Analyze this! (1 Viewer)

[JonStern]

Today I tested the in-room bass frequency response of my system. I used:

1. Stereophile's Test CD #3, Track 17 (bass warble tones)
2. Sony 222 SACD/CD player
3. Denon 3803 receiver
4. Polk LSi15 front left/right speakers
5. Radio Shack analog SPL meter

I played the test CD through the front left and right speakers only. The speakers are 4' out from the wall and 8' apart. The SPL meter was set to C-weighting/slow response and mounted on a tripod at the listening position, pointed forward and angled up. I set the receiver's volume so that the 200Hz tone from the Stereophile CD was showing 75db on the meter.

Anyway, the readings for the bass tones were as follows (I normalized the 200Hz reading to 0 for reference):

200Hz.....0
160Hz.....+0.5
125Hz.....+1
100Hz.....-7.5 (!)
80Hz......-9.5 (!)
63Hz......-1.5
50Hz......+0.5
40Hz......0
31.5Hz....-8.5 (OK, we're below Polk's lower limit)

So there's a huge dip in the 80-100Hz area. Thing is, I was standing several feet to the side of the meter and both the 80Hz and 100Hz tones sounded pretty much as loud as the other ones. I decided to move the meter away from the listening position and while I was moving it I saw the needle start to climb. When the meter was 4 feet to the right (or left) of the listening position the levels for both 80Hz and 100Hz were in line with the other frequencies.

There's obviously huge cancelation going on at the listening position for certain frequencies. But how representative are stereo "warble" tones of typical source material? Obviously, if the sound were coming from one speaker and not the other there would be no cancelation. Also, this seems to argue for setting the front speakers to "small" and letting a single subwoofer handle the lower bass (or even 2 subwoofers since subwoofers can be moved around pretty freely until the ideal spot is found).

This is probably familiar ground to old timers but I'd appreciate hearing any comments.

Thanks!

PS I don't use the "sub out" channel; I have an old Velodyne which I drive from the Denon's preamp L/R outs. I set its low pass to 40Hz and it blends pretty well with the Polks. I'd hate to have to use the LSi15s as "small" speakers, and my sub doesn't have an unfiltered input so it wouldn't be happy attached to the sub channel.

 

[Guy Usher]

I call that a 50 hertz room bump. Looks like the resonance frequency of your house. I dont think you have a hole at 80, what you have is a big room problem. You dont live in a moble home do you??
Your room is so hot at 50 that it makes it look like a hole around 80. I had the same problem and no matter where I put the sub or how much EQ I used I couldnt fix it, finally I had to go to a sealed box that fired at the floor , no vents to couple with the walls, move it out as far as I could from the wall, managed to pull it down where I was 6db hot and left it there. When I started I was 20db hot at 50, went without a sub for awhile. I tried a bandpass afair that only output at 20-30hz, big bottom but very thin around 80-100.
May not be your problem but it certainly looks like what I had going on. In the end I lived with a very small sub that would only go to 50 hz period, no use trying to get to the bottom , my house wouldnt let me. . .
I reread your post and I see you are not using a sub for those measurements. Looks like you will have to run your Polks small, it will be easier on your amp and will clean up any unwanted stuff down there, a sub is more forgiving when moving around. Put your sub where you sit then walk around with the meter when you find the place where your meter is working best, put your sub there, sounds backwards but works. . .

 

[JonStern]

Guy and Mike:
Thanks for the responses. The Radio Shack meter seems to be working OK--it shows the drop in level at the listening position (which is readily confirmed by the ears), but adjusts once you move a few feet left or right, where there's no dip. I'm aware that below 50Hz the meter readout needs compensation (I may have gotten that from one of your earlier posts, Mike) but I'm dealing with 80Hz and above here. Also, since I posted, I've read that room anomalies of 5-10db are common; I just wish mine occurred at 50KHz

 

[Guy Usher]

Jon, Do try putting your sub where you sit and then walk around with your meter to find the best spot, a lot easier than moving the sub all around, I have done this before and it works well. I too have the Stereophile CD, I know some of those were way off, I see yours is #3 so you are aware of this. Also I have been told that warble tones are not accurate enough for any good evualuation. I think we all agree the meter is just a tool for "Quick and dirty" setup, final tweaking for bass response still by ear.

 

[JonStern]

Guy,

In fact, at 50Hz the response in my room is flat according both to the meter and my ears. It remains flat down to the limit of the Polks (40Hz) and then my sub takes over (if it's turned on, which it wasn't when I was doing the testing above) and I've tuned the low pass and level so that everything's flat down to the low 20s. I also don't have any serious dips/peaks above 125Hz. I've established that I have that sizable dip at 80Hz but no other problems. I agree that the Radio Shack meter is only sufficient for rough analysis but that's more than enough to reveal the problem in my room.

 

[GregLee]

You could test whether the dip at 80Hz is due to cancellation between the fronts by disconnecting one of the fronts and repeating the measurement. Well, this is obvious, but you don't mention having tried this. Also, it would be interesting to see what happens when you reverse the speaker connections on one of your fronts.

 

[Guy Usher]

At first I thought you were using your sub, being at your 0db level at 50 doesnt mean it is not the one thats hot but since you are not using the sub it rules that out. I know in my case if I was at 0db at 50 I would be down over 12db at 80-125 all because my house was so hot at 50 hz, when I would run the test tones in my NHT super ones I would be 12db hot at 50hz, now I know the NHTs are really struggling to make 60 much less 50 . . . my head is starting to hurt again. . .

 

[Thomas J. Coyle III]

Hey Mike and Jon
Mike is right that the Rat Shack SPL meter can be way off in certain areas of the audio frequency band. However, several members of HTF have provided offset adjustments for the SPL meter to correct it's inaccuracies. If you do a search on the SPL meter, I am sure that you will find them. I know that Sonnie Parker has them on his website.
Regards,
TCIII