• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

Seeking advice on analysis of REW measurement

feicipet

New Member
Joined
Nov 13, 2022
Messages
4
Likes
0
Hi,

I had recently started to analyze my room acoustics using REW and a MiniDSP UMIK microphone. I would like to get some advice on what I need to do to fix some of the more noticeable peaks and troughs in my room frequency responses. This is my first time measuring my room acoustics so I am very much a newbie at this. Thank you very much for any advice given.

My room is a slight L-shape with a width of 3.25-4.25m / 10.7-13.9 ft (depending on which end of the room I'm at) and depth of 5.5m / 18 ft. Height of the room is 2.54m / 8.3 ft. All walls are made with concrete while the ceiling is made with a dry wall material (fake ceiling).

Room Measurements.png


My studio desk is placed at the narrower end of the room, I've attached a picture of it as follows:

Picture of Studio Desk.jpg


Some notable facets of my room layout:

  1. The acoustic panels that you see in the pictures are self made panels using wooden frames and Rockwool Safe n Silent Pro 350 slabs (90mm thick, 600mm width, 1200mm length). I have 2 of them at the front of the room and 4-5 more running along both the left and right sides of my sitting position.
  2. There are windows to the right of my sitting position but I've covered them up with the acoustic panels
  3. There's also another window at the far end of the room behind my sitting position, I've left them uncovered except for a bamboo curtain
  4. I do not have any ceiling cloud acoustic treatment currently
  5. For bass trapping, I had simply bought 2 packs of Rockwool 350 (4 slabs each, creating a thickness of 320mm each pack) and threw them behind my studio monitors at the left and right corners in front of my sitting position (they're visible in the room picture above) - I did this just recently to test out how bass traps would improve acoustics in my room, if they work out I will build more permanent structures at the corners
  6. I have an Ikea shelf rack of guitar amp heads right in the middle of my room behind my sitting position. They're not that densely arranged such that they create an artificial wall in the middle of my room, but I'm not sure if the amp heads are enough to create very significant reflections from my monitors. Here's a picture taken from my sitting position facing the back of the room.
Picture of amp rack.jpg


In terms of measuring the room, I had done two rounds of measurement in REW with the following parameters:
  1. My listening position is slightly pushed back from my sitting position at my desk so as to create the equilateral triangle position (this is cos I had positioned my speakers a bit to the front in order to place them in front of my computer screens)
  2. Each side of the equilateral triangle is 1.5m.
  3. From my listening position, I took 6 REW measurements:
    1. Right at the listening position (labelled "center" in my measurements)
    2. About 10cm to the right, left, front and back (4 positions labelled "right", "left", "back" and "front" respectively)
    3. And finally one additional position to simulate myself sitting right at my editing position at the desk (not equilateral, labelled "lean", as in lean forward)
  4. During my first measurement, the average value in the All SPL window showed this:
    Screen Shot 2565-11-06 at 14.36.19 - Average All SPL on 6-Nov.png
  5. This graph showed a pretty significant trough and peak between 60 - 120 Hz. All graphs are set to Var Smoothing.
  6. At this point of time, I already had all of my acoustic panels installed but I had not yet put in the makeshift bass traps.
  7. I didn't know better on what to do so I just ordered 2 more packs of Rockwool and then threw them behind the speakers to test and see how things would change.
  8. So I did this yesterday and the measurement I took today shows this:
    Screen Shot 2565-11-13 at 11.43.56 - Average All SPL on 13-Nov.png
  9. If I overlayed the 2 average lines, it would look like this (Green is before the bass traps, Orange is after):
    Screen Shot 2565-11-13 at 11.50.19 - 6-Nov and 13-Nov Average All SPL Overlay.png
  10. So I guess the effect of adding the bass traps were:
    1. It remedied some sort of phase cancellation at the 20-30 Hz range.
    2. It somewhat stabilized the trough between 70-100 Hz range.
    3. But in general the trough and spike between 60-120 Hz range is still there.
  11. I've uploaded the REW MDAT files (named "Nov 6 vs Nov 13.mdat"). This file contains both sets of measurements made on 6-Nov and 13-Nov. The file is too large for me to upload into the forums directly but I've uploaded into my Google Drive account here: https://drive.google.com/file/d/1Um_NlzovIPqXP_Vvv7L6axB-HwO0R4R6/view?usp=share_link
  12. In case it's difficult to visualize how I've described my room, I did a 360 degree picture of my room and made it available here: https://photos.app.goo.gl/FbPKguDNnBtXDxPL9
Would very much appreciate it if someone could give some advice or pointers on what I should do next to improve my room acoustics.

Thank you!
Wong
 

ppataki

Major Contributor
Joined
Aug 7, 2019
Messages
1,216
Likes
1,355
Location
Budapest
Other people here will advise against this but I would just simply try to fill that gap between 60 and 120Hz forcibly with EQ and test how it sounds
It might get horrible due to increased ringing or it might not - nothing to lose if you try it
If it sounds horrible you can also try not to fill the gap completely but gradually, let's say in 3dB increments

The peaks you can easily flatten with EQ though (and that will decrease ringing at those frequencies too)
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
It seems that your early reflections are pretty messed up, FDW of 5 cycles shows a huge gap in high frequencies.

1668337072813.png


I have never seen something like this with decent speakers. I would pull that up and bring the 100 to 200Hz region down.
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
Oh you're measuring both speakers at the same time, that's probably the reason behind the big cancelation, don't do that.

Try to get a single measurement for both L and R respectively exactly at the listening position.
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
3,106
Likes
2,313
Location
Canada
Oh you're measuring both speakers at the same time, that's probably the reason behind the big cancelation, don't do that.

Try to get a single measurement for both L and R respectively exactly at the listening position.

I don't think that's the (only) possible big issue there:

1668338935188.png


I'd endeavor to find a better speaker and listener position.
 
OP
F

feicipet

New Member
Joined
Nov 13, 2022
Messages
4
Likes
0
Oh you're measuring both speakers at the same time, that's probably the reason behind the big cancelation, don't do that.

Try to get a single measurement for both L and R respectively exactly at the listening position.
Ah, thank you for that.

I've done a single measurement of L and R speakers each. I'm afraid I really didn't understand your comments about the EQ filters yet and I'm trying to read up on that first, but I've uploaded the separate L / R measurements below here in the meantime.

Thanks for the advice!
Wong
 

Attachments

  • Nov 13 - L R Separate.mdat.zip
    2.5 MB · Views: 61

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
Ah, thank you for that.

I've done a single measurement of L and R speakers each. I'm afraid I really didn't understand your comments about the EQ filters yet and I'm trying to read up on that first, but I've uploaded the separate L / R measurements below here in the meantime.

Thanks for the advice!
Wong

Okay the cancellation is gone as expected, try to experiment with the filters i provide below for the 100Hz to 200Hz region, determine Gain value based on how neutral sounding pink noise is. If you don't know how to apply equalization let me know, if you're finding it difficult to follow up with what i'm saying let me know.

1668340050336.png


Try the filter below

1668340159830.png
 
OP
F

feicipet

New Member
Joined
Nov 13, 2022
Messages
4
Likes
0
Okay the cancellation is gone as expected, try to experiment with the filters i provide below for the 100Hz to 200Hz region, determine Gain value based on how neutral sounding pink noise is. If you don't know how to apply equalization let me know, if you're finding it difficult to follow up with what i'm saying let me know.
Thanks. There are a few things that I would like to ask for clarity:
  1. You said to apply the EQ settings. From what I had watched of a Julian Krause video, it seems that the EQ window in REW would only help to generate an EQ file that I need to load up into Equalizer APO, which seems to be a global system sound EQ tool. If I'm just trying to test out the EQ filters, is it possible for the EQ filters defined in REW to be applied to the measurement sweep tone in order to test the results directly as part of iterative measuring?
  2. If I need to use a separate EQ software: The Equalizer APO app is only for Windows. Is there any similar software that I can use for Macs? I know that 3rd party apps such as Rogue Amoeba's SoundSource has an equalizer built-in and can also load in AU plugins, but I'm not sure whether they can be used with the EQ file generated by REW.
  3. Is there anything specific that I should be setting the bunch of options in the EQ window ie
    1668341867517.png
  4. When setting the EQ, is the objective to bring up both the obvious troughs and peaks to roughly align with the rest of the curve? If the troughs are occurring because of phase cancellation from reflections, wouldn't EQ'g those frequencies just increase the volume of those reflections which end up cancelling each other out again when they meet?
  5. How do I get that FDW view that you've been showing me?
Apologies for the newbie questions. I've been trying to watch and read help content as much as time allows, but at this point of time, a lot of the content is still going over my head while I'm trying to absorb as much as I can.

