REW measurements - discussion

[insider9]

A place to post and discuss your measurements. Might be useful to have it's own thread for that.

 

[insider9]

I've managed to get some room treatment in. Not DIY and not yet sure about placement nor performance. So I've looked at the impulse response graphs to better understand where the treatment would do the most.

Here's some graphs before:

Right speaker - here

Left speaker - here

Here's after:

Right speaker - here

Left speaker - here

I'm mostly interested in understanding how to work out points of reflections based on these graphs. Time delay = distance? Right?

The most curious fact is the after treatment right speaker graph shows reflection after 18ms that wasn't there before. That's sound traveling 6.2 meters. Indicating the only (at present) not treated wall. How can it be that this was not the issue before treatment?

 

[luckylion100]

sadly I won't be able to contribute anything but I imagine there'll be quite a lot of informative posts by others especially Ellisdj.

Would be great to see some pictures of the treated room when you get a chance.

 

[insider9]

It's far from finished. I've only plonked them in where I flet they could benefit. Also I'm yet to determine what actual frequencies these panels affect. Plus no diffusion so far.

 

[newlash09]

I will be very interested to read and learn REW measurements. Ellisdj had generously calculated my issue with bassboom at 60 Hz. Been very interested ever since.

 

[insider9]

I will be very interested to read and learn REW measurements. Ellisdj had generously calculated my issue with bassboom at 60 Hz. Been very interested ever since.

Feel free to post your own room measurements whatever they may be. It will be a learning excercise for all of us to discuss how different rooms behave and how they react to changes and what can be done to treat them.

 

[ellisdj]

I find analysing the impulse response very difficult - you need to look at the graphs in certain ways.

Matt Khan on AVF is the man to ask this to

 

[insider9]

Thanks Ellisdj

I think I'm getting the hang of it in fairness. I find it very interesting to look at. Will post some thoughts later.

EDIT

So far managed to figure out which are floor and ceiling reflections. Contemplating the others at the minute.

 

[ellisdj]

You can teach me then please if you dont mind

 

[insider9]

For the time being lets look at one speaker only (left) in my setup.

Looking at the impulse response before treatment with comments - here

This is with treatment how it's currently placed - here

And that's how it was when I just placed it by hunch (already linked before) - here

Sadly no way to overlay these in REW. I've managed to get get rid of the reflection at 3.74ms by sliding the treatment about 12 inches to the right. Between the last two grahs already treated room.

I'll explain how I've managed to figure out which reflections are which in the next post.

 

[insider9]

So just looking at this graph - here and how I got the comments.

I took notes of time delay of the reflections. The four annotated ones are as follows:

1.74 ms / 3.76 ms / 5.10 ms / 13.51 ms

I have converted these into distance

0.6 m / 1.28 m / 1.74 m / 4.6 m

These distances are where the sound has to travel further than the direct (non-reflected) impulse.

I then measured the distance from speakers to the mic and added it to the above distances. Getting these values

2.33 m / 3.01 m / 3.47 m / 6.33 m

That's the distance the sound traveled from the speakers to the mic reflected of a surface.

I've also compared the pre/post measurement and reaslised that the two reflections on both speakers remain. Also both speakers had reflections in mostly same places. That indicated that these are front to back and not so much sides.

The two reflections that remain are floor and ceiling. To check which is which I measure the distance from the speaker to the ceiling and ceiling to the mic.

Continued...

 

[insider9]

The other two reflections were not easy to determine. The front wall reflection could've been the back wall as the distances are very similar. With the back wall mostly covered they would disappear completely though. They've not on one side slightly remaining.

This way I knew it's likely to have been the front wall (behind speakers) I have moved one of the panels to confirm this and the reflection was pretty much gone.

The last reflection was a nightmare to determine and I asked my son for that as measuring wasn't the easiest by myself. But we finally got to grips with it and given the symetric nature of it we concluded it's the only plausible explanation. Sound bounicing of back wall then to fron wall and back to the mic.

EDIT

I hope it makes sense. All you need is a measuring tape and a calculator (as long as you rememeber the speed of sound )

 

[insider9]

Also for the eagle eyed you may have noticed the first "spike" is at 0.2 ms super close to original impulse. That's mightily close and accounts for 7cm distance and is a direct reflection from the mic stand. Yes I have moved the mic in it's stand forward (moved the stand back the same amount) and re-took the measurements. It pretty much removed that spike :biggrin:

 

[ellisdj]

How did you determine the time delays?

Also the spikes on the graph you feel are the distances from the mic?

 

[insider9]

The X axis on the impulse response graph is in miliseconds. The main impulse is at 0 ms (X axis) 100% (Y axis). In an ideal world it should be just that and nothing else. All else is reflections, diffractions, etc.

From there you can convert the time delay from original impulse into distance. Then you just need to unleash your inner Sherlock to figure out which is which. Tape measure helps a lot but also differences between measurements when you move panels around.

 

[insider9]

How did you determine the time delays?

Also the spikes on the graph you feel are the distances from the mic?

The mic stand was only a guess as it was the only explanation I could come up with. Since the reflection was in cm. Having measured it it's now fine

 

[insider9]

Here's the newest measurement with one panel on the floor. Treating the first reflection. If only I knew the ceiling could handle the weight.

Also notice much smaller notch at 0.2 ms from moving the mic forward in it's stand.

And that's before and after

 

[ellisdj]

Time in milseconds converted to seconds x 343 = distance in cm.

Then add the distance to the mic to get the reflection distance.

Then work out what it is and try and cover it

 

[ellisdj]

Have you got photos of what you have done?

Also how do you know what freq the reflection is at?

 

[ellisdj]

Have you got photos of what you have done?

Also how do you know what freq the reflection is at?

 

[insider9]

I thought I already posted this. It's likely that not all panels are necessary. Picture link here in case it doesn't show up.

As to your question regarding frequencies you can't get that info from impulse response graph. That's purely to establish where the reflections points are. I've not yet managed to get the treatment to target specific frequencies, only reflections.

Time in milseconds converted to seconds x 343 = distance in cm.

Then add the distance to the mic to get the reflection distance.

Then work out what it is and try and cover it

Yes, distance you will get distance in meters though but the therory is sound (s x m/s = m)

Then add the distance speaker to mic. This is the distance the sound has to travel. Which is the distance you measure

speaker > reflection point > mic

and it's for you to figure out where the reflection takes place. Fun, fun, fun

 

[ellisdj]

You measure from the speaker or mic?

 

[insider9]

From speaker to reflection point and from reflection point to mic. I guess measuring other way round will give the same result. Total distance has to be what you got from converting time delay in impulse response graph.

EDIT
That's for first reflections. You may have a couple reflections that will make it even more interesting. Got one on right speaker i still cannot figure out (18 ms)

 

[ellisdj]

Right so total distance is from speaker to reflection back to the mic?

 

[insider9]

Right so total distance is from speaker to reflection back to the mic?

Yes