Home Theater Forum and Systems banner

1 - 15 of 15 Posts

·
Registered
Joined
·
13 Posts
Discussion Starter #1
Hello,

I am in need of help! I am currently in the process of treating my room to utilize my new Focal Alphas, and I am new to REW.

First, I am having some difficulty calibrating my REW. For some reason, when I try to calibrate REW's SPL meter, using a behringer ecm8000, the spl rating is not matching up with my hand-held SPL meter set to Slow, C-weight. Do I need to be using different settings on the REW SPL meter since I am using a mic and not a SPL? Due to this, I have gotten some extremely wonky results, especially since my monitors have been calibrated to 75dB with pink noise.

Additionally, I was told that when actually doing the measurements, it should utilize both speakers. For some reason it is only letting me choose one at a time. Will this affect my results, or do I just take an average between left and right?

Finally, once I have my measurements, would anyone be willing to help interpret my results, or enlighten me by way of tutorial or other means to analyze it myself? I have had some pretty accentuated bass frequencies in my room when I was using different studio monitors. Now that I have some beefy new Focals, I am in desperate need to figure how much more treatment I need to purchase or how much I need to utilize the built in lowpass filters.

Any help is greatly appreciated!
 

·
Banned
Joined
·
4,838 Posts
Hello,

I am in need of help! I am currently in the process of treating my room to utilize my new Focal Alphas, and I am new to REW.

First, I am having some difficulty calibrating my REW. For some reason, when I try to calibrate REW's SPL meter, using a behringer ecm8000, the spl rating is not matching up with my hand-held SPL meter set to Slow, C-weight. Do I need to be using different settings on the REW SPL meter since I am using a mic and not a SPL? Due to this, I have gotten some extremely wonky results, especially since my monitors have been calibrated to 75dB with pink noise.
Not quite surie what you mean. With a SPL meter as your reference, you are telling REW what the SPL is for a given level of input from your mic for a given gain setting on the mic preamt of your audio interface. Once done, as long as you do not change that gain, the result should match exactly. What is the "difference" you are seeing?

Additionally, I was told that when actually doing the measurements, it should utilize both speakers. For some reason it is only letting me choose one at a time. Will this affect my results, or do I just take an average between left and right?
You were told incorrectly. Meaningful results require that speakers be measured one at a time. The human psychoacoustical brain does the "separating" of signals from different directions as we listen, a microphone cannot and at high frequencies gives eratic - and useless - and misleading - results due to phase cancellations at high frequencies.

Finally, once I have my measurements, would anyone be willing to help interpret my results, or enlighten me by way of tutorial or other means to analyze it myself? I have had some pretty accentuated bass frequencies in my room when I was using different studio monitors. Now that I have some beefy new Focals, I am in desperate need to figure how much more treatment I need to purchase or how much I need to utilize the built in lowpass filters.
Post your measurement .mdat file and we'll take a look.
 

·
Super Moderator
Joined
·
9,252 Posts

Welcome to the Forum, Roughroom!

You can’t calibrate REW’s SPL meter with a mic directly. You must have a calibration file for the mic (loaded under Preferences -> Mic/Meter) that includes a sensitivity figure. Otherwise you should just use your hand-held SPL meter to do it, a shown in Section 8 of the REW Help Files.

As for measuring with both speakers or one, that depends on the measurement signal and your intent for measuring. As AudiocRaver mentioned, using the sine wave sweep with both speakers can register a droop in the graph at high frequencies, caused by comb filtering resulting from the mic not being perfectly centered between the speakers. For full range measurements, pink noise is a better signal source for measuring with both speakers.

However, since you’re primarily interested in the lower frequency range, that won’t be an issue for you. If you want to measure both speakers simultaneously, just split the analog output of your sound card to the inputs of your speakers.

You didn’t go into details on what you hope to accomplish with your treatments, but if it’s for the low frequencies their main benefit is reducing ringing (aka low frequency decay times), which can make the bass sound tighter. However, if you have large low frequency peaks or troughs you’re trying to address, it takes a lot of treatment to minimize those, typically more than most people want in a room.

Regards,
Wayne
 

·
Banned
Joined
·
4,838 Posts
A teeny, tiny, minor correction Wayne. I believe the ECM8000 is an analog mic, so the calibration file will not have a sensitivity number, since it has to run through an audio interface.

Wayne P is correct, for subwoofers you can - and should - run the signal through all subs together for LF measurements. Above sub frequencies, I recommend running the speaker measurements separately.
 

·
Super Moderator
Joined
·
9,252 Posts

A teeny, tiny, minor correction Wayne. I believe the ECM8000 is an analog mic, so the calibration file will not have a sensitivity number, since it has to run through an audio interface.
Thanks for that – somehow I had not made the connection that the sensitivity was only given with USB mics. :T

Regards,
Wayne

 

·
Registered
Joined
·
13 Posts
Discussion Starter #6
Thanks for the replies!

The main confusion about the SPL calibration was that I thought it was asking for me to put in the value being registered by my mic, not my SPL meter. Luckily, I just put in my SPL meter's output.

Hopefully the upload worked. I have been having some issues with my internet lately.

In regards to the room that I am in, it is a home studio. Luckily I do not do any live recording. All of my work is entirely within the computer, so I need to be confident that any changes I make to a mix are not caused by my speakers/room. I was ecstatic to find wholesale acoustic foam/treatment, so I have no qualms about the room looking like a padded cell >:)

Thank you again for the help.
 

Attachments

·
Banned
Joined
·
4,838 Posts
Your soundcard calibration file is very noisy, that should not be the case, and has a steady downward slant, that should also not be the case.(it gives your measurements the illusion of rising response at high frequencies)

The L and R impulse diagrams show pretty good reflection control, except within the first 1.5 sec or so. Surfaces close to the speakers &/or listening position are causing some irregularities that could affect imaging.

Freq Response:
  • vertical settings should cover a 60 dB range
  • measurements need not be done nearly so loud, use 75 or 80 dB level for sweeps
  • don't know what curve 1 is, "soundcard?"
  • do not try to fix freq resp before addressing other questions above, which affect the FR measurement accuracy
  • the dips at 60 to 80 hz will be hard to get rid of, the rest should be easier; the 60 to 80 hz dips are probably from reflection cancellations, getting rid of the reflection at that frequency is really tough
 

·
Registered
Joined
·
13 Posts
Discussion Starter #8 (Edited)
Thank you.

In regards to your questions:

1) I have no clue what to do about the soundcard calibration. I have tried using both inputs and outputs. I am curious if it is being caused by the drivers on it. It is an older Focusrite2i4 which is not fully supported on Win10, only works with beta drivers.

2) Unfortunately, the speakers are close to the wall, 11" to be exact. I am unable to preposition them. Would this be the cause of the imaging irregularities?

3) From my understanding, I thought I had the sweeps being produced at 75dB, at least that is what my external meter was showing. Since I know the meter is calibrated correctly, I most likely am ignorant to something within REW. I followed several tutorials while completing this process.

4) Curve1 was the measurement it produced while my soundcard loop was still connected. It showed up once I completed my calibration.

Finally, these measurements were taken while utilizing the Focal's built in low pass filter due to their position near the wall. It attenuates anything below 300Hz.

How should I proceed?
 

·
Premium Member
Joined
·
6,711 Posts
The soundcard cal file is wrong, you'll need to clear in the Soundcard preferences that so it isn't used for any further measurements. You can remove it from existing measurements using the Change Cal button. Perhaps the loopback connection was on the wrong channel, but the measurements look OK so it isn't a big deal to just get rid of the cal and carry on.

The idea of calibrating REW's SPL meter is to tell REW the real SPL, as read from some external source (i.e. your own meter) so that REW can offset its own meter to read the same. After you've done it the REW meter and your external meter should be showing the same levels.
 

·
Registered
Joined
·
13 Posts
Discussion Starter #10
Hello,

I recently put up the last of my room treatment. Here are the results. After the problems I have had with room eq, I decided not to add a calibration file because my most recent ones are no better than the original.

What did change:

1) Additional Room treatment at the first reflection points on the sides of the wall, above the listening position, and directly behind the listening position. All absorption. I have some additional treatment spread out along the side walls.

2) Studio Monitors calibrated at 67.5dB. For some reason it seems that the measurement was recorded at a much higher amplitude than my monitors produced.

3) Measurement sweep was done at -20dB instead of -12. For some reason it was clipping at -12. Nothing has changed in my setup, so I figured it was not going to be the end of the world.

