Try to learn REW and room treatment physics or just enjoy the music?

[Multicore]

I think you've already done very well, Jeff. Getting in-room sound right is complicated.

I have no means of utilizing DSP. Last year I did look into it.

Hm. Thanks for the photo. Take a look at the thread.

Let's share diagrams (and photos) of our total physical audio system and the whole signal path, with a few words and/or links

Prior to starting this new thread, I have searched all the thread titles within ASR Forum by a key word "diagram" and found we have no similar thread like this; if you know/find some, please let me know. I always believe one diagram or one photo would worth more than 1,000 words especially for...

www.audiosciencereview.com

It might inspire you to start making system diagrams, which might help us find ideas about how to incorporate equipment for compensation.

A software called draw.io is very good for this but I quite like using paper, pencils and pens.

 

[Raymond_61]

the dip up there can happen this way if distances from speakers are identical at MLP measured with laser,etc.

Sokel you may have nailed it! I believe these measurements were the first time i took measurements with both channels at the same time. Because I'm sure i dont have that on any of my left only or right only measurements. I'll con

What test signal did you use?

I believe its called "Sweep" It was not white noise.

 

[FrantzM]

Hi

You can make this system provide SOTA performance. This will require work (perhaps a lot ), but it is doable, without resorting to spending a lot of money.
You have good components. Especially the speakers, the others are also competent but on the pricey, side. You already have them , so ... keep them.
You may need to add one more .. thing. I'd suggest a mini DSP 2x4 HD. Bear with me...

I don't fully understand your seating position, it seems there is a wall, on the left ? :

The FR as measured is not good. I apologize for being so blunt.
You have issues in the bass and also the mid treble :

These are correctable.

I would suggest to start by removing most of the "Room Treatments"/ Especially the bass traps which as you can see on the REW graphs do not do much in the low bass (or elsewhere, . actually they can't do that much in the low (<100 Hz) bass, a subject for another thread or post. FOr the purpose of this exepriment could you remove all of these?

Move the seat forward say about 60 cm, past the wall on the left,
Measure and post FR and .mdat file

We can take it from there and make some suggestions.

This system can provide SOTA results, especially the complements of speakers (Kef R11) and subwoofers. (SVS SB 1000) ..

Waiting for your reply

Peace.

P.S. Can you share your floor plan? Especially your seating position and distances to boundaries, walls, etc?

 

[Raymond_61]

If you share the REW mdat file you measured, other contributors will be able to provide more detailed comments. Your listening environment looks cozy and pleasant. Are you currently sitting close to the back wall?

I tried to share the mdat file but is was not allowed because of an improper file extension.

 

[Mort]

Zip it first

 

[Lion♡]

I tried to share the mdat file but is was not allowed because of an improper file extension.

you can upload google drive or onedrive things.

 

[Raymond_61]

what microphone that you use for measurement, and how do you perform your measurement?

Hi

You can make this system provide SOTA performance. This will require work (perhaps a lot ), but it is doable, without resorting to spending a lot of money.
You have good components. Especially the speakers, the others are also competent but on the pricey, side. You already have them , so ... keep them.
You may need to add one more .. thing. I'd suggest a mini DSP 2x4 HD. Bear with me...

I don't fully understand your seating position, it seems there is a wall, on the left ? :
View attachment 453451

The FR as measured is not good. I apologize for being so blunt.
You have issues in the bass and also the mid treble :

View attachment 453453

These are correctable.

I would suggest to start by removing most of the "Room Treatments"/ Especially the bass traps which as you can see on the REW graphs do not do much in the low bass (or elsewhere, . actually they can't do that much in the low (<100 Hz) bass, a subject for another thread or post. FOr the purpose of this exepriment could you remove all of these?

Move the seat forward say about 60 cm, past the wall on the left,
Measure and post FR and .mdat file

We can take it from there and make some suggestions.

This system can provide SOTA results, especially the complements of speakers (Kef R11) and subwoofers. (SVS SB 1000) ..

Waiting for your reply

Peace.

P.S. Can you share your floor plan? Especially your seating position and distances to boundaries, walls, etc?

Thank you! Yes there is a wall. The entire space is two rectangles. 15'x15' then 9'x15'. I can provide requested info and more photos. The picture I sent has a chair for listening which was an experiment. The room now has a recliner/loveseat set up, popular for watching movies. I'm trying to ditch the loveseat and just have the lounge chair. But there's more on that topic I'd rather not get into. lol

I'm off to a Memorial Day ceremony but I will be back.
Thanks again to all !!!

