[pjchappy]

Here are 2 pics of the first measurement of my untreated room, from my listening position  (see my system below for details). 

Right now, it's of just my monitors.  I will be (trying to) integrate my sub this weekend.

I used Room EQ Wizard with their calibration settings for the RS digital SPL meter.  I will later try the same measurements with a microphone.

This is of both channels at once (will post individual later).  I know the RS meter (even calibrated through REW) isn't accurate past 3KHz.  So, the first graph is from 15 to 9KHz and the second one is the same measurements, but the graph goes out to 20,000KHz. 

I plan on doing major room treatments in a month or so.  I will be hanging up some old Eighth Nerve stuff later today and then take more measurements.  I'm open for suggestions.

[TRADERXFAN]

Could you post the waterfall plot, too?

-Tony

[pjchappy]

I have never done this before.  Not sure if the graph is set correctly.  Let me know if I need to change it.

I have no clue what this plot even means. . .   

[youngho]

I made some suggestions in your other thread.

In it, I discussed the significant cancellation you'd expect around 150 Hz (depending on how accurate your measurements were) because of the distance of the speakers from the front wall. This is why I recommended treating the front wall behind the speakers.

In your graphs, you see the effects of the first- and second-order length modes, which you can calculate by 565 divided by the length of the room (11.4'), so ~50 Hz and ~100 Hz. In your plot, you see that they ring, particularly the ~50 Hz one. This is why I also recommended treating the back wall behind the listening position.

The third-order length mode (150 Hz) is cancelled out by the adjacent boundary effect of the front wall behind the speakers, but the fourth order mode (200 Hz) is a minor peak. At 300 Hz, the adjacent boundary effect results not in cancellation but reinforcement (bounces off and arrives in phase, not out of phase).

I have no new suggestions to make. Best of luck, and happy listening.

[pjchappy]

Thanks again.  I have some Eighth Nerve treatments up now.  I also can treat the wall behind the speakers (a little, for now. . .) with some 4" acoustic foam.  I will post measurements shortly.

Next month, I'll have a lot better treatments.

[youngho]

Okay, then, one more suggestion:

Measurements are EXTREMELY sensitive to placement, particularly for things like modes and comb filter effects, so any changes that you measure may be from not having the microphone in exactly the same place. Take your results with a grain of salt.

[youngho]

I have no clue what this plot even means. . .   

As I understand it, a waterfall plot is essentially a three-dimensional graph. The X-axis, which runs horizontally, is the frequency spectrum from 20-500 Hz. The Y-Axis, which runs vertically on the left, is the amplitude from 30-88 Hz. The Z-axis, which runs back to front but is notated by the numbers and bars in the upper right side of the plot, is time, from 0-300 msec.

Imaging taking the graphs that you got in your first post, measuring the response again 10 msec later, overlaying them on the original graphs but moved slightly down and to the right, measuring them again 10 msec later, overlaying them on the original graphs but moved slightly down and to the right from the last overlaid curve, repeat and repeat and repeat until you reach 300 msec.

This is why "back to front" (0-300 msec) runs down and to the right. It's basically a two-dimensional representation of a three-dimensional graph. The "contour lines" each represent a different "slice" in time. You can follow the contour lines to look at what happens in the frequency spectrum at that point in time. You can look perpendicular to the lines ("back to front") to see what happens over time for a particular frequency.

Looking at time=0, you can see that there are actually two relatively close peaks around 50 Hz (one slightly below, one slightly above). Over time, these form one peak that "rings" (has high amplitude that persists over time), due to pitch bending or shifting. I don't know why there are two originally. I wonder whether your room may be slightly smaller than you measured. I also don't why the peak is higher, rather than lower, than 50 Hz.

At 300 msec, you can see more peaks at ~22 Hz, 30 Hz, 32, and 38 Hz (I'm guessing, but I'm really bad at logarithmic graphs). These may represent the effects of some tangential and oblique modes, though you don't really see most of them at time=0. Again, you see evidence of pitch shifting here, as the frequency of the developing peaks moves downward.

You can't have missed the big "boom" around 50 Hz, and that's because you and the speakers are positioned to maximally excite this mode (and, to a lesser extent, two of the tangential ones). You may not have noticed the peak around 100 Hz as much, as it is less in amplitude and also Q (the width of the peak), and you may not have noticed the cancellation as much, since we tend to perceive peaks more than nulls.

Interestingly, although one might predict a problem with the height mode at multiples of 63 Hz, you don't see that reflected in the graph. As I've argued with Bryan before in the past, most listeners sit with their ears relatively close to 36" off the ground. For standard construction with 8-9' ceilings, that means that they're positioned between 33-38% of that distance, which positions them nearly midway between the node of the first height mode and an antinode of the second height mode. As your graphs and plots suggest, the height modes don't seem to play an obvious detrimental role in your room.

