Next steps for speaker placement and room correction

[bcurtin]

After buying a miniDSP umik I've gone down the rabbit hole of speaker placement and trying to correct what I believe are nasty room modes and SBIR, but have one in particular that doesn't seem to respond too differently to speaker location.

I have a pair of KEF 104/2s placed in a condo living room, which includes a staircase, dining room, kitchen - a large, non-rectangular volume. The speakers aren't centered along the wall they are against, with the back of the left speaker 23 cm from the rear wall and the left side 75 cm from the left wall. The back of the right speaker is also 23 cm from the rear wall and the right side is 152 cm from the right wall. There's no specific room treatment and unfortunately the most convenient listening position is ~400 cm away from the speakers and about halfway along the long axis of the room.

Below is the left (red) and right (blue) speaker RTA measurement with full range pink noise before applying any DSP correction with var smoothing. Low frequencies are slightly boosted and there are large nulls in the right speaker near 65 Hz, 150 Hz, and 300 Hz. I swapped the left and right speakers to make sure it wasn't a driver issue and confirmed it's location and not speaker dependent.

Below is the left (green) and right (orange) speaker RTA measurement with full range pink noise with DSP correction to 500 Hz and a slight bass boost, avoiding nulls and in particular the right speaker drop around 60 - 70 Hz. I also have a set of filters to 1k since there seemed to be some issues beyond 500 Hz but have not measured this one yet.

DSP is definitely an improvement in tonal balance after knocking down some of the low frequency peaks, but that massive null in the right speaker bothers me and I'm curious if there's other feedback on these measurements. Any suggestions or recommendations are appreciated!

 

[dougi]

What DSP are you using? I'm surprised it gave no boost at all to the nulls as the seem fairly wide. Mine would have attempted to correct those somewhat. How do you think it sounds? There seems a fair hump in the treble which may make it edgy. It may be worth seeing if that responds to EQ.

 

[bcurtin]

What DSP are you using? I'm surprised it gave no boost at all to the nulls as the seem fairly wide. Mine would have attempted to correct those somewhat. How do you think it sounds? There seems a fair hump in the treble which may make it edgy. It may be worth seeing if that responds to EQ.

I used REW to take averaged RTA data with the mic swept around my listening position, and then used REW EQ to define the DSP filter type, frequency, gain, and Q. I also used REW to generate filter data that I imported into Roon. The EQ settings I used were:

• Match range, 20 - 500 Hz
• Individual max boost = 8 dB
• Overall max boost = 8 dB
• Flatness target = 1 dB
• Allow narrow filters < 200 Hz
• Do not vary max Q above 200 Hz

The right null is ~12 dB lower than the target so it doesn't attempt to boost.

The treble was surprising since I don't have them toed in much and they are known for having clear but not sharp highs. I can't say that I've been bothered by treble and if anything noticed that the bass, in particular kick drums, sounded lacking.

 

[Hipper]

Whilst my room is symmetrical and I employ lots of room treatment I had a similar issue with a deep wide null at 50Hz. See here bottom graph on my post #60:

https://www.audiosciencereview.com/...in-room-measurements.13540/page-3#post-411614

Before I EQ'd this dip I tried everything else - moving the speakers, trying to find where in the room this frequency might be using test tones, my ears and an SPL meter (a rough plan was that if I could find 50Hz louder then it should be in some spot in the room I might employ room treatment targeted to that frequency and location). Nothing worked so in desperation (as we are told that you should not really fill in big nulls with EQ) I just tried a bit of EQ - 5dB - and amazingly 5dB was removed from the dip. Although I said I used REW filters, in fact these didn't work properly so I filled in the null manually by adding 10dB then using REW I measured and generated the filters, applying them manually to my equaliser.

The result looks good but more importantly, it sounds better.

 

[alex-z]

You need acoustic treatment and subwoofers. Treatment for the 100-300Hz region, subs for below 100Hz.

Don't use EQ to fix a 15dB null, that is like trying to drive a nail with a balloon. That eats into valuable headroom, and worsens the reflections which also reach your ears. EQ should be applied in such a way that you are fixing the speakers overall radiation pattern, not the room.

