• 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!

REW measurements done. Please help interpret. Amazing? Horrible?

Snoochers

Active Member
Joined
Mar 28, 2020
Messages
225
Likes
77
Hi all. I did a lot of EQ a couple years ago and I’m wanting to re-examine things. I have an Anthem MRX 1140 and two 21” subs on opposite centres of walls. I have KEF R7s as my LCR. I used Anthem ARC to do some EQ followed by measurements only in REW with a calibrated UMIK-1 microphone. I have a lot of acoustic treatments.

The ARC results have a lot in them so I won’t go over everything here but here is a PDF with settings. This includes corrections up to 20k. They say to leave it to 5k but I pushed it up and things looked a bit smoother. Crossovers and roll-offs etc. are in there. If you have questions I can answer.

I did the moving mic method (MMM) at my MLP. It's close to seat backs behind it but I tried to move around up and down front to back.

L and R have a dip at 60hz but not C which is curious. Seems to be a general bump in the 10-15khz range.

Is this amazing? Is this horrible? I have no idea. What should I be focusing on? How does the general house curve look?

I am happy to change settings and to take more measurements etc.

L and R 20-20k
L+R 20k.jpg

L and R 20-200hz
L+R 200.jpg

L and R combined 20-20kz
L+R combined 20k.jpg

L and R combined 200hz
L+R combined 200.jpg


C - 20-20k
C 20k.jpg

C 20-200hz
C 200.jpg

LFE 20-200
LFE 200.jpg


rt60.jpg
 

Attachments

  • arc results 2 oct 2024.pdf
    1.6 MB · Views: 43
Last edited:
Always include the legend when exporting graphs and make sure to give your measurements meaningful names.

Always use logarithmic frequency scale for full-range response graphs.

For in-room response measurements, do not use more than 1/12 Oct smoothing, or use Var smoothing.

Do not show more than one or two response graphs at the same time.
Too many overlapping graphs and they all become illegible.

For in-room response measurements, make sure that your vertical dB range (max dB minus min dB shown) does not exceed ~50dB.

Please use REW's Light theme for exporting graphs, as it is much easier to read.
You can choose Light mode only for graph exports and keep everything else in Dark mode, if you prefer that.

Use the Moving Microphone Method for in-room response measurements.
Single-Point sweeps are next to useless for judging frequency response.
 
Always include the legend when exporting graphs and make sure to give your measurements meaningful names.

Always use logarithmic frequency scale for full-range response graphs.

For in-room response measurements, do not use more than 1/12 Oct smoothing, or use Var smoothing.

Do not show more than one or two response graphs at the same time.
Too many overlapping graphs and they all become illegible.

For in-room response measurements, make sure that your vertical dB range (max dB minus min dB shown) does not exceed ~50dB.

Please use REW's Light theme for exporting graphs, as it is much easier to read.
You can choose Light mode only for graph exports and keep everything else in Dark mode, if you prefer that.

Use the Moving Microphone Method for in-room response measurements.
Single-Point sweeps are next to useless for judging frequency response.
Thanks for this. OP has been updated. Let me know what you think.
 
Perhaps enter your room’s physical dimensions into REW’s ‘room sim’ feature you can discover the cause of the 60Hz cancellation.
Keith
 
Perhaps enter your room’s physical dimensions into REW’s ‘room sim’ feature you can discover the cause of the 60Hz cancellation.
Keith
I've done it and the null seems to be more in the 30-40hz range which I don't have so much. Also interesting that the centre speaker doesn't have this.
 
But when you virtually drag the speakers/ your listening position ( not forgetting seating height) isn’t there a position where the cancellation disappears?
Keith
 
But when you virtually drag the speakers/ your listening position ( not forgetting seating height) isn’t there a position where the cancellation disappears?
Keith
If I get rid of the subs in the simulation then I can replicate the 60hz pretty well. But I'm not moving them lol.
 
But when you virtually drag the speakers/ your listening position ( not forgetting seating height) isn’t there a position where the cancellation disappears?
Keith
My crossover is at 50hz. Maybe if I move that it will help.
 
