Mucking about with MiniDSP UMIK-1, REW and JRiver

[Count Arthur]

After languishing in a warehouse in China for about three weeks, the UMIK-1 I ordered a while back finally arrived.

I placed the microphone roughly where my head is whist sat at my desk and this is what I measured:

I was surprised by the big drop-off below 100Hz, I've never considered my speakers particularly bass shy before; am I doing something wrong.

In any case I set about generating the EQ settings:

As my speakers are placed asymetrically in the room, I generated a separate set of EQ settings for the left and right channels and set about plumbing them all into JRiver, which was quite a tedious process, but here we are:

The +1.5dB at the begginging was so that the volume sounded, subjectively, roughly the same as without the EQ settings aplied, so that I could switch it on and off and compare the sound.

The differences:

Greater clarity - I thought my setup sounded fine, but now when I bypass the EQ, it sounds a bit "muddy".
Imaging - This is quite different, spearate instruments seem to be easier to locate within the soundstage. In addition, on some tracks the singer seems to be further in front of the rest of the band.

Overall, I'm pleased with the result. I might try messing around with a subwoofer later and see what that does.

Has anyone else tried, what were your results/impressions?

 

[Matias]

After languishing in a warehouse in China for about three weeks, the UMIK-1 I ordered a while back finally arrived.

I placed the microphone roughly where my head is whist sat at my desk and this is what I measured:

View attachment 58996

I was surprised by the big drop-off below 100Hz, I've never considered my speakers particularly bass shy before; am I doing something wrong.

In any case I set about generating the EQ settings:

View attachment 58997

As my speakers are placed asymetrically in the room, I generated a separate set of EQ settings for the left and right channels and set about plumbing them all into JRiver, which was quite a tedious process, but here we are:

View attachment 58999

The +1.5dB at the begginging was so that the volume sounded, subjectively, roughly the same as without the EQ settings aplied, so that I could switch it on and off and compare the sound.

The differences:

Greater clarity - I thought my setup sounded fine, but now when I bypass the EQ, it sounds a bit "muddy".
Imaging - This is quite different, spearate instruments seem to be easier to locate within the soundstage. In addition, on some tracks the singer seems to be further in front of the rest of the band.

Overall, I'm pleased with the result. I might try messing around with a subwoofer later and see what that does.

Has anyone else tried, what were your results/impressions?

A few tips as I have been doing the same last few months:

1. Don't use a flat curve on the listening position. Flat is for measuring a speaker near field. All house curves have slope decreasing high frequency.
Mine is here:
https://www.audiosciencereview.com/...is-your-favorite-house-curve.2382/post-341190

2. On the EQ window on top there is a total headroom required to compensate the peaks.

The difference indeed is great.

 

[Count Arthur]

Don't use a flat curve on the listening position. Flat is for measuring a speaker near field. All house curves have slope decreasing high frequency.

I actually listen to my speakers near field, they are about 1m away on the desk; does that make a difference to my choice of house curve?

It should also be noted that I can't hear anything much over 14KHz, and probably nothing meaningfull over 12KHz. My ears were damaged as a child and age hasn't helped matters.

 

[Matias]

Btw EQing adds peaks to the signal that may clip it, that is why the headroom required field says you should lower the signal a few dBs. Yes, it will sound lower so you have to increase the volume a little too make up.

 

[Count Arthur]

Ah, I think I may have moved the target the wrong way, I should set it so there is minimal dips below, so that it's mainly cutting, rather than boosting. Correct?

 

[DanGuitarMan]

Making eq corrections from a single sweep may cause more problems than they solve. Have a look in to moving microphone measurements, which average the responses over a listening area. Kind of like what popular room correct softwares do when they have you measure multiple locations. Here's a video and here's a paper on the idea. Also, consider avoiding positive EQs values at first, and only subtract with the EQs - you might cause clipping, distortions, blown drivers, etc...

 

[Soniclife]

2. On the EQ window on top there is a total headroom required to compensate the peaks.

This.
Ensure you attenuate by at least this much before applying the corrections. It will be quieter than without DSP, unless you create an attenuated normal version and flip between them, that's what I do.