Thanks
Wong
 
OP
F

feicipet

New Member
Joined
Nov 13, 2022
Messages
4
Likes
0
I don't think that's the (only) possible big issue there:

View attachment 243084

I'd endeavor to find a better speaker and listener position.
Thanks for the response. That would be an endeavor that takes a fair amount of time, given how cluttered my room is and the fact that my wife has given me strict instructions to keep the clutter within my room and not move any of my shit outside.

Still, can I ask how do I interpret the spectrogram to arrive at the conclusion you've stated above? That would help a lot in me understanding my problem.

Thanks!
Wong
 

abdo123

Master Contributor
Forum Donor
Joined
Nov 15, 2020
Messages
7,423
Likes
7,940
Location
Brussels, Belgium
Thanks. There are a few things that I would like to ask for clarity:
  1. You said to apply the EQ settings. From what I had watched of a Julian Krause video, it seems that the EQ window in REW would only help to generate an EQ file that I need to load up into Equalizer APO, which seems to be a global system sound EQ tool. If I'm just trying to test out the EQ filters, is it possible for the EQ filters defined in REW to be applied to the measurement sweep tone in order to test the results directly as part of iterative measuring?
  2. If I need to use a separate EQ software: The Equalizer APO app is only for Windows. Is there any similar software that I can use for Macs? I know that 3rd party apps such as Rogue Amoeba's SoundSource has an equalizer built-in and can also load in AU plugins, but I'm not sure whether they can be used with the EQ file generated by REW.
  3. Is there anything specific that I should be setting the bunch of options in the EQ window ie
    View attachment 243096
  4. When setting the EQ, is the objective to bring up both the obvious troughs and peaks to roughly align with the rest of the curve? If the troughs are occurring because of phase cancellation from reflections, wouldn't EQ'g those frequencies just increase the volume of those reflections which end up cancelling each other out again when they meet?
  5. How do I get that FDW view that you've been showing me?
Apologies for the newbie questions. I've been trying to watch and read help content as much as time allows, but at this point of time, a lot of the content is still going over my head while I'm trying to absorb as much as I can.

Thanks
Wong

1) your understanding is very much correct. and yes with APO you can test your improvements and remeasure. However, and i highly recommend this, for any correction above 100Hz use your ears and pink noise. What we measure (steady state response) is not really what we hear in the 100Hz to 20KHz range, frequency dependent windowing narrows it down but it's very important to check if the filters improve things by ear. Below 100Hz the steady state response is fine.

2) Check out this thread https://www.audiosciencereview.com/...ows-linux-macos-ios-ipados-and-android.18450/

3) feel free to explore the options on your free time, my target curve of choice (for steady state correction below 100Hz) is the following

1668343128199.png


4) you're absolutely correct, which why it is important to keep corrections in general to negative gain and low Q and always double check with pink noise that you're not actually doing more harm than good.

5)

1668343251385.png

If you want to understand what Frequency dependent windowing actually does, and learn in general about effective room correction check out this video:
 

ernestcarl

Major Contributor
Joined
Sep 4, 2019
Messages
3,106
Likes
2,313
Location
Canada
Still, can I ask how do I interpret the spectrogram to arrive at the conclusion you've stated above? That would help a lot in me understanding my problem.

No need to look at the wavelet spectrogram to arrive at that conclusion. However, the wavelet view does make it easy to see for any discontinuity and/or time "smearing" in the transient performance of the system which is most often caused by the room and not the speaker itself.

One can make reasonable predictions as to the outcome of whatever your desired EQ is by clicking the ff. button in REW:

1668345726442.png

*disable any manual windowing before generating new measurement


1668345735355.png


And some more decay views...
1668345760939.png 1668345764061.png 1668345769399.png 1668345773078.png


Room treatment is not my forte so can't give much more advice than mere generalities, although, it may be benefit you to watch some of Jesco's (Acoustic Insider) videos over your spare time e.g.

 

Sokel

Master Contributor
Joined
Sep 8, 2021
Messages
5,840
Likes
5,772
Don't want to waste a thread so I'll ask here:
Is there any way to export REW EQ settings so can be integrated with Mathaudio?
The later uses .snr files to save presets but opening them with notepad,etc doesn't seems to be configurable as is not like a normal text.
 
Top Bottom