First go at room measurements using UMIK-1 and REW

[C0mbat]

Hi all,

So today is my first attempt at doing some measurements using my new UMIK-1 and REW. Below are 7 attempts and, as you can probably tell, I've got a sub and speaker combination, otherwise I'm sure the low end would never measure that much higher than the upper frequencies.

My set up is a WiiM Pro Plus feeding a Kef KC62 that then feeds two Aiyima A07 Max, running in mono mode, that finally feed my Kef LS50 Meta.

I'm using the High Pass Filter on the KC62 that is set at 80hz. I originally had the crossover on the KC62 set at 70hz (on run one) and have continually changed it to the point that I had it set at 50hz on run 7 (highlighted in purple). The other thing that I did during the course of those 7 runs is p;ay around with the volume on the KC62 directly using the volume control on the sub. At no point during the runs did I change the High Pass Filter from 80hz.

Given that those manual tweaks resulted in such a big change, I'm keen to know if you think I should make any further adjustments to the sub volume, the crossover or the high pass filter because considering using the 4 parametric eq channels on the WiiM Pro Plus.

And does anything else stand out as concerning? This measurement was only taken from the left speaker so I can do the right speaker measurement as required.

 

[5-pot-fan]

Hi,
and welcome to this particular audio rabbit hole!
I am not an expert in this field having spent only about 100 hours with REW, but will kick off some responses to you.
1) You will need to measure the RH speaker as well, and it will be worth measuring both together so you can see the differences between averaging the L+R measurements in REW and the actual L+R results.
2) Do practice both types of measurement - M(oving) M(icrophone) M(easurement) and single point. You will also want to experiment with both the 0 and 90degree Umik calibration files and holding the mic accordingly.
3) When you have more results do look at @OCA (ObsessiveCompulsiveAudiophile) Youtube series
4) Do you have any method (or desire) for incorporating more DSP in the system (such as CamillaDSP)?
If this is saying stuff you have already done, my apologies.
If my LH speaker result looked like yours I would be fairly happy that I had a good starting point for making some further improvements.
Hopefully others will be along with more advice and suggestions before long.

 

[ZolaIII]

Try with positioning to eliminate the uper bass hole (170~280 Hz). Do the sub measurements alone and speakers separately. Unfortunately the 80 Hz isn't a very good crossover point for those speakers. Idea is 140 Hz but then you need two sub's (each chenel one). I need to see what's happening with that deficit at crossover point. On a bright side it looks you don't have major room modes so you can spend those 4 PEQ's on first four bumps up to mids. Let's hope WiiM will increase number of PEQ's available soon enough.

 

[ZolaIII]

@5-pot-fan use 45° call and positioning of UMIK-1 vertical. He is very limited regarding DSP so solving base tonally is priority with what he has and to the listening position.

 

[C0mbat]

Hi,
and welcome to this particular audio rabbit hole!
I am not an expert in this field having spent only about 100 hours with REW, but will kick off some responses to you.
1) You will need to measure the RH speaker as well, and it will be worth measuring both together so you can see the differences between averaging the L+R measurements in REW and the actual L+R results.
2) Do practice both types of measurement - M(oving) M(icrophone) M(easurement) and single point. You will also want to experiment with both the 0 and 90degree Umik calibration files and holding the mic accordingly.
3) When you have more results do look at @OCA (ObsessiveCompulsiveAudiophile) Youtube series
4) Do you have any method (or desire) for incorporating more DSP in the system (such as CamillaDSP)?
If this is saying stuff you have already done, my apologies.
If my LH speaker result looked like yours I would be fairly happy that I had a good starting point for making some further improvements.
Hopefully others will be along with more advice and suggestions before long.

Thanks for this. I'll get on with more measuring over the next few days and report back. When it comes to DSP, I'll start with the 4 channels that the WiiM provides and just wait for them to provide more (as they have stated that they will) because I'm not keen to get too far down the rabbit hole

 

[C0mbat]

Try with positioning to eliminate the uper bass hole (170~280 Hz). Do the sub measurements alone and speakers separately. Unfortunately the 80 Hz isn't a very good crossover point for those speakers. Idea is 140 Hz but then you need two sub's (each chenel one). I need to see what's happening with that deficit at crossover point. On a bright side it looks you don't have major room modes so you can spend those 4 PEQ's on first four bumps up to mids. Let's hope WiiM will increase number of PEQ's available soon enough.

Thanks for this. I would assume that the upper bass hole of 170-280hz is the speakers rather than the sub? It's strange because the left speaker is already close to the corner of the room. 15" from the rear and 8" from the side. What's weird is that the hole was not as pronounced on the 2nd run. But it was pronounced on the 1st run. That's despite the fact that I don't think I changed anything between them. So I maybe had the mic in a slightly different position or place. Some more runs will confirm.

Once I'm happy with a set of measurements do you (as the user) define the number of peq channels that you have available to alter and then the software chooses which bumps to resolve? I'm also assuming that you can define the frequency window to tell it to choose bumps between 20-800hz?

Sorry, one other thing. I'm not going to get a second KC62 anytime soon as my wife would divorce me, plus it would likely be over kill for my 6m x 5m (Edit: It's actually only 3.6m x 3.7m) room. But are you suggesting that I should increase both the low pass filter from 50hz and the high pass filter from 80hz? What would be your suggestion?

 

[ZolaIII]

Well they aren't chenels but slots. Main thing is to solve the uper bass hole (with speakers placement) or at least minimise it as podible and it's in speakers range and probably caused by flour or sealing refractions. We will see is 80 Hz deep room or bad crossover (probably room) and how to minimise it. Mic will do good bass measurements no matter of angle of microphone, putting it at 45° helps with highs measurements (with the same call file for your own UMIK-1). Just uper peeks from 40 to 400 Hz (four of them). Would be good if you could address and one at 600 Hz (the fifth one).

 

[RickyC34]

If your room is not a very large open-concept room REW's room sim could be of great value for optimal speaker/sub/ listening position. Plug your room dimensions and current placement into the room sim and star by moving your subwoofer around the room and see the effect it has on your response. (pic included)

Measure Front right, left, and sub all by themselves along with a sweep of your subs + mains and we will be able to get a better understanding of how things are working.

I'm not super familiar with the KC62. Does it offer PEQ or a way to adjsut delay? If so, adjusting the delay could help smooth out some of those dips.

What's your main source? If pc, equalizer APO is free software that would allow you more PEQ filters.

 

[HarmonicTHD]

Aside what good advice you already got, could you post your graphs with either VAR or 1/12th smoothing. This way you see the room modes more pronounced and can better what to do.

Yes I second, what has been said before, try to get a proper PEQ (APO or Camilla).

 

[C0mbat]

If your room is not a very large open-concept room REW's room sim could be of great value for optimal speaker/sub/ listening position. Plug your room dimensions and current placement into the room sim and star by moving your subwoofer around the room and see the effect it has on your response. (pic included)

Measure Front right, left, and sub all by themselves along with a sweep of your subs + mains and we will be able to get a better understanding of how things are working.

I'm not super familiar with the KC62. Does it offer PEQ or a way to adjsut delay? If so, adjusting the delay could help smooth out some of those dips.

What's your main source? If pc, equalizer APO is free software that would allow you more PEQ filters.

Wow, this is interesting. I think I'm a way off learning how to use this but I've nonetheless filled in some of the measurements for you to see below.

1. The parts circled in orange haven't been filled in as I don't know what they mean.
2. The room is smaller than I thought it was
3. There doesn't seem to be a way of individually adjusting the left and right speaker. My seating position is off centre as I have my work desk in front of me and near the window. It means that I have to have the left speaker close to the corner and the right speaker more towards the centre of the room. I couldn't have the right speaker in the corner anyway as the door is there (closed position straight yellow line and open position dotted yellow line
4. I always have my door open
5. The sub is on the left but in front of the left speaker. There isn't really room to put the sub behind it
6. The purple lines indicate the roof line that sweeps up from the back of the room and cuts in about 30% in to the room

I'm not sure what I would do next but I thought it would be of interest to show my room layout.

 

[Purité Audio]

The figures circled are absorption coefficients ( how absorbent the surfaces are) I wouldn’t worry about those and you can move speakers independently maybe untick ‘keep left and right symmetrical’.
Keith

 

[C0mbat]

Tried some measurements this morning but now I'm frequently getting a warning message after running it of "high measurement distortion" (I'm getting it on every run and not just the last one). Is that because of the low end as it's reaching near 85db? Also could that damage my sub because you'll see from the last measurement in pink that the low end just completely dropped off. However, I have been playing around with the subs volume and crossover so I'm hoping it's just that. Either way, I've stopped taking measurements for now in case I'm doing damage (which I would doubt as the sub volume isn't that high and I think it should be able to go higher).

That last pink line is all over the place to be honest so I'm not sure what is happening. Maybe the mic?

Now showing in 1/12 smoothing rather than the 1/6 from last night

These are the measurements without the last run. They're showing a huge dip at 85hz that wasn't as pronounced last night. I'm only showing some of the runs below. What I started to do was increase the crossover on the sub (the low pass filter) and it started to help reduce that dip. But by the end I had the low pass at 70hz and the high pass was still at 80hz, which I'm not sure is right. The other thing that I've previous seen (maybe on Erin's review of the LS50s) is that the LS50s drop off 5db at around 80hz and then drop off significantly from 60hz. But that would unlikley explain it. My thoughts are that there is a gap in the crossover but maybe it's something else? Also, there are a lot more bumps in general from last night. Is that due to the distortion that it's picking up (as per the message on screen) or is that due to the switch from 1/6 to 1/12 smoothing in the graph display?

 

[ZolaIII]

You just made hole lot of mess.
Use 1/6 smoothing as psy and no smoothing for raw response. By the way that's only how it shows it all measurements are done no smoothing and you can change it as you like in all SPL - controls. Do measurements on high 70's dB so that moods are active but not going wild. Do separate subwoofer (turn off power amplifiers or dial them down for mains) and speakers (set volume to 0 on sub) without timing reference. Then adjust volume so that they match as better as you can and do a full range measurements with timing reference. And show them, actually upload saved .mdat files of measurements (at least full range ones with timing reference) so that we can see all graphs and what is happening. What's the purple one, the only one without sub bass ringing as hell and that's why you had a warning. I don't really want to comment on overpriced toy sub with two 6.5" driver's which whosent better even then cheap 8" one. Of course things will get worse when you mismatch the crossover and make hole with it.
I told you to try with moving mains to minimise at least hole in uper bass. Whan you show separate measurements of sub and mains crossed at 80 or more Hz both we will see how they actually behave and what can be done (if one is missing to give output as it should). After all of that one of us will help you to by hand calculate PEQ's as you only have 4 slots for tham on WiiM. Then you will check how that translate and adjust them more (± dB) and that's it. The speaker that is in corner will have 2~3 dB higher gain then other one so if there is a balance control on amp use it. Edited: adjust the gain knob of monoblock amplifier for that speaker accordingly.
Don't worry you will get it done and then you can play with it regarding personal preference and how it sounds to you.

 

[C0mbat]

Hi all,

It took me a while to find the time to get back around to this due to family and work life. I did spend some time in the first part of January messing around the found that I needed to move the left speaker back as I had it too far out due my concerns about it being in the corner. That filled one of the bass holes slightly.

Today I've been playing around again and can show a couple of graphs below. Maybe it's me but the measurements look ok. I realised that when measuring the right speaker on it's own I could see a big dip at 80hz when the door (2m to the right of it) was open vs when it was closed. However, when measuring both speakers and sub together it doesn't seem to be very noticeable because the sub and left speaker are clearly making up for it. the dip at 80hz is still there (door open and closed) but it's not huge. The one at 115hz is bigger and then there are two bumps at 50hz and 140hz.

I feel that I'm at a stage now when I need to use the 4 PEQ slots on the WiiM so some help would be really appreciated.

I wanted to attached the measurement files but I don't seem to be able to attach that file type. Anyone got any tips?

The below is a select view of all of the measurements that I did today. These were done with the LPF set at 60hz and the HPF set at 80hz on the KC62. See the above posts for more information on my set up. You can see quite a few bumps but I'm thinking I might need to deal with the ones at 50hz, 80hz, 115hz and 140hz? I did mess around with the crossovers but couldn't really achieve anything better than this. For example, moving the LPF to 70hz reduced the dip at 80hz but caused a bigger bump at 65hz so it was six and half a dozen as to what was better. If anything stands out as worth trying prior to using the PEQ slots let me know.

 

[staticV3]

I wanted to attached the measurement files but I don't seem to be able to attach that file type. Anyone got any tips?

