Would you get a more accurate response since it is coming from the player that would be playing the movie?

In couple of days called Corona Recovery or Coming Out of Isolation. Waiting on final test results.

Johnny Blaze not really comparable. Audyssey and DIRAC are automatic FIR based correction systems. REW is just a measurement software. It isn’t an EQ. However it does include a feature to calculate IIR filters which can be used with a manual PEQ system. I use REW for much of my measurements. At this point it’s a superior measurement system to most pay softwares as long as all you need is single channel measurements. For room correction type work that is mostly fine. It isn’t comparable to DIRAC in sophistication but it’s capable of very good correction in the bass. I hear no difference. On the other hand, even John strongly recommends against the use of REW to do full bandwidth correction. Using PEQ based on in room spatially averaged measurements is a very bad idea. It will typically lead to worse performance above Fs than had you done nothing at all. REW is best for bass correction and a little fixing up to maybe 500hz. I would be very careful of anything else.

@@PoesAcoustics Thank you for clearing that up Matthew. I'm looking for a single system I could use that can do the much better room correction for a home theater. I have a Denon with the Audyssey xt32 but would you recommend the more expensive Dirac over that? I find with the Audyssey, I tend to do a lot of tweaking to get it to my liking. Would you say Dirac is a significant upgrade from the Audyssey?

Johnny Blaze yeah I like DIRAC better. It’s not a fix for bad speakers, but used correctly, I find it does more good than harm. I’ve had very good results with DIRAC. I must warn that DIRAC is much harder to setup. Audyssey did a really good job of making setup Easy. As a beta tester of Dirac, I had thought the problem was my use of beta products. However I’ve had similar experiences with final production products too. Many folks I know have complained as well. I mean, I ended up setting these up for various audio shops because They couldn’t get it working.

@@PoesAcoustics Thanks a lot. I'll stick with Audyssey and do the final tweaking. If down the future I end up buying a product that comes with Dirac, I'll figure it out then.

Dirac on miniDSP sits on top of all other settings (xover, peq, delays). You can theoretically set only the xover point and then let Dirac do all the work but in practice it is much better to manually eq first and time align the subs and then let Dirac make the final small corrections towards the target you choose.

Under non-pandemic times I can offer that as a service.

pb24dagrk no this was XT Pro. I dont have anything with XT32 at the moment. It won’t be any different really. It has finer resolution, but the problem I showed is inherent in the general approach. Any eq system will do this. Some will be worse than others. I was having trouble generating the charts with DIRAC but I was able to carry out the correction. It did the same thing. Looked flatter but still had the same errors. I just kept getting an error trying to capture the transfer function. That was needed to create the graphics.

@@PoesAcoustics Thanks for taking time to respond. I use the app and limit EQ to 500hz because, well, I feel inclined after reading some articles and watching a couple of your videos. I'm not skilled enough to measure with REW and not trained enough to tell a difference (I have XT32 and take time to at least set it up correctly and methodically). I keep hearing and reading how much better XT32 is compared to previous versions, and how little it does up top and how much more it does with bass...hence why I'm interested in your potential future comparison video but hope it would be with the "better" version of Audyssey.

L J yes disable YPAO