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

[Tupisac]

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.

New mdat file is too big to attach here so I uploaded it to my google drive: https://drive.google.com/file/d/1_rWvPyL_OdJlXZBDxKrQG2-s_Z-T-COK/view?usp=sharing

This time I did all measurements using acoustic reference. Each raw measurement (raw meaning no xo, no delays) was repeated 4 times for L, R, Sub and L+R. I managed to get each one without catching any cars passing by
I also did some different xo measurements with summed sub and speakers. LP stands for Low Pass, HP is High Pass. No delays were used.

I used this tutorial as a config reference for MMM measurements: https://www.audiosciencereview.com/...phone-method-mmm-for-dummies-using-rew.51333/

I tried calibrating SPL levels but apparently failed again. I'm clearly missing something. First I did the calibration thing in SPL meter using speaker cal signal. Adjusted the volume on the MOTU until it showed 75 dB on my phone positioned next to a mic:

Then I set the input gain on the MOTU until I got to this in the check level window (around -30). I've tried setting the gain to get -12 but it was clipping during stereo measurements:

Some screenshots for the non-rew people:

 

[Tupisac]

If you want to tame and focus highs you want diffusers and it will work.

I don't care about highs that much. I'm a hobby guitarist and my wife is a hobby singer. 100 - 1000 Hz region is where all our range happens .

 

[ZolaIII]

I don't care about highs that much. I'm a hobby guitarist and my wife is a hobby singer. 100 - 1000 Hz region is where all our range happens .

I noticed the one on the wall didn't know about wife, nice. What I wanted to point out is that you can settle up with highs easy and relatively achievable to mids. Lows and up to scholder are problatic regarding accustic treatment and better dealed off with DSP. You wouldn't believe in extraordinary cases what whose a span of female vocal, up to 15~16 KHz. Take Uma Sumac as example. Putting foam plugs in speakers make them something in between open/close buffle by declain order and by behaviour. Box F isn't completely eliminated but even with cabinets that aren't rigged enough it's enough to don't make them show. Port has minimal flow but function ain't there so you lose good part of it's extension. But you don't need it (sub does that part anyway) and back port's close to the wall won't be a good thing anyhow. You correctly assumed first refraction would be ceiling and lower then predicted by measurements because isometric shape and proximity. You are lucky those speakers actually behave very good regarding such still better approach would be to iron out beaming with phase shift then trying to treat refraction to null the holes you have in range you said it's important too you. I really don't want you to spend a lot of money on accustic treatment that in the end won't work but try to make it deacent as it is and with what you already have.

 

[Tupisac]

better dealed off with DSP.

My DSP is fairly limited. Only 7 band PEQ + high and low pass and only on the outputs. I can set the delay down to 0,02 ms.

But I've been eyeing some tutorials and wondered if it's worth reverting back to software equalization and start playing with convolution plugins. It is all wayyy above my head tho.

 

[ZolaIII]

My DSP is fairly limited. Only 7 band PEQ + high and low pass and only on the outputs. I can set the delay down to 0,02 ms.

But I've been eyeing some tutorials and wondered if it's worth reverting back to software equalization and start playing with convolution plugins. It is all wayyy above my head tho.

Unfortunately for unlimited capabilities that's still the only option and not for sakes of convolers or convolution kernels alone. Lot of other stuff too as: EBU R128 too source and ISO 226 2003 (ELC) to SPL target (program loudness) and so on. Both of us are still on JRiver tho use different third party convolers as embedded one is bad.

Edit: even your old MiniDSP has only 7 PEQ banks per chenel still that's about 10 to 15 useful trough full range and quite enough. Tap based EQ including FIR are not scaling symmetricaly trough octaves and you need a lot of taps to get to bottom ones with some precision. On the other hand PEQ Q (quality) factor is determinant how precise you can aim with it and it's best used (disregarding it's phase impact) up to a shoulder frequency. It's not computing expensive if you don't use very large number of them so minimal latency, on the other hand FIR with lot of taps is and do to latency not really for latency sensitive tasks like live streaming/performing with syncing it's OK for let's say video and cinematics but still taking care about latency. There are and phase manipulation tools (DAW) after all as RePhase.

 

[Tupisac]

You are lucky those speakers actually behave very good

Well, not quite. There are audible issues. Left speaker is randomly popping during tone sweeps. Probably it's electronic guts are on it's last legs. Both speakers also rattle from the woofer ring thing at various frequencies. It's high time to retire them.

