First REW Measurement - Room Mode at 40Hz and SMSL A300 Inverted polarity

[Kix1991]

Hello, was hoping to get an opinion from some of the pros on my first REW measurement on an untreated room in a 2.1 system (Acoustic Energy AE1 40th Anniversary and SVS SB1000 Pro).
I've done the test on a Behringer ECM8000 and Scarlett Solo 4th Gen, and turned off all EQ/AMP adjustments, with a low pass filter on the sub at 60hz/12dB slope.

Could I also ask if the impulse / filtered Impulse IR charts look correct? Reason being when I first tested I noticed it was inverted and went to -100% at 0ms. After testing it with another mic and speakers I determined it was actually my amp which I'd been using for about 2 years now with my old pair of speakers, and only realized now after using REW and trying to tone in my new speakers that it looked inverted and swapping my speaker wire polarities for both speakers brought the IR to 100% at 0ms. Just asking to make sure I didn't prematurely or wrongly assume things and it's now in the correct phase.

From what I could tell by looking at the various charts, these are my observations:
1. Big room mode at 40hz (I'll try redoing my sub-crawl to see if there's a better position but not many options available)
2. A dip slightly above 100hz which could potentially be SBIR which I will need to test out. (No acoustic panels behind speakers at the moment)
3. Very bad reflections with RT60 above 1s

Could I check from an experienced soul what else I may be missing?
I suppose I didn't really need a mic since the consensus is that the first thing after getting good speakers is to treat the room with bass traps in all corners ideally and absorbers in first reflections, but is there anything else that seems like a big problem?

Thank you

 

[Devnull]

Don't do a sub crawl. Wrestle the sub into your main listening position and then just move the microphone from position to position.

 

[Flaesh]

An unusual and strange room

 

[Keith_W]

1. It does not matter if the polarity of your speakers is inverted, as long as both of them have the same polarity. It may be a problem if you are trying to integrate speakers with inverted polarity with a subwoofer which has correct polarity, however. You can spot this in the impulse or step response. These measurements (here, comparing L speaker to R speaker with sub off) show that both speakers have inverted polarity and the tweeter alignment to the woofer is very slightly off. I wouldn't worry about it, you won't hear it and the only consequence is a slightly less pretty graph.

2. This measurement compares L+R only (blue) with L+R+Sub (yellow). You can see that the subwoofer impulse is arriving too late, my guess is about 15ms. If you want to find out how late, you need individual measurements of each speaker and sub with a timing reference.

3. Since the RT60 was mentioned, we should take a closer look at it. Select the measurement with the subwoofer (L+R+Sub), go to the "RT60 Decay" graph, and click "Generate". I'll train your eyes a bit with this graph:

In the lower graph, we have a waterfall. Take note of where I placed the pointer, at 1.5kHz. Now look at the upper graph. Observe how the SPL drops fairly rapidly then flattens out at -45dB? And also see how REW has extrapolated the drop-off (green line) to give you an RT60 of 390ms at 1.5kHz?

Now I move the pointer to 105Hz. The SPL drop is much slower, and the noise floor is higher at -30dB. And REW has extrapolated the RT60 to 1.229s. You can also see a ridge on the waterfall graph. This tells you that sound is bouncing back and forth like a ping-pong ball - it's a nasty room mode, and it's prolonging your decay time. Examination of all your measurements show that the culprit is the right speaker.

I don't know why your right speaker is doing this, but it's most likely a room issue. Do you have a photo of your room? I will need to get an idea of where the speakers are positioned and where the walls are. The solution would be to re-orient your system or install room treatment. Room treatment to absorb 105Hz is no joke, it has to be very thick (and ugly and expensive!) so I try to avoid recommending it when I can.

I would also suggest re-measuring at a quieter time and with appliances turned off (fridge, computers with fans, etc) to make absolutely sure that it is NOT random noise causing spurious results. Choose a longer sweep (1M) to increase noise rejection.

Anyway I can tell that the room seems rather large. That huge spike at 15ms is the reflection from the rear wall, assuming a rectangular room. A 15ms time lag is equal to a distance of 5.15m, meaning the distance between mic and rear wall is about 2.55m. You can also see that the reflection from the right speaker (green) is stronger than the left (red).