I am by no means fluent in analyzing this data, but from what I am looking at, there are dips at 60-80dB, 600-800dB, and 2k-5k? How much of this could be caused by the output level of my monitors? If this is just a reflection of my room, or the less than desirable placement of my monitors that I cannot fix, then I guess it is not worth adding additional treatment. (Although I have space for more basstraps along the wall my monitors are placed.) Furthermore, would diffusion be of value, or do you feel that I have hit the maximum amount of treatment before I begin to just chase my tail? Lastly, are these fluctuations are enough to warrant a corrective eq?
 

Attachments

·
Banned
Joined
·
4,838 Posts
Hello,

I recently put up the last of my room treatment. Here are the results. After the problems I have had with room eq, I decided not to add a calibration file because my most recent ones are no better than the original.

What did change:

1) Additional Room treatment at the first reflection points on the sides of the wall, above the listening position, and directly behind the listening position. All absorption. I have some additional treatment spread out along the side walls.

2) Studio Monitors calibrated at 67.5dB. For some reason it seems that the measurement was recorded at a much higher amplitude than my monitors produced.

3) Measurement sweep was done at -20dB instead of -12. For some reason it was clipping at -12. Nothing has changed in my setup, so I figured it was not going to be the end of the world.

I am by no means fluent in analyzing this data, but from what I am looking at, there are dips at 60-80dB, 600-800dB, and 2k-5k? How much of this could be caused by the output level of my monitors? If this is just a reflection of my room, or the less than desirable placement of my monitors that I cannot fix, then I guess it is not worth adding additional treatment. (Although I have space for more basstraps along the wall my monitors are placed.) Furthermore, would diffusion be of value, or do you feel that I have hit the maximum amount of treatment before I begin to just chase my tail? Lastly, are these fluctuations are enough to warrant a corrective eq?
Your RT60 is getting very low through mid frequencies, I would not add more absorption.

The dips at 60-80 will only become horrible monsters if you try to boost them with EQ. Speaker/LP position change &/or truck-sized traps are the only fix for them. Guessing you have neither option, plus traps cannot just go where there is room, the cause must be determined and addressed in the right area of the room. (GIK would love to build you some of those truck-sized traps;)

Cause of the 620 Hz dips looks like reflection/cancellation, the reflected path length is about 0.9 ft longer than the direct path length, maybe your desk top? Or front wall, as mentioned, although that would be too long a reflected path. Look at your work surface. Only a guess.

You could put a parametric EQ stage on the 110 Hz peak (6 to 8 dB, same applied to both channels), on the 1.2 kHz peak (2 to 3 dB, same applied to both channels), and shelf down the broad area from 4k to 10k by 2 to 3 dB (same applied to both channels), and it would sound pretty good, I'll bet.

Edit: The 55 Hz difference is best left alone. They average out quite nicely.

Except for the 620 Hz, forget about fixing dips and focus on fixing peaks.
 

·
Registered
Joined
·
13 Posts
Discussion Starter #12
Thanks for the help.

Should I remove some absorption from my room, in comparison to the first measurement? If it isnt worth messing with it anymore, I am more than happy to call it quits, as it seems that my listening position is more of the problem at this point. If it would be best to remove some absorption, where would you suggest to remove it first? What I have added is placed above, behind, and to the sides. Would it be worth swapping some of my absorption for diffusion?

To the EQ:

Based on the levels of the spikes, are they drastic enough to warrant the suggested eq to monitor at the flattest response? I am currently away from my exact computer at the moment, so if I should take the time to adjust my templates with an eq on my stereo out bus for when I am working, I will. To double check if the eq choices match that of your suggestion, I have attached a picture with what I believe to be your suggested settings. Should I use sharper q's to isolate those specific frequencies? Since I have never used a corrective eq for monitoring purposes, I assume I turn these off for my mix-down.

Finally, how am I able to test my room with these eq settings, and is there a way to utilize these eq settings when I am not in my DAW, this way I can have a flatter response without having to be within a specific project?
 

Attachments

·
Registered
Joined
·
13 Posts
Discussion Starter #13
Also for my sanity, why are my readings so high, in comparison to what my monitors are tuned to? Where do I find my reference point, looking at the graph, to determine what is a drastic peak/dip?
 