I've been eyeing Neumanns since forever. There is one tempting deal on KH-120s II. I wonder how much difference would they make with it's MA1 magic. My wife would probably kill me tho...

 

[Tupisac]

Unfortunately original measurement data doesn't come with waterfall plots.

Apparently MKIIs are the same thing. Just different looks: https://www.erinsaudiocorner.com/loudspeakers/jbl_305pmk2/

 

[ZolaIII]

Well, not quite. There are audible issues. Left speaker is randomly popping during tone sweeps. Probably it's electronic guts are on it's last legs. Both speakers also rattle from the woofer ring thing at various frequencies. It's high time to retire them.

I've been eyeing Neumanns since forever. There is one tempting deal on KH-120s II. I wonder how much difference would they make with it's MA1 magic. My wife would probably kill me tho...

I told you how to deal with Box FS already and that's the problem they got while it will improve and cuple other things on the way. They would do more on their own then MA1 module and you need their sub to go that way (which all do very good is rather expensive). Port is in front and they are very linear and even spread prioritising direct sound (relative narrow dispersion). Didn't look predicted refractions for them in the long time. What I ment directly is inhabit refraction response and impact of such on JBL LSR 305.
Hire with illustration how much worser it is with Control 1

https://www.audiosciencereview.com/forum/index.php?attachments/jbl-control-1-pro-compard-to-jbl-305p-mark-ii-early-reflections-response-measurements-png.45241/

And I know lot of much more perfect speakers than exhibit much wilder ones especially from ceiling ones.

 

[ZolaIII]

Apparently MKIIs are the same thing. Just different looks: https://www.erinsaudiocorner.com/loudspeakers/jbl_305pmk2/

They are and Erin never did waterfall plots, more so it's same measurement equipment. At least there is a close field measurement there.

 

[Tupisac]

I plugged the ports with socks. Here are the measurements:

Left:

Right:

Left+right:

MMM left+right:

Not much difference really - besides some bass reduction.

 

[Keith_W]

If your measurements are too big to upload, it means you haven't curated your MDAT enough. All I did was glance at all the repeated measurements and deleted all the unnecessary ones.

Re: adjusting SPL. Here is another method:

1. Turn on REW's generator and play pink noise.
2. Place the SPL meter next to the microphone and adjust the volume to 75dB. Make sure the SPL meter is on the "C weighted" setting and peak capture is on.
3. Turn on REW's RTA and start recording. Let it run for maybe 10-20 seconds.
4. Take note of the SPL of REW's RTA. Let's say it's 20dB. Now you know that all your measurements are off by 55dB. Go to all your measurements, right click, "SPL offset" and type +55dB.

Anyway you have to wonder why you need that sub at all. Your main speakers roll off at 50Hz. Your sub also rolls off at 50Hz. In other words, your sub isn't contributing anything meaningful except to complicate your setup. Maybe you could try positioning your sub elsewhere in the room so that it goes lower. If it won't do that, then there's no point keeping your sub.

As for the rest of the measurements - i'm not sure what you want me to look at. They look OK.

 

[Tupisac]

Anyway you have to wonder why you need that sub at all. Your main speakers roll off at 50Hz. Your sub also rolls off at 50Hz. In other words, your sub isn't contributing anything meaningful except to complicate your setup. Maybe you could try positioning your sub elsewhere in the room so that it goes lower. If it won't do that, then there's no point keeping your sub.

I've been asking myself the same question for quite some time, including this very forum. But tbh even if it doesn't show on the graphs it does make an audible difference. Adds a lot extra gravity and makes kick drums way more tactile. Adding delays really helped with that too.

Theoretically it goes flat down below 40 on the driver alone and below 30 with the port. I've measured it's outputs from 5 cm away:

The only other possible placement for it is under the left speaker. Here are some old measurements when I was experimenting with sub placement and Sonarworks (red is in the corner, green is under left speaker):

Those were heavily squished by EQ and I don't even remember the exact positioning but it shows possible extension. I just remember there was something I didn't like about it. More smeared? Less snappy? Something along those lines. I guess I should try it again with the DSP and delays.

As for the rest of the measurements - i'm not sure what you want me to look at. They look OK.

The main question I was hoping to get some hints on is what to expect from acoustic treatment. Is it possible to get few Hz more of extension if I stuff two meters of 30 cm thick bass trap column in the opposite corner where most of the bass buildup happens. Is it possible to bury some of the dips around 250 and 600 Hz with a ceiling cloud or thick panels on the walls. I guess it was broadly answered already but I still wonder if maybe there are some obvious issues visible in my measurements that can be addressed by treatment.