Your raw REW graphs looks like a typical untreated room. Not terrible but not great...

I would try to EQ to cut bass humps below 250 Hz, but leave the dips alone (those are likely acoustic nulls due to standing waves).

If you want to affect the nulls, a strategically placed overstuffed sofa may help.
 
Your raw REW graphs looks like a typical untreated room. Not terrible but not great...

I would try to EQ to cut bass humps below 250 Hz, but leave the dips alone (those are likely acoustic nulls due to standing waves).

If you want to affect the nulls, a strategically placed overstuffed sofa may help.
Thanks for this. If that’s true, then I don’t understand. I have considerable treatments. Maybe 25% of my room is covered or more. I have huge hemi cylindrical diffusers, and huge bass traps. I have 4 inches of insulation at a minimum at every point of absorption including first reflections.
 
Thanks for this. OP has been updated. Let me know what you think.
Nice graphs! Thank you for the extra effort.

Your L/R measurements look really nice, except for the treble blip and the unavoidable room modes.

I'd try to diagnose the treble peak using close-mic'd, gated sweep measurements of just the tweeter.
If those show the same peak, then I'd remove that with EQ

I'd also look into room treatment to alleviate the choppiness at 200-1000Hz.

Edit: Plenty of room treatment already installed, I see.
 
Nice graphs! Thank you for the extra effort.

Your L/R measurements look really nice, except for the treble blip and the unavoidable room modes.

I'd try to diagnose the treble peak using close-mic'd, gated sweep measurements of just the tweeter.
If those show the same peak, then I'd remove that with EQ

I'd also look into room treatment to alleviate the choppiness at 200-1000Hz.

Edit: Plenty of room treatment already installed, I see.
Yes perhaps I should've included that in the OP, sorry. Room is plenty treated and I'm probably not going to make changes in that regard, at least not any time soon.
 
Your raw REW graphs looks like a typical untreated room. Not terrible but not great...

I would try to EQ to cut bass humps below 250 Hz, but leave the dips alone (those are likely acoustic nulls due to standing waves).

If you want to affect the nulls, a strategically placed overstuffed sofa may help.

When I looked at that 250ms RT60 for the high frequencies with up to 750ms for 50Hz I thought to myself that this guy has plastered acoustic foam on his walls for sure. I personally think 250ms sounds too dead but the optimal target varies depending on room size and taste.

To the OP: one graph which I do not see is LFE + mains together. Looking at your mains graphs alone does not tell the whole story. Yes, there is that huge dip at 60Hz. What is more striking is how little bass there is, and the asymmetry between left and right speaker (there is a broad +4dB shelf between 65Hz - 130Hz). The reason why can be found in that PDF (I am only showing one side because the other is a similar story). I am also going to compare it to the graph that you posted.

1727997263295.png


1727997992604.png


Notice that neither the raw measurement nor the predicted outcome has the 60Hz dip that you observe in your verification measurement. Either the mic / speakers / furniture were moved between the raw measurement and verification measurement or the DSP on your Anthem did something that produced these artefacts. You said that your verification measurement was done with an MMM. I don't know what you used for your initial measurement but it was likely not an MMM. I am also highly sceptical that the mic was at the same distance to the speaker between initial and verification measurements. If it was, then your raw measurement would show a 60Hz dip.

Also, the bump that you see from 10kHz - 15kHz is missing in the initial measurement and the correction. A bump like that is typical from in situ measurements where the mic picks up reflections from nearby objects. How were you doing the MMM? MMM's should be done with the mic mounted on a pole (you can use a broomstick and duct tape) with the mic pointing at the speaker and waved over the listening area where your head would be. If you hold the mic in your hand in front of you, reflections from your body might produce that bump.