 

[bcurtin]

Whilst my room is symmetrical and I employ lots of room treatment I had a similar issue with a deep wide null at 50Hz. See here bottom graph on my post #60:

https://www.audiosciencereview.com/...in-room-measurements.13540/page-3#post-411614

Before I EQ'd this dip I tried everything else - moving the speakers, trying to find where in the room this frequency might be using test tones, my ears and an SPL meter (a rough plan was that if I could find 50Hz louder then it should be in some spot in the room I might employ room treatment targeted to that frequency and location). Nothing worked so in desperation (as we are told that you should not really fill in big nulls with EQ) I just tried a bit of EQ - 5dB - and amazingly 5dB was removed from the dip. Although I said I used REW filters, in fact these didn't work properly so I filled in the null manually by adding 10dB then using REW I measured and generated the filters, applying them manually to my equaliser.

The result looks good but more importantly, it sounds better.

Thanks for the link to your thread and sharing what helped. I might try to add some EQ to the dip to see if I can improve it without a sub, but had tried a much wider boost range that "corrected" the null but instead did what @alex-z mentioned and made things worse (distortion/clipping).

Given the speaker arrangement, measured response, and not having any existing treatment, is there anything from the RTA measurement that points to treatment that would help 100 - 300 Hz? Happy to take some pictures of sketch out a floor plan, but we have some furniture and other things that I may be able to move around if a corner or wall needs some padding.

 

[Hipper]

In my case I moved the speakers and it made no difference. After checking that it wasn't something faulty in my equaliser or some other gear I resorted to using EQ.

Boosting a null in the bass region often does not make any difference. This is because of the nature of low frequency sound waves - they have wavelengths similar to the size of your room. For example, 65Hz is 5.3 meters.

Your problem is clearly the left speaker position. You could experiment by temporarily repositioning your left speaker and measure if it makes any difference. If it does that would seem to me to suggest a sub is required. I don't know much about using subs but the idea is that you can position it where you like whereas a full range speaker needs to be positioned mostly for mid and high frequencies which often compromises bass sound.

You could download this test tone CD to listen if your current arrangements really make a difference to what you hear:

http://realtraps.com/test-cd.htm

 

[abdo123]

tbh you can fix everything with a little bit of patience and money. but that 2KHz to 10KHz messiness?

what is going on? did you find any spinorama data / measurments for the speakers?

 

[abdo123]

the messiness in the areas highlighted in red is purely SBIR. you need to move the speakers till they're ~1.5 meters away from the back wall.

or put acoustic absorbers (bass traps) behind them that absorbs frequencies all the way down to 100 Hz.

If i were to guess, your system sounds as if the bass is disattached and the voices are hollow. perhaps also a bit bright.

 

[bcurtin]

tbh you can fix everything with a little bit of patience and money. but that 2KHz to 10KHz messiness?

what is going on? did you find any spinorama data / measurments for the speakers?

Interesting you pointed that out - something I've been puzzled by after doing room measurements on these speakers. I don't believe there's any spinorama data since they're about 30 years old. I also restored the woofers, added fluid to the tweeters, and replaced some of the gaskets so hopefully a generic spinorama would be relevant to my speakers but not sure. I'll go through old measurements and see if I can make sense of what's happening there.

I had recently placed the speakers closer to the wall to see if that help some of the nulls, and have measurements adjusting the position around what I had used nominally so can look at that or move the speakers back too and remeasure. The speakers have a very good phantom center image and vocals do sound echo-ey, but not too bright to me (but it could be because I'm used to them by now).

 

[puppet]

My take is (300-600hz) floor bounce.
(150hz range) wall(s).
(65hz) a room null.

Move them around and get clear of the walls if you can.

 

[Daverz]

My take is (300-600hz) floor bounce.
(150hz range) wall(s).
(65hz) a room null.

Move them around and get clear of the walls if you can.

Agree that 300-600 Hz dip is probably floor bounce. The dip will move around depending on the speaker to listener distance. Treat with some thick rugs.

 

[MarkHall]

Anyone in nyc want to help me run a ReW test and advise some clouds and placement of current GIK panels .. i live in the space .. can pay some money..

 

[AdamG]

Anyone in nyc want to help me run a ReW test and advise some clouds and placement of current GIK panels .. i live in the space .. can pay some money..

Welcome Aboard @MarkHall.