 

[Kix1991]

View attachment 498019

1. It does not matter if the polarity of your speakers is inverted, as long as both of them have the same polarity. It may be a problem if you are trying to integrate speakers with inverted polarity with a subwoofer which has correct polarity, however. You can spot this in the impulse or step response. These measurements (here, comparing L speaker to R speaker with sub off) show that both speakers have inverted polarity and the tweeter alignment to the woofer is very slightly off. I wouldn't worry about it, you won't hear it and the only consequence is a slightly less pretty graph.

View attachment 498020

2. This measurement compares L+R only (blue) with L+R+Sub (yellow). You can see that the subwoofer impulse is arriving too late, my guess is about 15ms. If you want to find out how late, you need individual measurements of each speaker and sub with a timing reference.

3. Since the RT60 was mentioned, we should take a closer look at it. Select the measurement with the subwoofer (L+R+Sub), go to the "RT60 Decay" graph, and click "Generate". I'll train your eyes a bit with this graph:

View attachment 498022

In the lower graph, we have a waterfall. Take note of where I placed the pointer, at 1.5kHz. Now look at the upper graph. Observe how the SPL drops fairly rapidly then flattens out at -45dB? And also see how REW has extrapolated the drop-off (green line) to give you an RT60 of 390ms at 1.5kHz?

View attachment 498023

Now I move the pointer to 105Hz. The SPL drop is much slower, and the noise floor is higher at -30dB. And REW has extrapolated the RT60 to 1.229s. You can also see a ridge on the waterfall graph. This tells you that sound is bouncing back and forth like a ping-pong ball - it's a nasty room mode, and it's prolonging your decay time. Examination of all your measurements show that the culprit is the right speaker.

I don't know why your right speaker is doing this, but it's most likely a room issue. Do you have a photo of your room? I will need to get an idea of where the speakers are positioned and where the walls are. The solution would be to re-orient your system or install room treatment. Room treatment to absorb 105Hz is no joke, it has to be very thick (and ugly and expensive!) so I try to avoid recommending it when I can.

I would also suggest re-measuring at a quieter time and with appliances turned off (fridge, computers with fans, etc) to make absolutely sure that it is NOT random noise causing spurious results. Choose a longer sweep (1M) to increase noise rejection.

View attachment 498027

Anyway I can tell that the room seems rather large. That huge spike at 15ms is the reflection from the rear wall, assuming a rectangular room. A 15ms time lag is equal to a distance of 5.15m, meaning the distance between mic and rear wall is about 2.55m. You can also see that the reflection from the right speaker (green) is stronger than the left (red).

Oh wow, thank you so so much for taking the effort into writing such a detailed analysis, immensely helpful in knowing what to look for.

I've rerun the measurements today (Apologies, I did re-invert the speaker cables for the measurements today - just an slight OCD thing - hope it doesn't make things too confusing) and if I'm not wrong the same issues are still there, split the zip into speaker measurements and sub/background measurements (These are done on 1M as advised, I also did 2 sweeps each).

As far as background noise based on my sub measurements it seems my PC fan is indeed constantly humming and seems to produce constant noise. I do turn off my fan and AC during measurements. I can try retaking the measurements tomorrow with the fans tuned lower if that might help the analysis.

Attaching a couple of photos of my room in the exact format I'm taking the measurements in, the dimensions of the visible room are ~3.6m (Wall opposite and behind speakers) and ~3.3m on the window and wardrobe side. However the wardrobe depth is around ~50cm which I'm guessing might affect the measurement so effectively the measurements are 4.1m and 3.3m for front/back and sides respectively. Curtains on the left are thin.

Sub is placed on the right beside the PC and close to the wardrobe corner based on a sub crawl I did previously but will test out other areas over the weekends.

Again, really appreciate you taking your time to share your knowledge with a total newbie.

 

[DWPress]

Get a thick rug first and some tapestry or something for the walls - that looks (and measures) like a very reverberant room with sound waves bouncing everywhere!

 

[Keith_W]

Well, I was right about the distance between the mic and the rear wall. But the room was smaller than I thought

Something has gone badly wrong with your new set of measurements. These two graphs should be EXACTLY the same - L+R with sub off from your new set (purple) vs. your previous set (blue). I am sure you can see the problem. Ask yourself what has changed, do some diagnostic work. According to you, the only difference is that you inverted the polarity of your speaker cable.

