My room response is HORRIBLE!!!

[tboooe]

Today, I tried to measure my room response using the Radioshack analog spl and the Rives test cd. Before doing this, I honestly thought that my room was ok, I probably had some bass issues but that was it. Boy was I wrong!!! I plan on doing this test again since this was my first time but if this is any indication of my room response then I think I may need some serious help! I've got huge bass peaks from 20 - 80hz. The readings were off the charts! I also have pretty big peaks from 120 - 315hz. From 400 - 1250 it is relatively flat (within 2db of 0). After that everything is just horrible. I get peaks 1600hz, then again at 4000hz (which is the crossover freq of the 802d), then huge valleys from 6300 - 20000hz. I think since high freqs are very directional I can solve this last issue by using more toe in so that the tweeters are pointed more at my ears.

Out of curiosity, the Rives cd tells me to calibrate my meter to read 0 at 1000hz. Is this a good baseline reference? I also noticed that the meter read different if I was standing off the side or behind it. Where should I stand when reading the meter? I would love to hear the procedure that other members used to measure their room response. How flat should I try to get my room? Is +-2db from 0 ok?

So it looks like I have a lot of work to do! HELP!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

[Briz vegas]

I think the crucial words here are "before doing this I thought my room was ok". If that is what your ears are telling you then I would be inclined to believe them over a cheap SPL meter and an expectation that a flat room response is ideal.

Our local Council spends more money on a single SPL meter than you did on your 802Ds. I would guess that the drop off at high frequency is due to your meter simply not being sensitive at that level. I have an analogue Radioshack SPL but only use it to blend in the sub by measuring near the crossover frequency, and even then I do not trust it further than I can throw it (actually it would probably go quite a way if I threw it maybe that is not a good analogy).

The instructions say you should not stand behind the meter when you measure, for similar reasons that you keep your 802Ds away from the wall. The professional SPLs come with their own stand to minimise interference from bodies holding the thing.

When I first tried to measure my room I was told that if I could get a response curve within 6db I was doing really well

PS Anyone for an SPL meter throwing competition? (not the expensive ones of course)

[tboooe]

Within 6db? I guess I am doing pretty well except for the extreme highs and lows. Not too bad then...

[audioqueso]

tboooe, if you really thought your room was ok, don't start to kill yourself over numbers. You don't want this to become as bad as the car scene where you might have 299hp and you push yourself hard for that one extra hp. For what? You can't actually feel 1hp of a difference. So if your ears were happy, why ruin that for them with a psychological issue? I took my readings of my setup, just so I can see what it looks like and let it be a guide. Yet my perfect setup actually looked bad in a graph.

[dknightd]

The radioshack SPL meter is not great, but, is probably good enough between about 40 and 8000 hz. It falls off above aand below that. It also have a peak at around 4-5k. Three are lots of calibration curves that people use to try and correct the readings, but, these are only approximations since no two meters are the same.

Is you plot a measurement of just one speaker?

I'm not familiar with the Rives test CD - does it use sin waves, or 1/3 octave warbles, or pink noise. I'd be willing to bet if you measured the low end with higher resolution you would find more peaks and dips.

Is your computer close enough to your stereo to allow connecting the meter to it using a long rca interconnect? There is software to help with these types of measurements.

[dyazdani]

I'm not familiar with the Rives test CD - does it use sin waves, or 1/3 octave warbles, or pink noise. I'd be willing to bet if you measured the low end with higher resolution you would find more peaks and dips.

I agree - from the graph, it looks like a fairly rough chart in terms of sampling interval.

If you can do like he mentioned and hook the SPL meter to your PC, you can do some 1Hz plots VERY easily. (You can do them by hand, but it is a real pain)

[tboooe]

audioqueso: thanks for the sanity check. we are all on a never ending search for audio perfection! I just want to make sure I get the most out of my system. I am not asking for an extra 1 hp as you put it. I am jsut trying to get everything that the speaker is capable of.

dknightd: thanks for the input. I did some research and the radioshack meter does indeed drop off dramatically after 10K hz which explains the far right of my graph. the graph is for both speakers. the rives test cd has test tones at the intervals shown on my graph (20, 25, 31.5, 40.....10k, 12.5k, 16k, 20k)

danish: i dont think the radioshack spl meter has an output for a laptop. in order to do what you suggest do I need some sort of card for my pc? can this work with a laptop? what exactly do i need? and how much will this cost? please excuse the ignorance...i am very new to this acoustical engineering thing....