You can also see that your speakers naturally produce a lot more bass. Your Anthem decided it would be best to cut it all off below 70Hz. This is not such a terrible idea if it is able to blend the subwoofers in. So let's look at the subwoofers:

1727997424426.png

1727998824283.png


Now this is very interesting. Notice that what the Anthem thinks it did vs. the measured outcome is very different :D Very strange that you said that your XO point is 50Hz. This is definitely not a 50Hz XO point. The anthem thinks it is crossing over at 150Hz, but your actual measurement shows the subs rolling off at 120Hz or so. The anthem is also boosting bass between 20Hz - 100Hz instead of cutting. This combination of high XO point + bass boost runs a high risk of producing distortion, so I would like to look at the distortion figures.

So we still have information missing. What I would like to see:

- a before/after comparison using the same measurement method. You can do an MMM, just make sure to do the MMM properly. Turn off processing in your Anthem and do an MMM, then turn it on and repeat the MMM.
- a FULL RANGE measurement including the subwoofer and the main speakers. L + LFE1, R + LFE2, and then L + C + R + LFE1 + LFE2 all together.
 
When I looked at that 250ms RT60 for the high frequencies with up to 750ms for 50Hz I thought to myself that this guy has plastered acoustic foam on his walls for sure. I personally think 250ms sounds too dead but the optimal target varies depending on room size and taste.

To the OP: one graph which I do not see is LFE + mains together. Looking at your mains graphs alone does not tell the whole story. Yes, there is that huge dip at 60Hz. What is more striking is how little bass there is, and the asymmetry between left and right speaker (there is a broad +4dB shelf between 65Hz - 130Hz). The reason why can be found in that PDF (I am only showing one side because the other is a similar story). I am also going to compare it to the graph that you posted.

View attachment 396442

View attachment 396445

Notice that neither the raw measurement nor the predicted outcome has the 60Hz dip that you observe in your verification measurement. Either the mic / speakers / furniture were moved between the raw measurement and verification measurement or the DSP on your Anthem did something that produced these artefacts. You said that your verification measurement was done with an MMM. I don't know what you used for your initial measurement but it was likely not an MMM. I am also highly sceptical that the mic was at the same distance to the speaker between initial and verification measurements. If it was, then your raw measurement would show a 60Hz dip.

Also, the bump that you see from 10kHz - 15kHz is missing in the initial measurement and the correction. A bump like that is typical from in situ measurements where the mic picks up reflections from nearby objects. How were you doing the MMM? MMM's should be done with the mic mounted on a pole (you can use a broomstick and duct tape) with the mic pointing at the speaker and waved over the listening area where your head would be. If you hold the mic in your hand in front of you, reflections from your body might produce that bump.

You can also see that your speakers naturally produce a lot more bass. Your Anthem decided it would be best to cut it all off below 70Hz. This is not such a terrible idea if it is able to blend the subwoofers in. So let's look at the subwoofers:

View attachment 396444
View attachment 396451

Now this is very interesting. Notice that what the Anthem thinks it did vs. the measured outcome is very different :D Very strange that you said that your XO point is 50Hz. This is definitely not a 50Hz XO point. The anthem thinks it is crossing over at 150Hz, but your actual measurement shows the subs rolling off at 120Hz or so. The anthem is also boosting bass between 20Hz - 100Hz instead of cutting. This combination of high XO point + bass boost runs a high risk of producing distortion, so I would like to look at the distortion figures.

So we still have information missing. What I would like to see:

- a before/after comparison using the same measurement method. You can do an MMM, just make sure to do the MMM properly. Turn off processing in your Anthem and do an MMM, then turn it on and repeat the MMM.
- a FULL RANGE measurement including the subwoofer and the main speakers. L + LFE1, R + LFE2, and then L + C + R + LFE1 + LFE2 all together.
Thank you for all this. I will do some thinking and report back. You say L+LFE etc. and I don't get it. I don't see that option in REW. I can do L or LFE but not L+LFE so I don't know how to do that. I thought LFE was included by simply having the AVR do crossover and sending the rest to the sub.
 
Thank you for all this. I will do some thinking and report back. You say L+LFE etc. and I don't get it. I don't see that option in REW. I can do L or LFE but not L+LFE so I don't know how to do that. I thought LFE was included by simply having the AVR do crossover and sending the rest to the sub.

