REW looking at hi-fi listening room

Hello everyone,

Been playing around with RoomEQ wizard for a couple of days. What a great program and free too! Well done to the creator!

I'm going to use the program to look at my listening room at home, I have a very good stereo hi-fi, but notice most people here are using it to look at 5.1 systems especially to get the correct sub (or .1) response.

If I'm looking at my full range speakers I guess I can use REW to see and eliminate early reflections and room mode problems? As there is no EQ whatsoever in my system (it's a Cyrus bi-amped system) I'll need to move speakers and try to apply absorption and diffraction materials to get the best response.

So am I aiming to get the flattest response possible using the software? From 20 - 20,000Hz. If I see peaks at, for example 80Hz, that's the frequency I should tune my bass traps too etc?

Thanks for the help!

Chris

[JohnM]

For full range work the Energy-Time curve is useful for showing reflections, which appear as spikes in the response after the initial peak, and the RT60 curves help to show whether the balance of damping in the room is appropriate. The frequency response with 1/3 octave smoothing applied gives an indication of overall balance but for a true representation you would need a calibrated microphone and take care about the orientation of the mic relative to the speaker being measured.

For low frequency measurements the waterfall and spectral decay plots help show modal resonances and the unsmoothed frequency response shows how even the low end is, peaks in the response are generally due to room resonances. The RTA plot can be very useful when moving things around to see how positioning (of speakers or listener) affect the result.

Thank you for your reply John. I'm using a calibrated mic and have measured the below response. I'd really appreciate your thoughts. I'm using a small room (11 x 10ft) and seem to have a serious mode at 60hz. The graph is smoothed to 1/3 octave as you suggested above.

Thanks Again!

[brucek]

seem to have a serious mode at 60hz.

Yeah, without a doubt.

Equalization can do wonders to room modes such as this when you're using a combination of bass managed mains and subwoofer, but with a set of full range mains only, you will need to get the best response possible using speaker placement. Try a few different positions and see if the peak at 60Hz will clear.

brucek

[JohnM]

That's quite a peak. You might have some luck if you could get a Helmholtz resonator tuned accurately enough, but generally speaking the kind of acoustic treatment you would need to get rid of that peak wouldn't leave any space in there for you. Try the mic in whatever potential listening positions you have to see what scope there is for finding a better spot (it can be handy to do that while playing the Periodic Pink Noise test signal and using the RTA to see the response). A useful alternative in your situation might be a Bag End bass trap (http://www.bagend.com/) or adding a small sub and adjusting its phase to counter the main speakers at the peak (which is more or less what the bag end does but not nearly as precisely, you would want the sub on the small side so it doesn't run too low and could use its crossover to limit its top end).

[bjs]

That *is* an impressive peak! An anechoic measurement of the woofer (microphone about 1/4" to 1/2" from dust cap) would be useful to ensure it is all due to the room.

Thanks for the comments. Having had a play about with speaker placement, I've found a position that sounds much better and smoother (even thought the graphs don't seem to change much - I'm trusting my ears here!). I'm looking into the bagend solution too, seems like a good idea, although I have a feeling it may sound a bit phasey or comb filtery?

I took a measurement very close to the dust cap of the speaker as recommended by bjs (good idea!) and it's brought up some interesting results. The anechoic measurement isn't as flat as I was hoping for. The graph below shows the measurement from 20 - 200hz of the response very close to the speaker (green line) and from the new listening position (purple line). There is a certain correlation at around 60Hz, that may not be entirely due to the room.



I have a feeling though that because my room is fairly cube shaped (11' x 10' x 10') I'm getting the first (lowest?) mode at around 60hz tripling up. Not good! But I guess inevitable for this room shape and size. Perhaps I should ask my boss for a raise so that I can buy a bigger house with bigger rooms!!

My speakers came with some foam plugs that could be put in the ports at the back. I tried remeasuring with these in. The close measurement to the speaker (blue line) is MUCH smoother, but the room response is similar to the inital measurement (brown line)



Here is the comparision of the room responses with the damping out (purple) and damping in (brown).



