need help: using REW to find reflections in a narrow frequency range

[zathrez]

I'm trying to finalize my "media room". I've got one big-ish problem left. The room has a response peak right around 815 Hz. I can't find the source. Nothing I do even changes it much.

This peak is why I finally downloaded REW (V5.31.3), bought a umik-1, and started trying to find it the right way. What I've found in the last month is that REW is large and complex, and I don't know enough to use it well. At least not yet.

I've attached a couple of REW mdat files. Both of these are using the single R speaker of the main L/R pair. This is with no EQ or subs (Audyssey off because Denon set to "pure direct" mode). The first file shows the big mid-range peak as heard by the mic set at the main listening position (MLP) at my ear height when I'm sitting there. The second file shows what the speaker is actually playing when REW does its chirp (from 1m (39") out front of speaker, at the height of the tweeter, and pointed directly at the tweeter, with the correct calibration file selected). Both speakers are about 1m (39") out from the front wall, and also from the side walls. Which puts them out in front of the TV. Slightly toed in.

Some questions come to mind:

1) How does a room build up a peak like that? That's in the range of 6-10 dB higher than what the direct sound from the speaker delivers. So, some sort of resonance in the room? Or maybe more probably, the layering on of reflections? Where is the extra energy coming from? I wish I'd had more acoustics training...

2) What's the most effective method of using REW to find the reflections that make up this narrow midrange peak? REW is so big that it probably has good tools for stuff like this, but I don't know what they are, much less how to use them.

3) Now that I know I've got a problem, how do I fix it? I don't want to just throw EQ at it. For one thing, I've tried that and Audyssey couldn't cope with it. It would only bring it down maybe 3 dB and it was still easily audible. That 820 Hz range is where a lot of female singers work, and lots of classical acoustic piano. With those it's hard to ignore (everyone who's heard this room coming together has heard this peak, it is not just me), even with Audyssey working its tail off. So in order to throw EQ at it, I first have to reduce it to manageable proportions.

Context:

Room is about 13.7' x 20.5' x 7.9'. Two walls are outside walls and are basic block and brick, finished with 2x2" furring strips and 5/8" wall board. Two windows in the long wall, one in the short wall that the MLP faces. These two outside walls have an HVAC duct (because of course). The soffit around the ducts is pretty stiff since it had to be remade. The tearing out and replacing of this duct work is what required the "remodel" of this room in the first place.

The long interior wall has two doors in it, both closed when in use. The rear wall has a double door that can open to the rest of the basement, but is closed when in use. I'm just saying this is a sealed room and not an "open plan" room.

The room is well treated with GIK panels. All four corners are floor to ceiling tri-trap bass traps. The walls and ceilings have about a 50/50 mix of absorption panels and diffusion / absorption panels. Maybe 28 panels in addition to the corner bass traps.

The room works great for HT. What I'm trying to do here is make it also work great for old school stereo. My wife likes 1940's Frank Sinatra, if you play some of those CDs this room sounds wonderful because practically everyting is below that 815Hz peak.

So... am I screwed? Or is this fixable?

 

[ddude003]

Since you have so many GIK products, I would suggest that you contact GIK and see what they think...

 

[solo541]

I'm trying to finalize my "media room". I've got one big-ish problem left. The room has a response peak right around 815 Hz. I can't find the source. Nothing I do even changes it much.

This peak is why I finally downloaded REW (V5.31.3), bought a umik-1, and started trying to find it the right way. What I've found in the last month is that REW is large and complex, and I don't know enough to use it well. At least not yet.

I've attached a couple of REW mdat files. Both of these are using the single R speaker of the main L/R pair. This is with no EQ or subs (Audyssey off because Denon set to "pure direct" mode). The first file shows the big mid-range peak as heard by the mic set at the main listening position (MLP) at my ear height when I'm sitting there. The second file shows what the speaker is actually playing when REW does its chirp (from 1m (39") out front of speaker, at the height of the tweeter, and pointed directly at the tweeter, with the correct calibration file selected). Both speakers are about 1m (39") out from the front wall, and also from the side walls. Which puts them out in front of the TV. Slightly toed in.

Some questions come to mind:

1) How does a room build up a peak like that? That's in the range of 6-10 dB higher than what the direct sound from the speaker delivers. So, some sort of resonance in the room? Or maybe more probably, the layering on of reflections? Where is the extra energy coming from? I wish I'd had more acoustics training...

2) What's the most effective method of using REW to find the reflections that make up this narrow midrange peak? REW is so big that it probably has good tools for stuff like this, but I don't know what they are, much less how to use them.

3) Now that I know I've got a problem, how do I fix it? I don't want to just throw EQ at it. For one thing, I've tried that and Audyssey couldn't cope with it. It would only bring it down maybe 3 dB and it was still easily audible. That 820 Hz range is where a lot of female singers work, and lots of classical acoustic piano. With those it's hard to ignore (everyone who's heard this room coming together has heard this peak, it is not just me), even with Audyssey working its tail off. So in order to throw EQ at it, I first have to reduce it to manageable proportions.

Context:

Room is about 13.7' x 20.5' x 7.9'. Two walls are outside walls and are basic block and brick, finished with 2x2" furring strips and 5/8" wall board. Two windows in the long wall, one in the short wall that the MLP faces. These two outside walls have an HVAC duct (because of course). The soffit around the ducts is pretty stiff since it had to be remade. The tearing out and replacing of this duct work is what required the "remodel" of this room in the first place.

The long interior wall has two doors in it, both closed when in use. The rear wall has a double door that can open to the rest of the basement, but is closed when in use. I'm just saying this is a sealed room and not an "open plan" room.

The room is well treated with GIK panels. All four corners are floor to ceiling tri-trap bass traps. The walls and ceilings have about a 50/50 mix of absorption panels and diffusion / absorption panels. Maybe 28 panels in addition to the corner bass traps.

The room works great for HT. What I'm trying to do here is make it also work great for old school stereo. My wife likes 1940's Frank Sinatra, if you play some of those CDs this room sounds wonderful because practically everyting is below that 815Hz peak.

So... am I screwed? Or is this fixable?

What is the distance between your speakers and the TV? If this distance is 40–42 cm, then this is a reflection from the TV. You can check this by taking a long string, measuring the distance from the speaker to your ears, and adding 42 cm to this distance. Place a stand next to the speaker or ask someone to hold one end of the string at the front panel of the speaker and the other end of the string at your ears. The string will sag. Take the free end of the string and try to see if it touches the TV or any other surface. Since the wavelength at this frequency corresponds to 42 cm, the distance travelled by the reflected sound will correspond to the direct distance travelled by the sound plus the full wavelength at 815 Hz.

 

[zathrez]

Since you have so many GIK products, I would suggest that you contact GIK and see what they think...

Yeah. Never would have thought of that. ;-)

Over the last month GIK and I exchanged maybe 30-40 emails about it. Pix, REW measurements, moved panels around, had me crawl all over the walls and ceiling looking for a resonance (didn't find anything buzzing or anything that I could damp with my hands that would even change the response, let alone move it in the right direction). GIK has been great to work with, and the room mostly does exactly what it should do thanks to their consultants. I'm happy with them and their products.

In the end however, they suggested that I'd have to have someone local come for a physical visit to this room. They might well be right. If I find anyone local that is a "room fixer" I'll do that. But I haven't found that level of acoustical consultant yet.

And that's why I'm here. Still trying to find out what the heck is going on in this room with the only viable tool I have. REW is big and complex and I need help with it. How to use it, how to interpret results. I'm really not here to waste anyone else's time. I'm here because the low hanging fruit has already been picked and found to be wanting.

 

[zathrez]

What is the distance between your speakers and the TV?

About 8 cm. It's a big tv, on the small dimension of the room. The speakers are about 0.9m off both the front wall (behind the TV) and off the side walls as well. The front of the speaker is about 19 cm in front of the TV screen. I've covered the both front and back of the TV with the two heaviest blankets we have here, REW indicated it didn't move the needle of the room response at all. At least not in this 815 Hz region.

