Help needed to analyze REW result and decide on treatment

[Kiran NC]

Hi All,

I have taken measurement using REW 5.18 and result file and SPL graph is attache here.

I have used Laptop HDMI port for sending Stereo audio to my Onkyo TX-SR 707 AV Receiver.
Front Left and Right speakers are Polk RTI A5 and Subwoofer is SVS PB12 Plus.
Audessey EQ in ON in AV Receiver and Crossover frequency for Front Left&Right channels are set at 80Hz.
Volume level in AVR was 65

I have used UMIK-1 with calibration file , mic was pointed towards center speaker

I have set the start frequency to 10 Hz and end frequency to 500Hz.

The below graph is without applying any smoothing.

I have tried using my AVR Eq to reduce peaks and dips, but there are very limited frequency options are available in AVR.
I also tried PEQ feature available in Subwoofer, but the max increase possible is +3dB which is not sufficient for the dip at 80Hz.
I am able to use PEQ to reduce the peak at 38Hz.

Please guide on what are the other options to be tried with measurement and also the room treatment options to be considered - like bass traps, absorbents, diffuses etc.

The setup is in living room LxWxH (ft) is 16.5x11x9. It is a 7.1 setup with Polk CSi A6 center and Polk RTi A3 (4 Nos) as surrounds and surround backs.

Currently there is NO room treatment done.

Any guidance/suggestions/pointers will be helpful

Regards,

 

[Wayne A. Pflughaupt]

Welcome to the Forum Kiran!

The problem at 80 Hz – and 65 Hz – are nulls. No amount of EQ will fix them.

Response below 60 Hz is very strange – two separate “plateaus.” Any other location options for the sub?

Treatments primarily reduce excessive reflections (upper frequencies) and ringing (low frequencies). They don’t have a huge effect on frequency response.

Regards,
Wayne

 

[Kiran NC]

Hi Wayne,

Thank you very much for your reply.

Can you suggest the options for removing the nulls?
I am planning to try measuring at other possible seating positions.
I am seeing those nulls even if I change crossover to 40Hz. Audessey selects crossover as 40Hz for all speakers after measurement. At 40Hz crossover, my Left&Right speakers are producing above 40Hz, still those nulls are there.

Attached 40Hz measurement and both in same screen for reference.

I will try to change the position of subwoofer for better results below 60Hz

Regards,
Rajesh

 

[Matthew J Poes]

Rather than playing too much with the subwoofer crossover, I would consider trying adjustments to the subwoofer phase and possibly its distance/delay. I don't expect it to fix things but doing that helps show if the problem is related to that integration. Some tweaking can make a difference.

I agree that the measurements look a little strange, namely the dip peak combination between 60 and 90hz. Let me take a look at your mdat file to see if there are clues. Resonances and such can cause that too and may explain what is going on.

As mentioned, treatments won't typically fix the bass response too much because they can't fully absorb a bass wave that low and stop its interference. At low frequencies your best bet for improving bass is subwoofer placement and eq. The biggest impact will come from multiple subwoofer's placed throughout the room along with eq. You need flexible PEQ.

Adding boost isn't a complete no-no, but generally you want minimal to no overall boost if possible. Boost removes dynamic range from the system by eating up power to boost those frequencies. Narrow nulls are not audible, but a clipping amp is. It is best to remove peaks and not touch nulls. You can shape the EQ curve by stringing multiple PEQ filters together which contain both cut and boost, and I find this ok. The Auto EQ function in REW does this. However, trying to just apply boost to a null is generally not a great idea.

I generally prefer running full frequency response measurements even if you don't intend to treat. That is because A) it gives you all the information you need to decide room treatment, and B) it lets you see the response of the low frequencies relative to the high frequencies and get a sense of the overall response shape. This is a good thing for figuring out bass levels. You generally want a wedge shaped response that is smooth with an increase of 1db per octave from 20khz to 20hz, which puts 20hz about 10db's above 20khz (this depends on the systems actual low end limit, if your subs can't do that, don't do that).

 

[Matthew J Poes]

You should try retaking the measurements at a much lower volume. Shoot for 75db's. I just took a look and your mdat shows both a distortion peak near the nulls and a phase anomalies. Try turning it down to see what is going on. Also, take full range measurements, it will help me see what's going on.

Since you are doing this, try flipping the phase 180 degrees and measuring that way as well. We can compare which looks better. There is a bit of a phase shift from minimum phase in this range, but I can't tell what is going on yet.

did the measurement sound clean at all frequencies or did you notice any distortion, rattling, or resonances at any particular frequencies?