I hope this helps a little in beginning to interpret the waterfall plots. The experts will be able to help you more. Keep in mind that there's generally a tradeoff between resolution in the time and frequency domains.

(edited to replace "front to back" with "back to front")

[bpape]

Exactly correct on the waterfalls.

Predicted modes don't materialize at times for a variety of reasons - not the least of which is an offsetting issue at the same frequency range from another dimension, SBIR, etc.

As for the height, the rule of thumb I use is 40-42" for seated ear level unless you're under 5'6" and/or you have extremely low seating. 

Std 8' high rooms (96") compared to 40-42" is approx 41-43%.  Anywhere between 40 and 50% is a place to be avoided if possible - which in the height, rarely can be.  9' ceilings don't present the problems that 8' ceilings do.

For measurements, try no smoothing and just run the sweep from 20-400hz.  That will give you a lot more detail as to what's really happening.  Young is exactly correct that mic placement is extremely important.  Get as close to your seated ear position as possible.  Try not to move the mic at all before and after treatments.  You can get huge swings in response without doing anything in a room but moving the mic a few inches.  This reinforces how important seating position within the room is.

[youngho]

As for the height, the rule of thumb I use is 40-42" for seated ear level unless you're under 5'6" and/or you have extremely low seating. 

Std 8' high rooms (96") compared to 40-42" is approx 41-43%.  Anywhere between 40 and 50% is a place to be avoided if possible - which in the height, rarely can be.  9' ceilings don't present the problems that 8' ceilings do.

That's right, I forgot why we disagree on this issue. I'm short (5' 6"), and I often slouch in my chair as I relax. Your clients must be taller with very good posture. Here were some of the sorts of information/links that had informed my thinking on the issue:

http://www.goodsound.com/howto/2007_08_01.htm
http://www.stereophile.com/floorloudspeakers/1208thi/index4.html

Well, this issue certainly highlights the need to have the microphone properly positioned at the level of the listener's ears.

[pjchappy]

Thanks again, guys.

I haven't even had a chance to really listen to my new room / set-up too much.  Yes, I definitely noticed some boominess, but with my monitors, it hasn't been too, too bad.  I REALLY notice it with my guitar amp, though.

I have some room treatments up now and took some measurements from an alternate seating position, which is sitting on a pillow, right in front of my couch.  I do need taller speaker stands and this position is actually more ideal in terms of my ears at tweeter height.  My ears are about 1.5" below the tweeter in this position.  If I scoot up even more, just for experimentation, it sounds very nice.  Basically nearfield listening at that point.

As for my room, I plan on taking your advice and treating the front and back walls.  I will put three 2' x 4' x 4" 703, 705 or wool Rockboard on each side of the window, behind the speakers.  I will put two 2' x 4' x 4" behind the listening position.  Probably will make two 2' x 4' x 6" pieces for the corners behind the speakers.  I will just have these kinda leaning in the corners.  One of them, I will have to move when I'm not listening.  I will make one 2' x 4' x 4" piece that I can move to set on the floor in front of me and another of the same to lean up against the TV.

The Rockboard looks to have better low-end absorption and is cheaper.

Here's a diagram of my room, not to scale.  The ceiling is actually 7.9'.

[Nyal Mellor]

Like others have said, that big dip at 150hz is speaker boundary interference off the back wall. Treatment is really the only good way to get rid of this - you'll need 4" of fibreglass to be really effective at 150hz.

The 50 and 100 hz modes you should be able to eliminate most of moving your seating position in the room. Try taking measurements in 6" increments forward and backwards of your current position and post the results here when you find something that looks flatter.

After that its going to be a case of applying EQ to get rid of the final modal peaks.

[pjchappy]

Well, I figured it was pointless to post any more measurements pre-treatment.

I currently have eight 2' x 4' x 4" frames ready for OC 703 (two 2" layers; back layer is the FSK type 703).  The frames are made from cheap 3/8" plywood.  I just used butt joints, wood glue and L-brackets to assemble them.  The wood was under $10 for a 4' x 8' sheet.

I plan on building at least two more of the same type of frame, plus 2 (maybe 4) that are about 3" deep w/ one layer of 2" 703.  May also make a couple skinny ones for the corners (thinking 1' x 4' x 4" or 6") . . . and maybe something similar for where the wall and ceiling meet (would just make the frames 2" deep for these).  The front wall will be treated with at least six 2' x 4' x 4" panels.

Do NOT order 703 off the Internet!!!  I got 192 square feet for $214 at a local SPI.  48 of that is the 703 FSK.  I likely will have extra 703 and may make my brother-in-law some traps.

Should finish these up this weekend (out of town again) and will hopefully get a chance to takes measurements (and pics).  Just need to get some burlap, as the local Hobby Lobby didn't have enough.  The lady there wasn't the sweetest person, either.   