I just went through a series of moves with the right speaker. Trying to see what I can do to change this 815 Hz peak. Moved this speaker forward and backward 20 cm and 40 cm, and moved it to the right 20 and 40 cm, so 9 new measurements. The response changes some, but that peak just sits there.

Since I was there, I put the speaker back to it's "normal" position and tried a series of rotations. Straight ahead (which doesn't do the stereo field any good, but I'd learn to live with it if that peak would come down any), then toed in a few degrees at a time. This did almost nothing to the response at all, especially this 815 Hz peak.

I can't believe that speaker placement doesn't effect this. Boggles my mind. Because reflections in the room is about all that's left, and speaker placement is usually all about the reflections in the room.

If this distance is 40–42 cm, then this is a reflection from the TV. You can check this by taking a long string, measuring the distance from the speaker to your ears, and adding 42 cm to this distance. Place a stand next to the speaker or ask someone to hold one end of the string at the front panel of the speaker and the other end of the string at your ears. The string will sag. Take the free end of the string and try to see if it touches the TV or any other surface. Since the wavelength at this frequency corresponds to 42 cm, the distance travelled by the reflected sound will correspond to the direct distance travelled by the sound plus the full wavelength at 815 Hz.

I've done the "string trick" too. Turns out nothing in the room is 42cm away from the "flight path" of the direct sound from speaker to MLP where the Umik-1 is listening. None of the first reflection points, not even the floor or the ceiling. None of them are even at 2x 42 cm. It's... odd. I don't know what else to call it. Not really believing that, I pulled a couple of panels off the walls and put them on the floor along the line between the speaker and the MLP couch. Did not move the needle. So it's not the floor bounce.

It's almost like it's got to be a double or triple bounce. But it doesn't seem like there's enough time for that.

So... how do I get REW to tell me about these bounces, and the timing of these bounces? It shows me *all* the bounces and I can't get a clear picture of the ones that are associated with this 815 Hz peak.

 

[solo541]

About 8 cm. It's a big tv, on the small dimension of the room. The speakers are about 0.9m off both the front wall (behind the TV) and off the side walls as well. The front of the speaker is about 19 cm in front of the TV screen. I've covered the both front and back of the TV with the two heaviest blankets we have here, REW indicated it didn't move the needle of the room response at all. At least not in this 815 Hz region.

I just went through a series of moves with the right speaker. Trying to see what I can do to change this 815 Hz peak. Moved this speaker forward and backward 20 cm and 40 cm, and moved it to the right 20 and 40 cm, so 9 new measurements. The response changes some, but that peak just sits there.

Since I was there, I put the speaker back to it's "normal" position and tried a series of rotations. Straight ahead (which doesn't do the stereo field any good, but I'd learn to live with it if that peak would come down any), then toed in a few degrees at a time. This did almost nothing to the response at all, especially this 815 Hz peak.

I can't believe that speaker placement doesn't effect this. Boggles my mind. Because reflections in the room is about all that's left, and speaker placement is usually all about the reflections in the room.

I've done the "string trick" too. Turns out nothing in the room is 42cm away from the "flight path" of the direct sound from speaker to MLP where the Umik-1 is listening. None of the first reflection points, not even the floor or the ceiling. None of them are even at 2x 42 cm. It's... odd. I don't know what else to call it. Not really believing that, I pulled a couple of panels off the walls and put them on the floor along the line between the speaker and the MLP couch. Did not move the needle. So it's not the floor bounce.

It's almost like it's got to be a double or triple bounce. But it doesn't seem like there's enough time for that.

So... how do I get REW to tell me about these bounces, and the timing of these bounces? It shows me *all* the bounces and I can't get a clear picture of the ones that are associated with this 815 Hz peak.

If this is not SBIR, then it could be a modal issue.

 

[solo541]

Regarding reflections, the very last message in your thread contains a link to a calculator that you can use to identify reflection surfaces. https://gearspace.com/board/studio-building-acoustics/1464707-weird-vexing-820-hz-peak-room.html