 

[Soniclife]

Making eq corrections from a single sweep may cause more problems than they solve. Have a look in to moving microphone measurements, which average the responses over a listening area. Kind of like what popular room correct softwares do when they have you measure multiple locations. Here's a video and here's a paper on the idea. Also, consider avoiding positive EQs values at first, and only subtract with the EQs - you might cause clipping, distortions, blown drivers, etc...

I'm not sure mmm is great for nearfield, the relative distance from mic to driver can change a lot as you move the mic.

 

[Soniclife]

I was surprised by the big drop-off below 100Hz, I've never considered my speakers particularly bass shy before; am I doing something wrong.

What speakers are they?
If the correction sounds better the measurement is probably close to right.

 

[DanGuitarMan]

Ah, I think I may have moved the target the wrong way, I should set it so there is minimal dips below, so that it's mainly cutting, rather than boosting. Correct?

Yes, I think it's safest to start by cutting at first.

As an example, here's a screenshot of my room's response before and after EQ:

 

[DanGuitarMan]

I'm not sure mmm is great for nearfield, the relative distance from mic to driver can change a lot as you move the mic.

I think it is valid for nearfield if you move it in a smaller pattern, like in a spiral around where your head would be in the listening position. This is essentially what Dirac does in having you measure many points, whether it be an auditorium, or in a chair at your desk.

 

[DanGuitarMan]

I was surprised by the big drop-off below 100Hz, I've never considered my speakers particularly bass shy before; am I doing something wrong.

Not sure what speakers you use, but they are probably much more linear than that graph. This is most likely a local phenomenon - they way your speaker and room modes interact with each other at that particular point in space.

 

[Count Arthur]

What speakers are they?

Dynaudio Audience 52SE: https://www.audioholics.com/bookshelf-speaker-reviews/dynaudio-52se/page-3

JRiver -> USB -> RME ADI-2 DAC -> Hypex UcD700HG HXR

 

[QMuse]

Dynaudio Audience 52SE: https://www.audioholics.com/bookshelf-speaker-reviews/dynaudio-52se/page-3

JRiver -> USB -> RME ADI-2 DAC -> Hypex UcD700HG HXR

There is something wrong with your measurements.
What was the distance between mic and speakers?
Were you standing behind your LP and holding mic in your hand where your head would be?
Was mic pointed between the speakers?

Can you provide a photo of your setup so we get an idea where are you sitting vs the speakers position?

Btw, how comes the level is so high (above 100dB)?

 

[Count Arthur]

I had the mic fixed to the back of my chair, pointing straight ahead between the speakers. My head and the two speakers form an equilateral triangle about 0.8m on each side.

I tried with a downward tilted house curve, but I prefer it flat. This may be due to my hearing, with its inbuilt rolled off high frequencies.

 

[Soniclife]

I tried with a downward tilted house curve, but I prefer it flat. This may be due to my hearing, with its inbuilt rolled off high frequencies.

I think flat is correct for nearfield, the typical curves are for further away.
If it turns out there is a researched curve for nearfield I'd like to give it a go.

 

[Count Arthur]

I found out why my speakers apparently produce no bass I have that huge drop off at 110Hz.

I've been merrily measuring the sound output from the crummy 5w sound bar beneath my monitor - what a muppet!

Here is a measurement of one of my actual speakers:

At my listening position, they actually go pretty low.

 

[Soniclife]

I've been merrily measuring the sound output from the crummy 5w sound bar beneath my monitor - what a muppet!

Ha. I had a good chuckle at that, then I realised I was laughing because it's exactly the sort of thing I'd do.

So before when you applied correction and it improved things, was that an improvement to the soundbar, or your speakers?

 

[Count Arthur]

So before when you applied correction and it improved things, was that an improvement to the soundbar, or your speakers?

Yes, oddly, despite measuring the soundbar, then applying the suggested EQ to my main speakers, I quite liked it.

I'm now mucking about with measuring with and without the subwoofer. The only problem is the time it takes to type all the settings into JRiver. I wish there was a means to import the REW settings.

 

[QMuse]

Yes, oddly, despite measuring the soundbar, then applying the suggested EQ to my main speakers, I quite liked it.

I'm now mucking about with measuring with and without the subwoofer. The only problem is the time it takes to type all the settings into JRiver. I wish there was a means to import the REW settings.

I think you shoud be able to use filter file in wav format with Jriver which REW can create from PEQ filters.