 

[Wayne A. Pflughaupt]

Can you suggest the options for removing the nulls?

I am seeing those nulls even if I change crossover to 40Hz.

If you’re still getting the nulls with the crossover set to 40 Hz, that means it's in the main speakers as well as with the sub (so you have the sub located close to one of the speakers?). If so, moving the speakers may help, but it may just end up moving the null to another frequency. Can’t hurt to try though.

Phase / delay adjustment is certainly worth pursuing, but assuming the sub is closer to one speaker than the other, it may not completely fix the problem.

Fortunately (as Matthew noted), nulls like this typically aren’t terribly audible.

Regards,
Wayne

 

[Kiran NC]

Thank you very much for the responses.

The subwoofer is not placed next to front speakers. I have attached the dimensions and positions of L/R and SW.
Front speakers are around 1.5 feet from front wall.

Since nulls are visible in both cases (crossover at 80 and 40), I believe these two positions are having similar rooms response at these null positions, I will have to try some other positions for sub. I am also planning to take measurements from other seating positions. Currently I am sitting around 1 feet from back wall because my screen size is 120". I will try to take measurement at 2 feet from back wall and see.

Based on the responses, I am planning to try the below options:
1) Reduce the volume level and take measurement
2) Change the phase to 180 and 90 and take measurements
3) Take full range measurement
4) Take measurement from other possible seating positions
5) Move Sub to other possible places and take measurements

Let me know if anything else to be tried

Regards,

 

[Kiran NC]

Hi All,

I have taken new measurements based on the inputs/suggestions.

I have attached the new measurement file and below is the SPL graph

After 1/6 Smoothing

This looks to be better than previous scenario. Looks like null moved to 95Hz (previously there were two nulls below 100 Hz).

The bump at around 38Hz is still there, and PEQ on SW at 40Hz with 2.0 Q value and -10db didn't help (tried higher Q values as well , but not much change).

Out of the options listed in my previous post, only thing left is changing subwoofer position.
Let me try various SW positions next time.

Please let me know the feedback , comments, suggestions

Regards,

 

[Kiran NC]

I switched off the subwoofer and took measurement to find out the SPL level from the front speakers.
I took measurements at crossover of 100Hz and 150Hz in AV Receiver.

I was hoping for a graph with increase in SPL level before the crossover point.

But there are higher SPL levels at some points including 38Hz

What could be the issue?
Is my AVR sending signal to front speakers which are much below crossover point?

 

[Talley]

Are you calibrating your mic SPL before doing the test? You were suggested to run the tests at a lower db but you increased it?

 

[Matthew J Poes]

Something seems amiss. 120db SPL is very loud. Like will do instant hearing damage loud. It seems like something isn't calibrated right.

This 38hz spike seems a bit odd, like a particular source is creating it or its in the electronics of your measurements somehow.

Can you also measure just the subwoofer and no mains?

 

[Kiran NC]

I don't have a SPL meter to calibrate the readings of REW.
Is there any way to properly check/confirm the readings of REW any other way?

As per the REW help, calibration is not required if USB Mic is used.
I am selecting AVR Receiver in preferences->Output device and Master Volume under Output.

I have started measuring at AVR volume 50, but REW said volume is low and I had to increase to 60 (similar to previous readings). I also noticed that SPL value is high compared to last time.

As part of the speaker position changes, I moved Front L/R speaker back, it is around 1 feet from back wall now.
And the Front Left speaker is in the corner now. Audyssey is reducing the level of that speaker after measurement to compensate. (above measurement is after disabling Audyessey)

How do I measure subwoofer alone without mains ? - please guide
Is it by physically disconnecting cable from Front L/R speakers?
or changing the level of front L/R to -12dB and keeping crossover to 200?

 

[Matthew J Poes]

I don't have a SPL meter to calibrate the readings of REW.
Is there any way to properly check/confirm the readings of REW any other way?

As per the REW help, calibration is not required if USB Mic is used.
I am selecting AVR Receiver in preferences->Output device and Master Volume under Output.

I have started measuring at AVR volume 50, but REW said volume is low and I had to increase to 60 (similar to previous readings). I also noticed that SPL value is high compared to last time.

As part of the speaker position changes, I moved Front L/R speaker back, it is around 1 feet from back wall now.
And the Front Left speaker is in the corner now. Audyssey is reducing the level of that speaker after measurement to compensate. (above measurement is after disabling Audyessey)

How do I measure subwoofer alone without mains ? - please guide
Is it by physically disconnecting cable from Front L/R speakers?
or changing the level of front L/R to -12dB and keeping crossover to 200?

Did you load a correction file? The sensitivity that is used to set the SPL level comes from that file I believe.

I've had situations where the USB mic is recognized but the SPL level isn't set or calibrated. One way to check is to click on the SPL button in REW and then Calibrate. If it lets you calibrate, it's not automatically calibrating it.

Do you have a smartphone? iPhone? If so there are SPL meter apps that range from very accurate to good enough for this.

 

[Kiran NC]

Something was wrong last time due to which high SPL was measured.

I tried measuring again, this time I got proper readings at Volume 55.

I installed SPLnFFT app in my iPhone and compared the reading in the phone with SPL reading in REW.
REW was not allowing to calibrate, got a message that Calibration is not required.
For 1KHz test tone, I was getting SPL 80 in REW as well as my in my iphone.
Below picture shows reading in REW

Below picture shows reading in iPhone - SPLnFFT app

First I measured with Audyssey OFF, Volume 55, Crossover at 80, Mic 90 degrees.
Below pictures shows the SPL reading

I enabled Audyssey, and took the measurement again.
Below picture shows SPL with Audyssey ON

Below picture shows Audyssey OFF vs ON

Audyssey is making a null at 38Hz, may be to compensate the peak it measured. Level of 15 to 35Hz also increased to matching levels.

In both cases there is a shallow deep from around 150Hz to 200Hz and at around 260Hz.

I have attached the measurement files as well.

Please go through and suggest any changes which can be done to further improve

 

[Matthew J Poes]

I'll take a look in more detail later but it looks like Audyssey did quite a bit to improve things. I'm impressed as I don't normally find that big an improvement. In my own system Audyssey on an Onkyo receiver tends to make things worse even.

 

[Wayne A. Pflughaupt]

It would be easier to analyze the graphs if you applied 1/3-octave smoothing.

Regards,
Wayne

 

[Kiran NC]

Please find the SPL graph after applying 1/3-octave smoothing

 

[Matthew J Poes]

Ok I've taken a look at the REW file and I think I have a good sense of what is going on. I need to know more about your room. Having recently measured a large post and beam open living room/dining room with no absorption other than a leather couch and thin area rug, I can now say you take the cake for most reflective room I've ever seen in a domestic space. That room was supposed to be my example of poor RT60 in a domestic space but turned out to have a nice flat RT60 in the .5-.6 range. Yours has a rising RT60 response as you go down in frequency that surpasses .6 seconds (600ms) by 300hz. What you want is flat and at or below .5 seconds if possible. This is normally not a big deal as most living rooms would have this quality. For some reason, your room is very reflective What this is doing is causing a few problems. First, your bass doesn't decay over a normal time frame. That will cause boominess, even though the response is flat. Second, it's causing a lot of late energy which impacts your clarify scores. I can't say those are super meaningful, but I like to look at them anyway. My opinion is that when you have too low of a ratio of early to late energy you end up with poor dialogue clarity and it becomes harder to hear voices in movies. With music it causes the instruments to be less distinct, you don't hear with as much resolution.

The fix is acoustic treatment. You need absorption at all frequencies, so broadband. There are lots of options, but I need to know more about your space. What is the wall construction, décor, room use, etc. All that helps to decide the best options. What you need more than anything is bass traps, but there are decay issues well up to 1khz, which is why I am suggesting broadband absorption.

Otherwise things don't look too bad. The response is relatively flat, distortion plots are very clean and relatively low.

 

[Matthew J Poes]

I'll mention the bass distortion levels below about 80hz are a little high for the level, but I'm not sure they would be audible. I compared them against some other measurements I had on hand and they do seem on the high end. At around 82db's at 50hz, I came up with .9% THD for you and .3% for me, as an example. The 2nd and 3rd harmonic are above 1% at that frequency, and 5th harmonic are .5%. For my own measurements I'm at .1-.15% at those frequencies. My only concern would be that if at 80db's you are seeing 1% or more, which is on the edge of audibility potentially, you might see very audible LF distortion at higher volume levels. I'm not totally sure why they are that high, taking measurements at about 6-10db's louder might help see. If they don't go up, it might be a measurement artifact. If they do go up, and they should, then it may be the subwoofer.

If it doesn't bother you, don't worry too much about it.