It may well be that if you use REW to send "L" to the speaker, your AVR might output "L + LFE". The only person who knows how your system is connected is you ;) If you are not sure, put your hand gently on the sub to see if it is moving when you send a sweep through it. If that graph you are showing is speaker + sub, then I would say that it looks very bass deficient.

I don't like automated DSP solutions, and I especially do not like DSP's in AVR's. It would be nice to have more control.
 
It may well be that if you use REW to send "L" to the speaker, your AVR might output "L + LFE". The only person who knows how your system is connected is you ;) If you are not sure, put your hand gently on the sub to see if it is moving when you send a sweep through it. If that graph you are showing is speaker + sub, then I would say that it looks very bass deficient.

I don't like automated DSP solutions, and I especially do not like DSP's in AVR's. It would be nice to have more control.
I think it's obvious (to me) since the curve is flat down to 20hz? My speakers wouldn't be able to do that on their own.

Also, check this out. Orange is NO EQ. Blue is WITH EQ. Wtf? EQ makes the low bass much better but messes with 60hz big time. It fixes a bump around 150 and a couple swings but doesn't change much otherwise.

Please note that this is with a STATIC mic because I didn't want to move the mic. Still, pretty strange to me.

1728046153204.png
 
Some observations.

1. It is possible that your Anthem is correcting over multiple positions instead of only one position. If this is the case, taking a verification measurement from one position is not a valid comparison.

2. It is very odd that your new measurement with DSP OFF looks different to the measurement curve that your Anthem is correcting. They should be exactly the same.

What was your procedure for taking measurements with the Anthem? Were you using the same mic?
 
Some observations.

1. It is possible that your Anthem is correcting over multiple positions instead of only one position. If this is the case, taking a verification measurement from one position is not a valid comparison.

2. It is very odd that your new measurement with DSP OFF looks different to the measurement curve that your Anthem is correcting. They should be exactly the same.

What was your procedure for taking measurements with the Anthem? Were you using the same mic?
The Antem is correcting over 5 measurement positions which were placed static with the anthem mic. The REW is the movement method +/- 12" around the MLP.

What do you mean with #2? What measurement curve are you referring to? Something in the PDF?
 
What I meant with point no. 2 is that this measurement here in yellow, i.e. with the DSP turned off:

1728103268715.png


looks different to this measurement in red (i.e. the "before" measurement which the Anthem is correcting). Notice the lack of the dip at 60Hz and lack of the bump between 10-15kHz, and the rest of the curves are subtly different:

1728103316365.png

What this means is that you are not comparing the same thing. There are two reasons why this may be so:

1. You are comparing a single point measurement vs. a multi-point averaged measurement,
2. You are using different microphones (Anthem ARC-1 for initial measurement vs. UMIK-1 for verification) and possibly making other mistakes with measurement as well. For example, vertical vs. horizontal orientation, using the mini-tripod and not a proper microphone boom stand, failing to move furniture out of the way, etc.

This also explains why the Anthem may have made that dip at 60Hz look worse after correction - it is optimising correction over a larger seating area.

There a few diagnostics you should consider if you want to move forward.

1. The first thing to do is find out whether the Anthem microphone performs the same as your UMIK-1. Do not assume that two calibrated measurement microphones will give you the same result. Condenser mics are fragile, too many knocks on the head and they will go out of calibration. If Anthem decided to cheap out on their measurement microphone, it may not even measure properly. Using a proper microphone boom stand, take a sweep with the Anthem. Then carefully replace it with the UMIK-1 and repeat the sweep from the exact same position. Compare the two sweeps.

2. Using the Anthem mic alone, perform "before and after" (i.e. DSP on and off) sweeps of the system from the same five points that you used to take the initial measurements with the Anthem. In REW, add all the "before" curves together (Trace Arithmetic - A + B, you will have to repeat the process). Do the same for all the "after" curves. This will give you a better idea of what the Anthem is doing since you are now comparing like with like.
 
Back
Top Bottom