·
Super Moderator
Joined
·
9,252 Posts

Based on the levels of the spikes, are they drastic enough to warrant the suggested eq to monitor at the flattest response?

Should I use sharper q's to isolate those specific frequencies?
The main problem I see is that everything above 1 kHz is too hot. I’d also recommend a broad shelving filter to bring it down.


I have attached a picture with what I believe to be your suggested settings.
No, you do NOT want to mismatch filters above ~3-400 Hz for the two speakers. It’s fine to do that below that point, but above it, it will do strange things to the imagine.


Since I have never used a corrective eq for monitoring purposes, I assume I turn these off for my mix-down.
Absolutely not. You are correcting your speaker’s EQ, you are attempting to smooth response as much as possible, with the goal being increased accuracy. You don’t want to negate that for your mixing – otherwise what’s the point of all this?


Finally, how am I able to test my room with these eq settings, and is there a way to utilize these eq settings when I am not in my DAW, this way I can have a flatter response without having to be within a specific project?
Can’t help you there – I have outboard EQ for my computer speakers.


Also for my sanity, why are my readings so high, in comparison to what my monitors are tuned to?
You probably did not do the Calibrate SPL routine.


Where do I find my reference point, looking at the graph, to determine what is a drastic peak/dip?
We’re typically looking for broad or sever trends. I’d say a peak or trough more than 5-6 dB, or a smaller one that’s very broad, should be looked at.

Regards,
Wayne

 

·
Banned
Joined
·
4,838 Posts
Should I remove some absorption from my room, in comparison to the first measurement? If it isnt worth messing with it anymore, I am more than happy to call it quits, as it seems that my listening position is more of the problem at this point. If it would be best to remove some absorption, where would you suggest to remove it first? What I have added is placed above, behind, and to the sides. Would it be worth swapping some of my absorption for diffusion?
I would probably leave it alone. The thing about going from absorption to diffusion is that it can end up diffusing the imaging of the system unless it is approached with care. If you like your imaging, and the overall sound of the room, then leave it alone. If you WANT it more live, then start moving to diffusion, but start at points farthest from the LP - in terms of the path from speaker to LP - most likely the rear wall, and carefully move forward from there. Again, unless you feel the need to, I would leave it alone.

To the EQ:

Based on the levels of the spikes, are they drastic enough to warrant the suggested eq to monitor at the flattest response? I am currently away from my exact computer at the moment, so if I should take the time to adjust my templates with an eq on my stereo out bus for when I am working, I will. To double check if the eq choices match that of your suggestion, I have attached a picture with what I believe to be your suggested settings. Should I use sharper q's to isolate those specific frequencies? Since I have never used a corrective eq for monitoring purposes, I assume I turn these off for my mix-down.
As Wayne P suggests, less EQ is usually better than more. The fact you are EQing for your DAW complicates things. I had this problem a few years ago and it led to a second system for speaker EQ, which also served as my music server, and it was built from existing hardware lying around, so it was almost free. Not so for your situation, I am guessing.

Changing all your templates is a BIG pain, I tried that, too, and ended up modifying the thing every time I started a new project. If your DAW allows the use of a Sub bus, you can set it/them up as the monitor EQ point. Another possibility is use the MathAudio plugin, but it is not free. Delays are very small with that plugin.

In my opinion, you can live without the EQ for tracking, and you want it most for mixdown, when longer latency delays are not as disruptive. You can almost run raw for tracking with your response, it is not all that bad. Then add the EQ for mixdown.

The big peak at 110 Hz is the one I would do something about for tracking, and you might be able to do that with one of the Win utilities that takes place of the Windows sound mixer, like the Equalizer APO, or the mixer from VB audio (free or donationware, I believe). If you only need a couple of bands, that might be a cheap, effective, and low-latency way to get there. I have not used them myself, but I believe they would do what you want.

Before you start to EQ, always look at the measurements without smoothing so you can see exactly what needs to be attacked.

Also, I always think of frequencies above 10 kHz as gravy, if they are there, great, if not, let 'em go, don't try to EQ them back up.

Finally, how am I able to test my room with these eq settings, and is there a way to utilize these eq settings when I am not in my DAW, this way I can have a flatter response without having to be within a specific project?
See above.
 
1 - 15 of 15 Posts
Top