Measurements with and without traps

[localhost127]

define 'treated'?
what type of material, thickness, how much area covered, etc?
photos of room?

[Rairun]

OK, sorry.
All panels are made of 4" thick, 50 kg/m3 fiberglass.
They are spaced approximately 10 cm from the boundaries. Some of them are FRK faced. (Blue ones on the picture.)
Approximate room dimensions are 4m X 3m X 2.7m
Please download the 360 degree Quicktime video that i made, you'll see a much better visualization of my room.
http://torecagli.com/Room0.rar (37mb)

Btw, how did that big dip at 80hz happen?

[takman]

Originally Posted by Rairun

Btw, how did that big dip at 80hz happen?

thats a good question. i am curious about that as well...

Notice at, hard to read.....say 120hz you have a peak for the untreated. Then the null at 80hz in after treated. They look like opposites of each other...very similar.

Fixing one problem can cause another surface. Since modes happen in multiples. Thats all i can tell you though.....as this is a subject im not well versed on.....

I assume these graphs are with both speakers firing? Run sweeps with one speaker at a time, and post graphs of left and right.

Also, looking at your room....you have a lot of absorption going on there. Looks like it would be a fairly dead room.

But, to me it appears you have too many panels on the walls, and not enough bass traps.

You should surgically treat first reflection points by using the ETC...and i think you would find that you dont need so many first reflection panels.

[Rairun]

I assume these graphs are with both speakers firing? Run sweeps with one speaker at a time, and post graphs of left and right.

Yes, the graphs are with both speakers firing. I'll do what you say. But i want to learn why that's important.

Also, looking at your room....you have a lot of absorption going on there. Looks like it would be a fairly dead room.

You're right my room is a bit too dead. May be I'll consider Ethan's suggestion:

Another option is to make the rear wall of a control room partially reflective and partially absorbent. You can do this by making the wall totally dead, and then covering it with thin vertical strips of wood to reflect some of the sound back into the room

Except first reflection panels, instead of removing some of the panels on the walls i may apply Ethan's suggestion to those panels. That may brighten the room up. And I'm considering FRK faced panels for the naked portions on the walls and remaining untreated corners. I suppose it improves low frequency response without making the room more muffled.

You should surgically treat first reflection points by using the ETC...and i think you would find that you dont need so many first reflection panels.

Instead of "surgically treating first reflection points" i preferred covering more area, that gives me the ability to move freely while mixing. Btw, what is ETC?

[Tom Higgins]

Envelope Time Curve, I believe

[takman]

Originally Posted by Rairun

Yes, the graphs are with both speakers firing.

Instead of "surgically treating first reflection points" i preferred covering more area, that gives me the ability to move freely while mixing. Btw, what is ETC?

Open REW, click the impulse tab. On the bottom, you will see ETC...check that box and uncheck all the others.

This is showing you the actual reflections in time. Vertical scare is DB's, and horizontal is time in milliseconds (ms)

During the first 20ms ( possibly shorter depending on the size of the room) you want to have all the reflections at least -20 of the initial signal ( 0 db, at 0 ms ) This is the Reflection Free Zone, shown on a graph.

if you fire both speakers to check the ETC. It will be difficult to determine which speaker was causing what reflection. It just complicates things. you fire one speaker, determine reflection points, then same for other speaker.

[Ethan Winer]

I got your PM. I'm not posting much because I'm up to my eyeballs writing my audio book. But I'll gladly chime in.

Originally Posted by Rairun

What do you think?

I think After is better, but it looks like the measuring microphone was not in exactly the same place as when you measured Before. If the microphone is moved even half an inch between one measurement and the other, a peak can become a null. Strange but true. Proof here:

A common-sense explanation of audiophile beliefs

See Figure 1 especially.

--Ethan

________________
The Acoustic Treatment Experts

[Glenn Kuras]

Originally Posted by Ethan Winer

I got your PM. I'm not posting much because I'm up to my eyeballs writing my audio book. But I'll gladly chime in.



I think After is better, but it looks like the measuring microphone was not in exactly the same place as when you measured Before. If the microphone is moved even half an inch between one measurement and the other, a peak can become a null. Strange but true. Proof here:

A common-sense explanation of audiophile beliefs

See Figure 1 especially.

--Ethan

________________
The Acoustic Treatment Experts

I would think the same thing. The other thing it could be is you had a null and a peak canceling each other before, but when you put the panel on the all it could have caused the null to show up. Hard to explain but it can happen.

[boggy]

Originally Posted by Glenn Kuras

I would think the same thing. The other thing it could be is you had a null and a peak canceling each other before, but when you put the panel on the all it could have caused the null to show up. Hard to explain but it can happen.

+1
Also I noticed gentle roll off at high frequencies, this might be different position of measurement microphone (missed sweet spot of loudspeaker), or different orientation of measurement microphone (microphone not point toward speaker).
Measurements might be informative, but only if it done very carefully... otherwise it's waste of time, for both sides... people that take (wrong) measurements, and for people that try to analyze it and explain what happened.

Cheers

Boggy

EDIT: I mean.... differences between graphs easily might be because different placement of speaker and microphone.... not because acoustic treatment

[Rairun]

Originally Posted by boggy

Also i noticed gentle roll off at high frequencies, this might be different position of measurement microphone (missed sweet spot of loudspeaker), or different orientation of measurement microphone (microphone not point toward speaker).

Yeah, i noticed the high frequency roll off. May be it's related to excessive use of broadband panels (absorbing high frequencies) in my room. Isn't it possible? There is really very little echo or reverb in my room (in terms of mid and high).
Even so, I'll recheck the mic position to see if there is any difference. But it is highly unlikely because i marked everything and took notes before moving things.

[boggy]

Originally Posted by Rairun

Yeah, i noticed the high frequency roll off. May be it's related to excessive use of broadband panels (absorbing high frequencies) in my room. Isn't it possible?.....

No, if your measurement microphone has a clear path to the tweeter and if it is pointed directly toward tweeter, and if tweeter is pointed directly toward your measurement microphone, there would not be any HF roll off (even in fully anechoic room).

EDIT: Orientation of microphone doesn't have influence to measured response in room below 300Hz. Only position of loudspeaker and microphone has this influence.

Cheers

Boggy

[Rairun]

Originally Posted by boggy

and if it is pointed directly toward tweeter
Boggy

The mic is at tweeter height but not pointed at tweeter. Does it make difference?

[boggy]

Originally Posted by Rairun

..... Does it make difference?

Not for low frequencies, only for high frequencies (already noted HF roll-off is because microphone orientation)

EDIT: try not to change height of microphone during measurements, also don't change position of microphone and speakers, if you like to know what exactly your bass traps doing.

Cheers

Boggy

[abechap024]

For what its worth, ARC instructs its users to point the omni directional microphone at the ceiling

[Rairun]

Thanks abechap024

Before and after Impulse Responses seem very different, is this normal?

[DanDan]

Yes, the graphs are with both speakers firing. I'll do what you say. But i want to learn why that's important.

Two speakers or one speaker for ETF,REW etc?
The paths from tweeter to mic will usually be slightly different in length.
This causes extreme HF comb filtering. Our ears do not hear in this manner. The R ear hears little HF from the L speaker and vice versa.
So if you are attempting to create a frequency response curve somewhat related to what we hear, two speakers running is simply wrong.
Similarly ETC is meaningless with two speakers running. Impossible to find individual reflections when there are two sources.

DD

[boggy]

Hm... two speakers firing in same time during measurements can cause HF roll off (I wrote something about this, earlier in thread.)... I didn't read carefully all posts... sorry.

Cheers

Boggy

[Tom Higgins]

Originally Posted by Rairun

Before and after Impulse Responses seem very different, is this normal?

Yes