[bpape]

If you have a local SPI, that's definitely the way to go.  Not much cheaper than I sell it for but no shipping.

Bryan

[pjchappy]

Would 10 oz. canvas be OK to use for covering the panels, instead of burlap? 

[bpape]

For bass absorbers it should be fine. For reflection panels, make sure you can easily blow through it or it may reflect upper mids and highs.

Bryan

[pjchappy]

Thanks!

[pjchappy]

Alright, due to a hard drive failure (had problems after something fell on my laptop. . .and went downhill from there), I won't be able to measure my room until Sunday at the earliest (and I will also post pics).

Anyways, I have done the following to my room:

- One 2' x 4' x 4" 703 panel (w/ backing) behind each speaker

- One 1' x 4' x 6" 703 panel (open back; fully covered in burlap) in each corner behind each speaker (angled into the corner, with about a 5" to 6" gap from the corner)

- One 2' x 4' x 4" 703 panel (open back) at each first reflection point on the side walls (have removed the bookcase on the left side of the room and moved my guitar amp about 3' farther away from the right speaker)

- Two 2' x 4' x 4" 703 panel (w/ backing) directly behind my listening position

- One 2' x 4' x 4" 703 panel (w/ backing) I can set on the floor at the reflection point

- One 2' x 4' x 4" 703 panel (w/ backing) I can put covering the TV (in brief listening sessions, I have found this ruins soundstage and best placement is on the floor lengthwise, leaning on the TV stand, just below the TV).

- Eighth Nerve triangles in all four ceiling corners

- One Eight Nerve seem above the window (I will replace this with two 1' x 4' x 4" 703 panels (open back) angled like a seem and butted together for an 8' "seem")

- May actually be able to put a panel or two on the ceiling at the reflection point (2" thick)

- May make two 1' x 2' x 6" (maybe 8"?) boxes to place on the floor against the wall, below each panel on the wall behind the speakers

I also scoot up in my seat, which puts my head about 2 1/2' from the back wall.  No boominess there and much better imaging and soundstage.

I'm VERY happy so far.  Immediate and dramatic improvement!    Spent A TON of time making these panels (eight 2' x 4' x 4" framed panels; two 1' x 4' x 6" framed panels; two 1' x 4' x 4" and possibly two 1' x 2' x 6" framed panels to go. . .)

I have enough wood and fiberglass to do even more, if necessary.  Also, big correction in the square footage / price.  I got 192 sq. ft. of 703 (without backing) and 48 sq. ft. of 703 (with backing) for $214.   

I will be re-integrating my sub soon. . .and should be able to make a MAJOR speaker upgrade in March or April.   

For now, I have also raised my speakers about 3" to make the tweeter at ear level (or at least close. . .may need to raise them a bit more. . .will need help to check.  I have a laser to help me out.)

Drilling 2" holes in eight panels was a MAJOR pain in the ass.  On the 6" thick corner bass traps (open back), I cut a 2" slot down the middle of the sides (may not have even been necessary, but was much easier than drilling those holes!)


      


 

[pjchappy]

A brief description of the panels' construction:

They are made of 3/8" plywood.  (Under $10 for a 4' x 8' sheet at Home Depot, which they cut into 2' x 4' panels for free. . . and I used a table saw to do the rest)  Put together with butt-joints, wood glue and two L-brackets ($.50 each) (screwed in) for each joint.  I used 4' x ~ 1 1/2" strips of wood on each side of the back to hold the fiberglass in place, which were glued and nailed in place.  Covered the fronts, tops and sides with one-piece of off-white burlap ($2.99 / yard - 40" wide), which was simply stapled in place on the back-side.  They are hung with two picture frame hangers each.

Took me three and a half VERY long days to make them.  Once I made a jig and got the construction down, things sped up dramatically (other than those damn 2" holes!!!!  )  The L-brackets squared things up nicely. . .and became perfectly square once I installed the fiberglass panels.  They look pretty good!

[Stylus]

Pics!!!  aa

After your help finding my 703 I'll be right behind you, it sounds like a lot of work, but worth it!

I am not understanding the 2" holes, what are they for?

[pjchappy]

So as much of the sound waves get into the fiberglass.  Made the holes on all 4-sides of the eight 2' x 4' panels I made. 

Now, if all that is totally necessary, beats me!  However, I've seen it done by others here and figured if I'm going to do it, I will do it right.

Jumping right in and just using the L-brackets to construct it would have saved me hours of time.  I was using a framing square, etc. and found just slapping on some glue, then just screwing on the L-brackets was all I needed to do to keep it (fairly) square.  Again, putting the fiberglass panels in did the rest.

I'll post pics Sunday.  I won't be home this week, as I'm house / dog-sitting.