To share the .mdat file on ASR, you need to Zip it first, then upload.

I feel that I'm at a stage now when I need to use the 4 PEQ slots on the WiiM so some help would be really appreciated.

I think you should focus on the frequency response below 300Hz or so.

Find the most prominent peaks in the response and reduce those.

You can use REW's EQ tool to either calculate the right filter settings, or find them manually.

Don't try to fill dips in bass response using EQ, that usually doesn't work.

Lastly, if you can, try measuring the in-room response not using REW's default measurement sweep, but using the Moving Microphone Method.

 

[boxerfan88]

View attachment 339600

Oooh.... you have an almost square room ... very similar to mine. My room is 4m x 4m.

Your room super huge mode at ~47Hz looks very similar to mine (@ 43Hz).

Allow me to share a few ideas to try:

• make the speaker-wall distances different (speaker-sidewall <> speaker-frontwall)
  • simultaneously keep the left-side & right-side symmetrical
• listening position away from the backwall (i think your diagram already shows it)
• use EQ to tame down those >10dB modes
• my square room can sound a little honky/boxy ... if so ... you can try to use EQ to tame down the 300Hz-600Hz region between 1dB to 2dB (Q anywhere from 1.0++)

Hope these ideas help...

 

[Grandzoltar]

This came from kef website

For an example, we’ll use LS50 Wireless II and a single KF92 subwoofer, and examine theKEF Connect app recommended settings in the KEF Connect app. Our engineers have determined that the best starting point when pairing LS50 Wireless II to KF92 is to set the HPF to 70 Hz, and the LPF to 50 Hz. This setting allows the slopes to complement each other so that the frequencies in between the two setpoints are properly passed to the appropriate drivers.

With the HPF filter set to 70 Hz, all frequencies above 70 Hz are sent to the LS50 Wireless II at 0 dB attenuation (full-strength). At 35 Hz (one octave below), the attenuation is -24 dB, which for all intents and purposes is fully attenuated. At one half-octave below the setpoint (around 56 Hz), the attenuation is -12 dB, which is pretty low but possibly still audible.

Setting the LPF for 50 Hz is the best complementary setting. This setting means all of the frequencies below 50 Hz are sent to the subwoofer, and frequencies at 100 Hz (one octave above) are fully attenuated. At one half octave above (75 Hz) the attenuation is approximately -12 dB.

When taken together, each driver is producing frequencies in between the set-points, but at an attenuated level that allows the drivers to work together to produce the proper volume at the frequencies in-between the setpoints.

 

[staticV3]

It's annoying that setting the HPF and LPF to the same frequency doesn't result in a perfect crossover, e.g:

 

[C0mbat]

To share the .mdat file on ASR, you need to Zip it first, then upload.

Find the most prominent peaks in the response and reduce those

Don't try to fill dips in bass response using EQ, that usually doesn't work.

Thanks for this. I didn't realise that fixing dips in a bad idea but, having read up on it, I get it. What that does suggest is that increasing the LPF crossover to 70hz (which I experimented with already) in order to reduce the 80hz dip is probably a good option even if it does increase the 65hz peak because I can use one if my peq slots to them tame 65hz.

Thanks for the information on zipping the files. I'll do that in the morning.

Im not sure how to get REW to suggest the peq settings so I'll look in to it.

 

[C0mbat]

This came from kef website

For an example, we’ll use LS50 Wireless II and a single KF92 subwoofer, and examine theKEF Connect app recommended settings in the KEF Connect app. Our engineers have determined that the best starting point when pairing LS50 Wireless II to KF92 is to set the HPF to 70 Hz, and the LPF to 50 Hz. This setting allows the slopes to complement each other so that the frequencies in between the two setpoints are properly passed to the appropriate drivers.

With the HPF filter set to 70 Hz, all frequencies above 70 Hz are sent to the LS50 Wireless II at 0 dB attenuation (full-strength). At 35 Hz (one octave below), the attenuation is -24 dB, which for all intents and purposes is fully attenuated. At one half-octave below the setpoint (around 56 Hz), the attenuation is -12 dB, which is pretty low but possibly still audible.

Thanks for this. Can you explain what you're trying to explain to me please? I have previously read up on the above and that's why I've kept the two filters 20hz apart.