REW measurements. Is it possible to fix 100-1000 Hz range with treatment?

[kyuu]

And yes, speakers were upside down but currently they are back to normal orientation.

I would actually put them back to the "upside down" orientation. That puts the tweeters at ear-level when seated, it seems like, which is what you want.

 

[Keith_W]

Erm... I'm ashamed to admit it but yours look nothing like mine and I don't even know where to start... All my REW knowledge comes from random tutorials.

1. Select "Impulse" tab
2. Click "ETC" and unclick all other graphs
3. On the top left, choose dBFS from the drop down box instead of %
4. Zoom in to the area of interest. Ctrl-Right Click-Drag to make a box, then click inside the box to zoom. The same box also shows time / distance between two points, so you can use it to automatically determine distance instead of manually calculating it like the book shows.

 

[Keith_W]

What about those popular absorbers with diffusers on the front?

Take a look at "MAC Talk Treble", in the second half of the presentation, room treatment is discussed.

I don't think you need more absorption. Remember that too much absorption is bad!

 

[Tupisac]

1. Select "Impulse" tab
2. Click "ETC" and unclick all other graphs
3. On the top left, choose dBFS from the drop down box instead of %
4. Zoom in to the area of interest. Ctrl-Right Click-Drag to make a box, then click inside the box to zoom. The same box also shows time / distance between two points, so you can use it to automatically determine distance instead of manually calculating it like the book shows.

Thanks.

What's that big downward spike at 26 m mark?

 

[Tupisac]

Come on now you are almost there and without dime spend. It doesn't matter put it to first output PEQ bank on each chenel you are using (L, R & Sub).

So just cut the peak across all channels? Sure, I'll measure it later.

But the phase reversal on the right side is quite unbearable to listen to. Incredibly weird. I'm not sure if I even care how it measures as it's unusable.

 

[Keith_W]

What's that big downward spike at 26 m mark?

Don't know. Maybe some external noise? It's not present in any of your other measurements, and notably it is absent when you measure your L+R speaker. In other words, if it was real, you would see it again when you repeat the measurement. Which you sort of did. I wouldn't worry about it.

 

[Keith_W]

BTW since I mentioned noise, I thought I would have a quick look at your noise floor. I use the waterfall and extended the time window to 1 sec. I also normalized the waterfall to peak at every frequency.

Random noise tends to look like mountains (you can see some random noise between 60-80Hz), and tones look like "sheets". You can see that you have some equipment generating tones in your room, at 250Hz, 480Hz, and 650Hz. These tones are too high for a computer fan - if a fan spins at 2000rpm, then it will generate a tone at (2000/60) = 33Hz. I don't know what's causing it. You might want to take a noise floor measurement. It's the same as doing an MMM except that you don't play any sound through your speaker and you don't move the microphone.

Note that a sweep does not tell you what the noise floor actually is, since the length and volume of the sweep suppresses the noise floor. It might look as if it's -60dB to the signal in this measurement (or about 35dB absolute), but that's not real.

 

[Tupisac]

BTW since I mentioned noise, I thought I would have a quick look at your noise floor. I use the waterfall and extended the time window to 1 sec. I also normalized the waterfall to peak at every frequency.

Random noise tends to look like mountains (you can see some random noise between 60-80Hz), and tones look like "sheets". You can see that you have some equipment generating tones in your room, at 250Hz, 480Hz, and 650Hz. These tones are too high for a computer fan - if a fan spins at 2000rpm, then it will generate a tone at (2000/60) = 33Hz. I don't know what's causing it. You might want to take a noise floor measurement. It's the same as doing an MMM except that you don't play any sound through your speaker and you don't move the microphone.

Note that a sweep does not tell you what the noise floor actually is, since the length and volume of the sweep suppresses the noise floor. It might look as if it's -60dB to the signal in this measurement (or about 35dB absolute), but that's not real.

Oooh, that's incredibly useful info.

So there is some random buzzing audible during sweeps. Some definitely come from the speakers. I've tried tightening the woofer ring screws and managed to kill some of the rattle. But I had trouble finding the exact frequencies and was just clicking through a sine generator. Now I can target specific ones

Some of the rattle comes from somewhere else, but I had hard time locating it with my ears. Might be the ceiling, might be some of the furniture. Otherwise I don't know what it could be. I turn off air purifier and fan for the tests. I have a Mac Mini so no loud fans there either. I'll try measuring again during the weekend so there is less traffic too. And maybe 5-10 dB quieter as is seems to reduce some rattle too and it's closer to my normal listening levels.

By the way - mine looks a bit different again (less detail) and I can't figure out why:

 

[ZolaIII]

So just cut the peak across all channels? Sure, I'll measure it later.

But the phase reversal on the right side is quite unbearable to listen to. Incredibly weird. I'm not sure if I even care how it measures as it's unusable.

I don't know what you mean by that, measurements will show, phase flips shouldn't be what you are hearing and it would be nice to share measurements and let's see. With accustic reference of course.

 

[Tupisac]

I don't know what you mean by that, measurements will show, phase flips shouldn't be what you are hearing and it would be nice to share measurements and let's see. With accustic reference of course.

I'll do the measurements. But the phase reversal of one side only sounds a bit like severely overblown stereo haas effect. It also feels like it wants to squish my eardrums in. Very strange. The soundstage is there but instruments and voices feel like they are smeared right next to my face (kinda like headphones) instead of coming from some distance in front of me. Hard to describe.

 

[Tupisac]

You can see that you have some equipment generating tones in your room, at 250Hz, 480Hz, and 650Hz.

Turns out most of those tones come from speakers, specifically the plastic ring around the woofers:

The rattle is a bit random. When ramping up the frequency to the measured rattle target it sometimes start to buzz earlier, sometimes later. Squishing it around and tapping fixes the problem for a while or moves it to another frequency when it starts buzzing again. The problem is that screws are already tightened. I wonder if some light DIY job could help here. Is it okay to add some layer of insulation under that plastic ring? Like a bead of silicone or a cardboard seal?

 

[Keith_W]

I don't see why not. It might make woofer replacement more difficult down the line though, but I can't imagine it would be impossible to remove the woofer. You'll need someone who knows more than me to advise you.

 

[Tupisac]

Whew, what a journey....

First I went reading through the Keith's pdf catalog. Inspired by the info I whipped out the measuring tape and dialed in the positioning. Changed the toe-in to around 25 degrees (was around 10 on previous measurements). Made sure that everything is level and symmetrical. The mic position and all the measurements were taken with power LED of the speakers as a reference point (previously mic was just pointed at the tweeter). Reflection distances are shown in meters and roughly measured from the drawing so they are not the most precise.

I also went hunting for rattle. I managed to lower it as much as I could and it's now almost inaudible during sweeps but it still shows in measurements. Maybe it's time to retire those speakers...

You might want to take a noise floor measurement. It's the same as doing an MMM except that you don't play any sound through your speaker and you don't move the microphone.

I've attached new mdat file. It contains the noise floor rta. Measurements were taken at night.

I don't know what you mean by that, measurements will show, phase flips shouldn't be what you are hearing and it would be nice to share measurements and let's see. With accustic reference of course.

Inverting right side indeed makes it all much flatter in the 100-1000 Hz region (inverted in red):

But the spectrogram shows more of the story and the weirdness I've been hearing.
Here is measurement of the whole system without messing with phase:

And here is with inverted sub and right speaker:

Both were made using the same xo, peq and delay settings.

The measurement with inverted phase is named "L+R+S 6,4 delays xo80 -peaks, inv" and it's in the attached mdat file.

 

[Keith_W]

That noise floor is pretty unbelievable, it is close to absolute zero SPL. I would believe it if it was 30-40B higher. I think someone forgot to calibrate the SPL of his microphone

Inverting right side indeed makes it all much flatter in the 100-1000 Hz region (inverted in red):

INVERTING ONE SPEAKER!?!?? Why the heck would you want to do something as crazy as that?!?!? I stared at my screen and wondered if you had completely lost your marbles, so I decided to check.

This is the impulse response of your L and R speakers with nothing inverted. Nothing crazy here, both speakers are the correct polarity. The reason you check this is because this is the ONLY reason why inverting one speaker would produce a better outcome - that your speakers have some kind of manufacturing defect where the polarity of the +/- speaker binding posts have been reversed by the factory.

So now we check the measurement of L+R with no inversion (yellow) and with inversion (green). See the obvious problem? The green measurement has a double impulse, with the second speaker offset by about 1.4ms. In other words, you are looking at a summation error.

Indeed, if we assume that the MMM is free of summation errors, we can see that your MMM correlates nicely with your non-inverted measurement ...

... more than it correlates with your measurement where you inverted one speaker.

In short: forget about inverting one speaker. It's ridiculous, and the so-called "improvement" is a measurement artefact.

There are in fact a lot of problems with your measurement, starting with the fact that the timing ref is all over the place. Like, why is the timing ref delta in the individual L and R speaker measurement 4.58ms, but when L is measured with R, it is consistently 0.02 - 0.1ms?

 

[Tupisac]

That noise floor is pretty unbelievable, it is close to absolute zero SPL. I would believe it if it was 30-40B higher. I think someone forgot to calibrate the SPL of his microphone

Yep. I'm attaching the measurements from yesterday, when I calibrated my soundcard and did the thing with dB meter app on my phone. It showed 75 dB right at the tip of the mic, running the pink noise according to guide linked in one of your tutorial posts. REW Operator error is also very possible, as it's still quite low. Frankly I was surprised at the result as the RTA thing was showing around 35-40 dB while taking the measurement.

There are in fact a lot of problems with your measurement, starting with the fact that the timing ref is all over the place. Like, why is the timing ref delta in the individual L and R speaker measurement 4.58ms, but when L is measured with R, it is consistently 0.02 - 0.1ms?

Ok, I was using acoustic reference for stereo since my interface only has two channels. Can I just split the right signal and feed it back to the right input for loopback during stereo measurement?

 

[Keith_W]

Ok, I was using acoustic reference for stereo since my interface only has two channels. Can I just split the right signal and feed it back to the right input for loopback during stereo measurement?

No you can't. I would simply use the left speaker for all timing references and leave it at that. Don't change the timing ref, and don't move the microphone. And another thing: REW's timing ref can be a bit unreliable. I suggest you repeat the measurement 3-4 times to make sure it's consistent, then discard all the other measurements.

 

[ZolaIII]

Only thing I agree for now is that measurements aren't right.
With phase inverted it comes much more to what it should be to anechoic response and in room response.

You didn't even hit the room fundamental properly. JBL LSR305P suffer from a lot of cabinet refractions (and it shows with inverted phase while it's better masked when it aint). Unfortunately original measurement data doesn't come with waterfall plots. Regarding cabinet refractions you should plug the port and try to stres them out even more (by raising the XO). It's not loosen up bolt or easy DIY fix. Would be nice if Keith_W would guide you on how to align impulse response and improve measurements. 30 to 40 dB residental noise is in quet isolated environment (country hause far from main road) your should be higher (and a problem when you get measurements right as it will make it harder to see what room is doing). If highs remain out of focus with port's plugged and impulse aligned you need mild diffusers treatment to the wall behind you. The measurement and review data:

JBL LSR305P MKii and Control 1 Pro Monitors Review

This is a review and detailed measurements of the JBL 305P Mark ii powered studio monitor and Control 1 Pro budget passive monitor. I purchased the 305P Pro Mark ii when it first came out and it currently sells for US $147.50 from Amazon including Prime shipping. This is the price for one so...

www.audiosciencereview.com

 

[Tupisac]

30 to 40 dB residental noise is in quet isolated environment (country hause far from main road) your should be higher (and a problem when you get measurements right as it will make it harder to see what room is doing).

Last measurement sessions were done at deep night during the weekend. I live in a semi detached house built from bricks in far suburbs. There was no traffic nearby (highway is around 3 km away). I had to hold my breath as it was registering it as peaks in the RTA plot.

During the day is a very different story. Traffic, dogs barking etc. You'll probably see it in my next mdat files, as my kids and wife are coming back today. No more nightly sessions in foreseeable future.

If highs remain out of focus with port's plugged and impulse aligned you need mild diffusers treatment to the wall behind you.

I was considering plugging the ports too. Unfortunately the back wall is tricky as there isn't much space for treatment. I've also heard that diffusers need more distance than I have there to work properly.

 

[ZolaIII]

You need about 1m behind you not to be a problem but If you want to tame and focus highs you want diffusers and it will work. Absorbers get thick as it's easy to impact highs but even getting to lows neads about 30 cm deep chamber. That's and only accustic treatment I will advise you (usually advise accustic curtains as absorbers). I live in the country side (very close to the capital center) and as it's mid summer it's so quet that crickets can be annoying at night. Another time please breed.

 

[Tupisac]

crickets can be annoying at night.

It's quite cold summer here this year. Crickets noped out.