[Paul H]

...
danish: i dont think the radioshack spl meter has an output for a laptop. in order to do what you suggest do I need some sort of card for my pc? can this work with a laptop? what exactly do i need? and how much will this cost? please excuse the ignorance...i am very new to this acoustical engineering thing....

The radio shack meters have an rca output which can feed your line-in or microphone-in on your laptop. You'll typically need a cable with a male rca on one end and a 1/8" male on the other end.

[tboooe]

paul, thanks for the response. i guess the only thing wrong with using the radioshack spl output is that (according to the manual), the response is not flat because of the a/c weighting...maybe i need to get a better meter???

[JKalman]

Today, I tried to measure my room response using the Radioshack analog spl and the Rives test cd. Before doing this, I honestly thought that my room was ok, I probably had some bass issues but that was it. Boy was I wrong!!! I plan on doing this test again since this was my first time but if this is any indication of my room response then I think I may need some serious help! I've got huge bass peaks from 20 - 80hz. The readings were off the charts! I also have pretty big peaks from 120 - 315hz. From 400 - 1250 it is relatively flat (within 2db of 0). After that everything is just horrible. I get peaks 1600hz, then again at 4000hz (which is the crossover freq of the 802d), then huge valleys from 6300 - 20000hz. I think since high freqs are very directional I can solve this last issue by using more toe in so that the tweeters are pointed more at my ears.

Out of curiosity, the Rives cd tells me to calibrate my meter to read 0 at 1000hz. Is this a good baseline reference? I also noticed that the meter read different if I was standing off the side or behind it. Where should I stand when reading the meter? I would love to hear the procedure that other members used to measure their room response. How flat should I try to get my room? Is +-2db from 0 ok?

So it looks like I have a lot of work to do! HELP!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!!

Your sampling interval between frequencies is way too large. You should do 1 Hz or even less, if you can. I wouldn't worry about testing anything over 300 Hz, because if you are going to do room treatment anyway, it will most likely take care of that automatically, if you use the right kind of material for your treatments. Also, when measuring, as you are, with such large intervals, you will be randomly getting the up or down dips caused by comb filtering in the higher frequencies. I'll post one of my measurements without treatments so you can see what it looks like with miniscule frequency intervals in a few minutes (it is on my laptop, not on this computer).

Once you get a more detailed measurement, you can start moving the speakers around the room looking for spots which give better frequency responses. You might be surprised how much of a difference speaker placement makes. Funny enough, the clapping method works pretty well for finding the best spots in a room (WASP method) used in the Wilson setup technique I gave the acronym for in parenthesis. Then follow that with equilateral triangle placement of the seating position.

It is well worth finding the best position for your speakers and your seat in your listening room... That is, if you have that freedom to do so in your listening space, i.e. if your wife, or significant other, won't castrate you for moving the speakers away from specific walls to diminish bass loading, etc.

[JKalman]

I'm having problems accessing my website so I'm going to post links to examples I can find on the net.

Bryston Example of Comb Filtering.

Figure D3 has the higher frequency comb filtering picture, but this page in general is excellent for seeing how you can use the measurement information in general to find a better speaker position.

[tboooe]

does anyone know where i can get a test tones cd with smaller intervals between frequencies?

[JKalman]

audioqueso: thanks for the sanity check. we are all on a never ending search for audio perfection! I just want to make sure I get the most out of my system. I am not asking for an extra 1 hp as you put it. I am jsut trying to get everything that the speaker is capable of.

If you wanted to make that analogy fit the issue at hand, then the 1 HP would be buying some fancy cables, 120 HP would approximate your speakers, 30 HP would approximate your components and the last 149 HP would be the room itself (dimensions/ratios, positioning, treatment, etc). Give or take 10-20 HP between groups. There is a reason experts call room acoustics the single most meaningful change you can make to your audio system...

[JKalman]

does anyone know where i can get a test tones cd with smaller intervals between frequencies?

Here is one. That test CD does 1 Hz intervals between 10 Hz and 300 Hz, but it is like ten per track, or something like that.