We 're definitely getting somewhere though! Thanks again for all the help! I'll try playing around with the sub cancelling idea later as I have a small sub up in the loft somewhere.

[Lobotomy^]

Weird nearfield response of speaker with unplugged port is normal to speaker tuned aproximately 48Hz. At tuning frequency cone doesn't move much and most of the sound is coming from port. If you make the same measurement from port opening, you'll see it will have it's peak at 48Hz. If you make the measurement outside one or two meters from speaker, it should have it's response flat to 48Hz and drop rapidly below that frequency.

Your room has really bad shape. You should really try to place some resonators tuned to 55-60hz in there. Of course you can eq your response to flat but it doesn't effect decay time.

[brucek]

Of course you can eq your response to flat but it doesn't effect decay time.

It does at the listening position that you applied the filter to correct. The EQ filter has a time response that matches closely to a modal time response. Both the amplitude and the ringing are counteracted by the filter.

brucek

[Lobotomy^]

brucek wrote:

It does at the listening position that you applied the filter to correct. The EQ filter has a time response that matches closely to a modal time response. Both the amplitude and the ringing are counteracted by the filter.

brucek

If you have 1sec decay time of -60dB without eq, you will also have 1sec dacay time of -60dB after eq. If you eq 85dB peak -15db@60Hz to 70dB, you have exactly the same decay time to silence than orginal signal played at 70dB level. Same ringing is still there, it just won't overhelm other frequencys same way the unequed signal did. On the other hand you will lose part of the kick you had at 60Hz with unequed signal as it is played lower volume.

I don't know any commersial product that would measure room acoustics and alter orginal signal to dump modal ringing. It would need massive amount real time calculations. For example to be able to play real one kick at modal frequency would need to play first that kick and then damp modal ringing by playing opposite phase echo of that kick to counteract modal ringing. Though acoustics are too complicated to be corrected with such simple method. Even if that would cancel modal ringing, I doubt that it would sound good. Your brain have got used to acoustics of room and if the sound doesn't behave like it is played in room, it just sounds weird.

But that was a bit off topic all ready. Do what you can with placement of listening post and speakers. After that improve acoustics and check again placement of listening post and speakers. After that with eq only what is needed.

[brucek]

If you have 1sec decay time of -60dB without eq, you will also have 1sec dacay time of -60dB after eq.

I think you're not considering that a filter has both a frequency and a time response. The filter has a time response that acts like a 2nd order biquad, just like the modal resonances of a room. The filter will reduce the amplitude caused by the mode's gain and the excess decay time. The center frequency of the filter and bandwidth have to closely match the mode, but REW is quite capable of this. It is listening position dependant, and only valid around the point where the response was measured, but we expect that as a side effect of equalizing. I can easily show you graphed waterfall examples.

brucek

Hi guys,

I had a play around with some bass traps, but struggled to hear the difference in my room, so I finally decided to buy a Beringer 2496. WOW. What a bit of kit!

I've now EQ'd out the HUGE mode at 58Hz and have made some other adjustments by ear up the frequncy range. It sounds great. To start with I was a little worried about putting a £200 EQ unit on my 7 grand hi-fi, but it's really been worth it. The width function is nice to play about with too.

So the bass response is still not perfect, according to the REW program, but I learned a long time ago to trust my ears and not rely completely on an EQ graph!

Thanks again for all the help,

Chris

[Wayne A. Pflughaupt]

So the bass response is still not perfect, according to the REW program, but I learned a long time ago to trust my ears and not rely completely on an EQ graph!

Good for you! A picture-perfect graph is not the goal. An improvement in sound quality is! Besides, shooting for a perfect graph, even if it’s achievable, typically requires over-filtering, which usually is not a good thing. Besides, a mic doesn’t “hear” the same way our ears to, so you can’t rely 100% on a graph...

Regards,
Wayne