 

[Raymond_61]

I tried to share the mdat file but is was not allowed because of an improper file extension.

Zip it first

Thank you! I zipped it but now the file is too big. I'll be back this evening.

 

[alankila]

I'm using the UMIK-1 in a vertical position. I watched some YouTubes on how to set up REW to take measurements mostly geared to HT setups. I just followed the steps they described and got it to create a measurement. I really didn't go much further than that. I wish I could articulate more detail but I just haven't invested the time to learn more about REW. But that's why I'm here.

Don't use vertical orientation with UMIK-1. Point them towards speakers, e.g. middle point, for stereo setup. The vertical orientation file is not measured but modeled, and I have found it to be quite inaccurate for precise work especially in treble, it is easily 1-2 dB off more than the mic is when pointing towards speakers.

Also, do not measure both channels at once. In room, this will add stereo comb filtering effects. Large dips in treble and general rise towards bass is common consequence of doing the measurement this way.

 

[Robbo99999]

Don't use vertical orientation with UMIK-1. Point them towards speakers, e.g. middle point, for stereo setup. The vertical orientation file is not measured but modeled, and I have found it to be quite inaccurate for precise work especially in treble, it is easily 1-2 dB off more than the mic is when pointing towards speakers.

Also, do not measure both channels at once. In room, this will add stereo comb filtering effects. Large dips in treble and general rise towards bass is common consequence of doing the measurement this way.

Yep, horizontal is best for UMIK, except if I remember rightly if you have a multi channel system then you use vertical. Of course you have to apply the horizontal calibration file to the UMIK in REW if you're using the UMIK horizontally, and likewise apply the vertical calibration file if using the UMIK vertically.

 

[Keith_W]

Don't use vertical orientation with UMIK-1. Point them towards speakers, e.g. middle point, for stereo setup. The vertical orientation file is not measured but modeled, and I have found it to be quite inaccurate for precise work especially in treble, it is easily 1-2 dB off more than the mic is when pointing towards speakers.

Also, do not measure both channels at once. In room, this will add stereo comb filtering effects. Large dips in treble and general rise towards bass is common consequence of doing the measurement this way.

EXACTLY!!!! I was looking at the measurement and thought to myself "measurement artefact". Fall in frequency response > 4-5kHz is a classic symptom of improper mic calibration or wrong mic orientation.

OP: please read this thread.

 

[boxerfan88]

Yeah, for stereo I prefer horizontal, less HF rolloff.

For HT, no choice but vertical.

 

[Audionaut]

I’ve never been able to move the needle much on the bass issue at 35-90 Hz.

You have two large room resonances at 37.5 Hz and 80.5 Hz. You have to equalize them first.

 

[Oddball]

Just looking at the photo, have you tried to put the speakers along the long wall? Logistically looks like there is an option to do so.

 

[Thomas Savage]

Looking at the picture the massive suck out in your frequency plots make sense ..

I'd buy a WiiM streamer , cost nothing and play with the PEQ based on the in room measurements youv taken .

Take down a few peaks in the Base and fill the suck out .. , leave it for a few weeks . See what you think ..

Looking at the picture I'd suggest ones chasing their tail and going mad . So chill out , try a bit of EQ and see how things are in a few weeks .

Good luck .

Edit , by filling the suck out with EQ ones likely just working against that random room treatment so first things first take that out and remeasure..

We've all gone mad at one point or other so don't feel bad.

 

[alankila]

You have two large room resonances at 37.5 Hz and 80.5 Hz. You have to equalize them first.

I think most people don't understand how to read these responses. I look for what the roughly anechoic level seems to be. I'm guessing it could be roughly around 76 dBSPL in this case, or at least I'm going to run with this claim until I see separate channel measurements which are more reliable for this sort of thing.

All the stuff beyond that is done, in some respect, by the room. So this gives the reference for bringing peaks down, if you for some reason desired entirely flat response. In reality, I'd target about 81 dBSPL < 100 Hz. This identifies that 100-300 Hz response is elevated, and there are those two large resonances that need some kind of targeted peaking band filters to reduce them towards some desired level.

I don't have a firm opinion of the room curve -- all that matters is that there's somewhat more bass than midrange, and probably most people prefer treble to roll off at least a few dB below midrange, but usually room acoustics takes care of that with good speakers and you probably shouldn't mess with anything above 300 Hz in this case with an equalizer, assuming that the single channel measurements that OP will no doubt post soon look reasonable.

So, to e.g. fit equalizer in REW, you'd create target shape from 30-300 Hz bandwidth which likely should have something like +3 to +5 dB worth of bass rise compared to 1 kHz midrange level across that range, and then just bring the peaks down and not even try to bring the dips up, if any. Keeping the overall gain of the filter at 0 is safe way to do that. I also prefer if I can have the entire curve be equalized only with filters with negative gain, so I set individual gains to 0 as well, because that makes kind of simple and obvious filters that look like someone might have designed them by hand. If you allow positive and negative gains, the optimizer runs with that and makes all sorts of goofy +12 dB and -18 dB filters next to each other and I dislike that sort of thing, even if it is in principle fine sonically. I also don't like using any narrow Q filters. Something like Q=5 that REW does by default is probably close to the maximum tolerable, but REW has the option to use way narrower ones and even recommends them for in-room responses, which I personally disagree with after trying them. It just sounded artifacted in practice, and I think the correction simply can't be made that precise, I think. (This is the same reason why I don't do FIR filters for room equalization: the bass region inherently requires long filters that are computationally inconvenient, and provides way too many control points at upper bass, so the correction tends to become more precise than data actually warrants.)

One way to reason your way through this soup is to realize that all measurements are approximations. To degree, they are accurate only if e.g. all doors are in similar positions, as reflections and bass leakage alter resonances. 10-20 cm differences in position may be all it takes to throw correction off by several dB. Especially if there are dips, they can be highly localized. Your very own ears are listening at approximately -11 and +11 cm offset positions relative to where you have placed the mic. So none of this is exactly very precise and it shouldn't be a goal to fit <1 dB accurate equalizer using high Q filters, because the frequency response of the system remains somewhat unknown despite we have a trace on the screen.

I prefer the moving mic method, using REW noise generator spewing pink noise and then I move the microphone around the head location in some kind of circle to average something like 20 full 65536 sample responses at 1/48 octave bins around the listening area. This helps to smooth out treble and typically is not very different < 300 Hz.

Anyway, this comment has turned into a sort of braindump. I'll stop here.

 

[Oddball]

Looking at the picture the massive suck out in your frequency plots make sense ..

I'd buy a WiiM streamer , cost nothing and play with the PEQ based on the in room measurements youv taken .

Take down a few peaks in the Base and fill the suck out .. , leave it for a few weeks . See what you think ..

Looking at the picture I'd suggest ones chasing their tail and going mad . So chill out , try a bit of EQ and see how things are in a few weeks .

Good luck .

Lol, picture says so much more than one would expect. I have gone mad, but there was nothing with EQ that could make me sane until I figured how to position speakers. Ever since, all not good, but great. Sucking out 20dB peak with EQ - good luck - could be done but your subs will be sounding like flat tires.

 

[Thomas Savage]

Lol, picture says so much more than one would expect. I have gone mad, but there was nothing with EQ that could make me sane until I figured how to position speakers. Ever since, all not good, but great. Sucking out 20dB peak with EQ - good luck - could be done but your subs will be sounding like flat tires.

The towel over the TV is what made me think ones gone mad .., random room treatment plus towel , one needs help .

Can't fill in bass with eq , pull down a few peaks yes , but one could fill that issue from 3 kHz- 13khz but then that maybe self inflicted ... take all the room treatments out then remeasure...

Also once at a certain stage , leave it and don't listen for a week or two ..,

A . This will help with expectation bias and general mental unhelpful things

B. If one can't do that , one might then consider, one has a more troubling condition. Likely therapy is needed .

 

[Mort]

B. If one can't do that , one might then consider, one has a more troubling condition. Likely therapy is needed .

Not mutually exclusive.

 

[Oddball]

The towel over the TV is what made me thing ones gone mad .., random room treatment plus towel , one needs help .

Can fill in bass with eq , pull down a few peaks yes , but one could fill that issue from 3 kHz- 13khz but then that maybe self inflicted ... take all the room treatments out then remeasure...

Also once at a certain stage , leave it and don't listen for a week or two ..,

A . This will help with expectation bias and general mental unhelpful things

B. If one can't do that , one might then consider, one has a more troubling condition. Likely therapy is needed .

Big rooms are difficult and agree that towel over the TV is a bit over the top.

There is no greater pain IMO than having a big room and lots of sub placing options. I tried that for several month with 2 subs and got nowhere. Not enough SPL or consistency in FQ response. Then got 4 subs and it was even more difficult exercise as there were more options. Luckily, I remembered what DBA means and finally can rest in peace with even a primitive implementation of the DAB.

What people often forget is that calibrating for the MLP does not really solve all your problems. Primitive DBA - all my problems solved in the entire room, including the odd hallway enclosures.