I also wonder if I could adress some of the issues by EQing individual speakers. Again, maybe there is something I'm missing which is apparent to a professional or someone way deeper in this rabbit hole than me. This also extends into the new speakers dilemma and how good is Neumanns correction...

But tbh I was also itching to get a full CSI style dissection of my measurements by proper nerds like in good old internet days .

 

[Tupisac]

I've noticed something interesting regarding plugged ports. Not much of a difference in FR but here is a step response against non plugged one:

I wonder how to interpret this.

 

[Keith_W]

I've been asking myself the same question for quite some time, including this very forum. But tbh even if it doesn't show on the graphs it does make an audible difference. Adds a lot extra gravity and makes kick drums way more tactile. Adding delays really helped with that too.

You are making an important observation. "Bass quantity" depends on two dimensions - (1) how loud it is, and (2) how long it rings. Everyone seems to think that bass ringing is bad and it should be killed with room treatment, Dirac ART, bass arrays, and so on. Sure, too much bass ringing is bad. But so is too little, as you are starting to find out. At some point, preference and taste takes over. I can't look at a graph and tell you whether you'll prefer A or B, you have to listen. I can only tell you if it's far off and not supposed to be there.

Case in point: your next post showing plugged vs. unplugged port. The plugged port has louder ringing, and it rings for longer. This comparison is only meaningful if both measurements are at the same volume. Assuming they are, the red measurement is objectively worse. But does it sound worse to you? What do you think?

Theoretically it goes flat down below 40 on the driver alone and below 30 with the port. I've measured it's outputs from 5 cm away:

That measurement is meaningless. If you want to know the performance of your subs, you can't just measure from 5cm away. You will need a quasi-anechoic measurement of your driver AND the port, and some way to sum them. You will also need to scale the SPL of the port measurement since the output of the diaphragm is concentrated into a small opening, the increased velocity will give apparent higher measured SPL. Or was it lower, I can't remember without looking it up You will then need a loopback timing reference to sum them. My usual advice is to forget trying to get that measurement. What matters is what your sub does at the listening position.

 

[Tupisac]

But does it sound worse to you? What do you think?

Frankly I don't know. I haven't had the time to listen. I just did quick measurements. It was definitely not the exact same positioning tho. Gotta do a second round of tests. I also wonder how much influence high pass has on both versions.

But the socks got noticeably warm. I guess port acts as a cooling vent for the electronics too.

That measurement is meaningless. If you want to know the performance of your subs, you can't just measure from 5cm away.

Well, it served it's purpose. Sub was second hand and I was wondering if it was broken or something.

 

[Tupisac]

I've been chewing through Keith's books and tweaking all the REW settings according to recommendations. And the wavelet showed something going on around 500-600 Hz in all of my measurements.

L+R:

Left:

Right:

It coincides with the wide dip in FR:

What is going on in there?

 

[Keith_W]

In that spectro view, what you are looking at is a reflection. If you look at the dotted line, it corresponds with a region of excess phase. The dip confirms it: it's a reflection causing cancellation between 500-600Hz. I thought that my book says something like that, but you have a nicer example than in my book

 

[Tupisac]

Feel free to use it in next revision if you want to. I'd be honored . You can use all my drawings too.

Could you please help me find the possible source or this reflection? All the distances are here: https://www.audiosciencereview.com/...00-hz-range-with-treatment.72700/post-2655767

By the way, it's also a point where RT60 measurements fall below the limit. Nominal RT60 was taken from amroc pro calculator.

 

[Keith_W]

At 500Hz, the wavelength is 0.686m, so a quarter wavelength is about 17cm. This means you only need to move your mic 17cm, and that dip will disappear. It's not worth trying to find where that reflection is coming from. I would simply check that it disappears and ignore it. Don't obsess over peaks and dips in the transition zone, there is very little you can do about it.

 

[Tupisac]

At 500Hz, the wavelength is 0.686m, so a quarter wavelength is about 17cm. This means you only need to move your mic 17cm, and that dip will disappear.

Nope, still there. It just moves around on the FR a bit. Dip is also quite visible in previous mmm measurements.


Red is listening position, green is 17 cm to the left and blue is 17 cm to the right:

Listening position:

17 cm to the left:

17 cm to the right: