That graph is merely showing the deviation in response based on a hypothetical "zero" point at the top. You're thinking of actual SPL level. If it's easier for you to understand, just add 85 mentally to all the numbers.

--Ethan

Thanks Ethan.
1)Could you also tell me what deviation (peak to peak) can be considered as good enough for a listening room at home where we cannot put a lot of treatment material due to reasons like rental- apt/cost/aesthetics.
2)Will the reasonable deviation levels be the same for low and high frequency ranges or should we have seperate them to two zones?

Quote:
Could you also tell me what deviation (peak to peak) can be considered as good enough for a listening room at home where we cannot put a lot of treatment material due to reasons like rental- apt/cost/aesthetics.

It depends partly on the size of the room. Most home-size rooms have peaks and nulls in the bass range spanning 30 dB or more. If you can get that down to a 10 dB span you're doing really well. This requires a fair number of bass traps. But if you can't drill holes they can be mounted on stands, or leaned against the walls etc.

Quote:
2)Will the reasonable deviation levels be the same for low and high frequency ranges or should we have seperate them to two zones?

I divide the ranges around 300 Hz or so. At mid and high frequencies, measured peaks and nulls are normal and not necessarily due to room acoustics. This is a deep subject that requires more explanation than I have time for. This article is about something else, but it talks about what causes peaks and nulls, and shows low and mid/high frequency graphs:

A common-sense explanation of audiophile beliefs

--Ethan

Ethan, I own a Velodyn DD 12 subwoofer that comes with a freq. sweep generator, video output (for displaying the measured room response) and a microphone.

When I first set up my system I ran the freq. sweep through my Hales towers full-range (no sub) and observed the response via my TV. According to the velodyn system the room response peaks at 32 hz then drops down approx. 10 db to a null centered at 50 hz. From there the response climbs back up to 90 hz where the level is equal to that of the 30 hz peak. The bass response is then more or less flat out to approx. 170 hz where a dip begins and continues down to a small plateau at the 200 hz. test limit.

While the above was displayed on my TV, I moved the Hales towers around to try and improve the above result. Strangely, there was very little change. But when I started moving the microphone (which was at the listening position) around the changes did ensue. It turns out that--in my room--the major room peak/null nodes are strongly linked with the listening position while the speaker position has little effect.

I should state here that the listening room encompasses the living room, foyer and kitchen. The speakers see an open space approx. 60 feet (from front of house to back of house). The ceiling slopes from 9 ft. to 11 ft., and the immediate listening area front to back dimension is 19 ft. The speakers fire across that 19 ft. space.

Now, while the speakers have no side-walls/corners near them the listening position has both. So, I ended up moving my couch out approx. 1 1/2 feet further from the back wall. This did fill in the null somewhat, but the overall problems are still there. I can't really move the couch any further out into the room, and the bass does get a bit muddy at loud volumes.

Can bass traps tame the peak at 30 hz.? Also, do you know if the Velodyn's room measurement system is sufficiently accurate?

PS: I wanted to post a photo of the room response graph, but I'm no computer wiz and can't figure out how to post a picture from iPhoto. Other sites are no problem, but this stereophile.com URL thing has me confused.

Quote:
The speakers see an open space approx. 60 feet

This is good and bad. Good because large spaces have fewer bass problems, Bad because there's much more surface area that needs treating.

Quote:
Can bass traps tame the peak at 30 hz.?

Not unless you have a lot of them. But 30 Hz is nowhere near as damaging musically as peaks and nulls above 80 Hz.

Quote:
do you know if the Velodyn's room measurement system is sufficiently accurate?

I have no idea. Do they say if it's 1/3 octave, or 1/6 etc?

Quote:
I wanted to post a photo of the room response graph, but I'm no computer wiz and can't figure out how to post a picture from iPhoto.

If the photos already exist somewhere, either post a link to them here, or you can embed them directly in your post using image tags:

Code:[img]http://www.website.com/photo.jpg[/img]

--Ethan

The Velodyne generates a sine wave sweep from 15 Hz to 200 Hz. The resolution appears to be 1/3 octave as measured by my Mk1 eyeball.

The adjustment is via eight 1/3 octave bands of EQ, the first band centered on 20Hz and the last centered on 100Hz (graphic, can also set bands to parametric and vary both Q and center frequency).