use the two rear panels that you have on the side walls, roughly centered behind the speakers.
Ok, so as mentioned above...
It looks like you misunderstood. When I say the panels should be centered behind the speakers, this is what I mean:
temp_room-1.png
In other words, each panel is centered behind each speaker.
Adam Audio A8X.
Nice! Good choice. Set the low shelving filter to roughly -5 dB. That's to compensate for the "baffle step response" issue. Since the speaker is now very close to the front wall, it no longer needs the built-in baffle step power imbalance adjustment, so you can effectively remove it by moveing the "low shelf" filter down. The speaker is now radiating into half-space, effectively, so the full-space power correction is not needed anymore. You can see the problem in this image:
bass-rise.jpg
The bass rise, starting at somewhere around 500 Hz very roughly, is clearly visible. That graph is smoothed to one-third octave, to highlight the overall effect, rather than the details.
I can't tell you how to set the other two, because you are cutting off your REW tests way short! Your tests only go up to 5 kHz, so you are missing the entire high frequency range. Your Adam's easily produce way above 20 kHz. (one of their impressive features), but even for lesser speakers that only go up to 20 kHz, you still need to let REW run the entire spectrum and some more, to see how the top end is doing. So adjust REW to run up to about 25 kHz.
You also do not need to have REW starting way down at 2 Hz, in the extreme low end of the subsonics! Even whales can't hear such low frequencies...
Your Adam's only start doing interesting things at about 30 Hz, so set REW to start an octave lower, at 15 Hz. There's nothing at all useful down that low, your speaker's can't actually produce tones that low, your mic can't capture them, your interface can't convert them, and your system can't process them. so there's no need at all to include such a low range.
In other words, let REW run from 15 Hz to 25 kHz, to capture the full spectrum of what your speakers are doing, and how the room is reacting to that.
So leave the other two controls at 0 for now: we might need to tweak them later, but the info for doing so is not in your data.
One other thing: In addition to the REW tests with only left and only right speakers, I also need to see tests with BOTH speakers on at once.
the acoustic axis of each speaker should be about 1.2m from the side wall. That means your speakers will be about 1.9 m apart.
Ok, done. Should be clear in the new sketchup drawing.
That's not what I'm seeing in the SketchUp! It shows that your speaker's acoustic axes are 127 cm from the side walls, instead of 120, and that they are only 180cm apart, instead of 190cm, and not equidistant from the side walls either! This is what you are showing:
temp_room-1b.png
Move each one about 7cm closer to the side walls, and make sure they are the SAME distance from the walls. Yes, 7cm matters! That is a full wavelength for 5 kHz, a half wavelength for 2.5 kHz, and a quarter wave for 1.2 kHz.... Yes, symmetry matters for the same reason. And accuracy too...
First an explanation of the position of those rolls. After checking the SPL and waterfall graphs in REW I then played a sine wave at the most troublesome frequencies and tried to find the peaks by ear. I placed the rolls there.
Yep, but that's the wrong place to put them! You want maximum absorption of MODAL energy: all modes terminate in room corners, not in the middle of the room, so placing bass traps in corners is where you get maximum effect. Also, you were listening for peaks, not nulls: Peaks are easier to deal with, and might not even be due to modal problems: that could be SBIR or other issues. Modal nulls are the problem, and you can't treat a null by putting absorption in it! You have to put the absorption where the peak is, which is always in the corner...
But now I changed the position as you suggested
... and you can see the much improved response on your graphs...
Only problem is the fact that in the left back corner there is a door.
That's why John's comment on the forum rules ask that you take care to specify such things in your first post!
You'll probably need to build a bass-trap on wheels to deal with that. One you can just push into place in front of the door when you are doing critical listening and mixing, but push out of the way to open the door.
You'll need to do a REW test with rolls in that corner as well.
What treatment do you have on the ceiling at present?
None... I was planning on making one other panel and hanging that above the mix position. It's a problem because I can even here some flutter echoes...
You will DEFINITELY need treatment up there! The reflections and vertical axis modal response are very, very clearly visible in your graphs. I would suggest that you make a very large, deep, hard-backed cloud to hang over the area between the speakers and your head, and angle it fairly steeply (lower over the speakers, higher over your head). Also put additional absorption on top, not just below.
I would also like to ask when you check the REW file that you maybe give some explanation as to how you interpret the data.
Wow! Well, there's a lot of things that I look at, in different ways, to tell me different things about the room, and compare that to what it should be ideally. For detecting the bass rise issue, for example, I smoothed the graphs as much as possible to get rid of the details and see only the overall curve, then I zoomed in to show only the lows and mids, the placed a cursor line at the average level (about 72 dB) so the rise is clearly visible above that.
Then, for the modal stuff I di NOT smooth it at all, and looked at the waterfall plots and compared those to a modal prediction for your room: the modes are very, very visible. First, your 1.0.0 mode at 34.4 Hz is a monster!
Modal-1--1.0.0.--34.4.jpg
That's the fundamental lowest mode in your room, associated with the long axis (front to back), and it is HUGE! It also coincides with the lowest note on a 5-string bass guitar....
Your 0.1.0 mode is not visible: probably lost inside that monster at 1.0.0. But its fundamentals are there, very strongly. Both the 0.2.0 (99 Hz) and 0.3.0 (118 Hz)are very well represented:
Modal-2--0.2.0--99.2.jpg
Modal-3--0.3.0--118.jpg
And your 0.0.1 mode is also very much on the charts, at 63.8 Hz:
Modal-4--0.0.1--63.8.jpg
So you have major modal issues in all three axes. There's also tangential modes clearly happening, and even e hints of oblique modes.
In other words, your bass trapping is not working.
Next, I looked at the decay times, one third octave:
decay-one-third.jpg
... and found that your room is way too live. Your decay times seem to be around 400 ms, but for a room that size it should be around 200 ms. (Of course, I can't see what is going on in the high end, since you didn't let REW do that, but from experience I'd guess it is fairly constant from 4k upwards, probably tailing off a little more at the high end.)
So you clearly don't have enough absorption in that room.
Then I looked at the ETC curve (smoothed to 0.1ms):
ETC-2.jpg
... and saw that there are multiple reflections happening within the Haas time and at high levels. Zooming in on that, you can see that there are at least 14 reflections above -20 dB before 20 ms....
ETC-03-z.jpg
So directionality and reverb-tails will be lost in there: you won't be able to hear subtle clues in the mix, and you won't be able to determine sound-stage locations very well.
Then finally I compared pure SPL levels of the "before" and "after" situation (smoothed to 1/12th octave), and saw a large improvement with the new layout and new treatment:
SPL-before-after--2.jpg
The gold colored curve is the original "before" situation, the purple curve is with your setup and treatment, and the green curve is with my setup and treatment layout. You can see how very much smoother the green curve is...
So, my overall conclusion is that your treatment is insufficient, and not working very well. You are lacking bass trapping, overall absorption, and reflection control.
What type of insulation did you use in your panels, and what type are you using in your "roll" traps? I have a feeling that it is not suitable for what you are trying to do. The graphs don't show that the room is responding as it should to that treatment.
Anyway, please do new REW tests with the correct frequency range set, also with both speakers, and with two of the rolls in front of the door.
- Stuart -
.
Carpets are no good for room acoustics. They do the opposite of what you need. If you just wanted a small piece under your chair and desk, that would be fine, but don't carpet the entire floor!!!