There is also an obvious comb filtering pattern above 1.5kHz which is present in your individual speaker measurements as well - left (yellow) and right (green). Comb filtering is VERY UNUSUAL in a speaker which is measured by itself because it relies on two drivers reproducing the same frequency range. For a comb filtering pattern like this to occur, the woofer has to go all the way up to 20kHz to interfere with the tweeter. So ... why is there a comb filtering pattern?

Because your new measurements are of your left and right speaker taken together! Here is your new set of measurements of the left speaker (green) vs the old set (red). See that double impulse in the green measurement? That's two speakers!

Anyway, measurement of two speakers together shouldn't give you a comb filtering pattern unless something is very off with the relative phase or delay. I would wager you made a mistake when you recabled your speaker. Please check it carefully.

You should examine your measurements WHILE YOU ARE TAKING THEM and reject any nonsense measurements. Something like this should be easily seen and should prompt you to investigate what is going on before you post the measurements on ASR.

Anyway, that room of yours is full of hard surfaces. It is no wonder that the RT60 times are so long.

ASR members might be sick of me plugging this book by now, but you should really download it and read it: free REW eBook.

 

[Kix1991]

Well, I was right about the distance between the mic and the rear wall. But the room was smaller than I thought

View attachment 498217

Something has gone badly wrong with your new set of measurements. These two graphs should be EXACTLY the same - L+R with sub off from your new set (purple) vs. your previous set (blue). I am sure you can see the problem. Ask yourself what has changed, do some diagnostic work. According to you, the only difference is that you inverted the polarity of your speaker cable.

View attachment 498223

There is also an obvious comb filtering pattern above 1.5kHz which is present in your individual speaker measurements as well - left (yellow) and right (green). Comb filtering is VERY UNUSUAL in a speaker which is measured by itself because it relies on two drivers reproducing the same frequency range. For a comb filtering pattern like this to occur, the woofer has to go all the way up to 20kHz to interfere with the tweeter. So ... why is there a comb filtering pattern?

View attachment 498228

Because your new measurements are of your left and right speaker taken together! Here is your new set of measurements of the left speaker (green) vs the old set (red). See that double impulse in the green measurement? That's two speakers!

Anyway, measurement of two speakers together shouldn't give you a comb filtering pattern unless something is very off with the relative phase or delay. I would wager you made a mistake when you recabled your speaker. Please check it carefully.

You should examine your measurements WHILE YOU ARE TAKING THEM and reject any nonsense measurements. Something like this should be easily seen and should prompt you to investigate what is going on before you post the measurements on ASR.

Anyway, that room of yours is full of hard surfaces. It is no wonder that the RT60 times are so long.

ASR members might be sick of me plugging this book by now, but you should really download it and read it: free REW eBook.

Thank you! And my apologies for posting nonsense - I'll be sure to check them against the original chart first in future.

That's super odd - I've just checked and the cables on speaker end seem to be connected fine (just inverted from the original chart). I will troubleshoot and vet through my tests over the weekends to see what might be causing it.

Indeed, I'd figured it definitely needed absorption/diffusion but was hoping to get an idea of the scale needed and potential trouble-spots so your inputs are a huge help which I will use to compare in a before/after manner while planning treatment.

Really appreciate the link to the eBook, I had indeed read Amir's guide but sadly it had dropped off at part 2 so this is a godsend!

 

[Keith_W]

Bear in mind, your measurement which is purportedly of one speaker, is actually both speakers swept together.

That REW eBook was written because ASR members have been pestering Amir for years to update his REW guide. I tried to fit it all in one post, but realized I could not get everything to fit. So it became an eBook.

 

[janbth]

Bear in mind, your measurement which is purportedly of one speaker, is actually both speakers swept together.

Can you rule out that the second peak in the impulse response isn't the reflection from the desk from a single speaker?

 

[Keith_W]

Can you rule out that the second peak in the impulse response isn't the reflection from the desk from a single speaker?

Good thought! It's possible, but unlikely. This is why:

Red: the previous measurement of his left speaker. Green = current measurement. Assuming both measurements were taken under the same conditions, the double impulse is much stronger in the green measurement.

BTW this set of measurements is a good example of why context is helpful. If you tried to read the original set of measurements without context of what the room is like, it's so much more difficult.