• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required. There are many reviews of audio hardware and expert members to help answer your questions. Click here to have your audio equipment measured for free!

My first REW measurement - Harbeth P3esr

Archsam

Senior Member
Joined
Apr 8, 2020
Messages
326
Likes
513
Location
London, UK
My minidsp UMIK-1 mic arrived yesterday, so after work today I did my first measurement, following Julian Krause's famous Youtube tutorial. The graph below is the average of 7 measurements taken with both L+R speakers.

What could have caused the big dip around 130Hz? Is that what is referred to as signal cancellation from L and R channel?

Average.jpg
 
  • Like
Reactions: mkt

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
What could have caused the big dip around 130Hz? Is that what is referred to as signal cancellation from L and R channel?

That's what a room does to the bass (varies, but something will likely be broken in the lower frequencies).

Here, around 48Hz, the standing waves from the left and right speakers, in my asymmetrical room, become 180 degrees out of phase, and cancel at the listening position.

The frequency that is affected will vary with the geometry of the speakers, walls, and micropohone.
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
Cancellation between left and right seems very unlikely. It looks like a typical room node at one speaker or both. I would suggest simple single point left and right sweeps at the listening position.

Here's the in-room bass response of my Vandersteen Quatro speakers. The speakers are 8 feet apart. Note the nulls between 100 and 200 Hz in the right speaker response (green) that aren't in the left speaker response.

20-200Hz-detail.jpg
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
Cancellation between left and right seems very unlikely.

Maybe.

My room does it:

Phase of left and right speakers at the listening position:

1594422361230.png


The left rear corner of the room is open.

I presume the standing wave phase difference is due to the longer path into which the left speaker fires than the right one does around the listening position.

Left and right each have a small dip (red), but combined, big dip (green).

1594422720720.png
 
Last edited:

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
I stand corrected, then. I should probably do a combined L-R measurement of my own speakers.

I'm no acoustician, but that's my analysis given the data at hand.

I didn't "notice" it until measuring.

I still don't really hear it.

When the bass is actually in stereo the loss is minimized, as the phase in the recording will change things.
 
OP
Archsam

Archsam

Senior Member
Joined
Apr 8, 2020
Messages
326
Likes
513
Location
London, UK
Below is the measurement of the left (red) vs right (blue), taken about 8 inches apart.

My living room is asymmetrical as well - on the left the room is a bit more open, on the right it is against the corner plus the fireplace protruding between the speaker and my position.

L and R.jpg
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
I think I should be able to see left-right cancellation with REW's vector average, no? I see a few points where this occurs, particularly 77 Hz and a bit at 43 Hz. Vector average is thick green line.

left-right-vector-average.jpg
 

RayDunzl

Grand Contributor
Central Scrutinizer
Joined
Mar 9, 2016
Messages
13,198
Likes
16,981
Location
Riverview FL
I think I should be able to see left-right cancellation with REW's vector average, no?

In my case, vector average of left and right shows the additional dip around 48Hz.

I don't use moving mic method. Moving mic wouldn't give an appropriate phase measure, I should think.


Vector Average (red) of the left and right speakers adjusted up by 6dB almost exactly matches the actual measured (black) response of both speakers.

1594428387768.png
 

Daverz

Major Contributor
Joined
Mar 17, 2019
Messages
1,294
Likes
1,451
Below is the measurement of the left (red) vs right (blue), taken about 8 inches apart.

My living room is asymmetrical as well - on the left the room is a bit more open, on the right it is against the corner plus the fireplace protruding between the speaker and my position.

View attachment 72731

Despite the asymmetry, the response above 200 Hz looks excellent. The dip at 130 looks like it might be wide enough to respond to digital room correction.
 

JohnYang1997

Master Contributor
Technical Expert
Audio Company
Joined
Dec 28, 2018
Messages
7,175
Likes
18,292
Location
China
Measure only one channel at a time. 1 meter from the speaker or pointing at the speaker from your listening position. The two measure different things. Also you can measure the two drivers individually from 5 to 10cm away from each driver and overlay the graphs.
 

Hipper

Addicted to Fun and Learning
Joined
Jun 16, 2019
Messages
753
Likes
625
Location
Herts., England
I'm not sure the first graph you posted shows 'no smoothing'. I would expect the graph to become more jagged as you go higher up the frequencies.

In your post 7, considering the asymmetrical nature of your room, both your speakers seem to have a very similar frequency response (FR). Are you sure you didn't have sound from both speakers when you did both measurements.

Anyway, I suggest you concentrate your measurements on the 0-500Hz region. You can do this on REW. To measure the left speaker only, I suggest you firstly get the REW signal to come only from the left speaker, and then place the microphone pointing directly at the left speaker from your listening position and placing it where your left ear would be. Do the same for the right speaker.

When looking at the FR graph use 'no smoothing' and the smallest range of parameters on the two graph axis that will include all the peaks and dips.

The big dip in your first graph is likely a room mode:

https://amcoustics.com/articles/roommodes

If you can move your speakers and listening chair about you might be able to improve things but it won't be perfect. If you can use room treatment that too would help.

Do you have DSP or EQ equipment?
 
OP
Archsam

Archsam

Senior Member
Joined
Apr 8, 2020
Messages
326
Likes
513
Location
London, UK
I'm not sure the first graph you posted shows 'no smoothing'. I would expect the graph to become more jagged as you go higher up the frequencies.

In your post 7, considering the asymmetrical nature of your room, both your speakers seem to have a very similar frequency response (FR). Are you sure you didn't have sound from both speakers when you did both measurements.

Anyway, I suggest you concentrate your measurements on the 0-500Hz region. You can do this on REW. To measure the left speaker only, I suggest you firstly get the REW signal to come only from the left speaker, and then place the microphone pointing directly at the left speaker from your listening position and placing it where your left ear would be. Do the same for the right speaker.

When looking at the FR graph use 'no smoothing' and the smallest range of parameters on the two graph axis that will include all the peaks and dips.

The big dip in your first graph is likely a room mode:

https://amcoustics.com/articles/roommodes

If you can move your speakers and listening chair about you might be able to improve things but it won't be perfect. If you can use room treatment that too would help.

Do you have DSP or EQ equipment?

Thanks for all the tips folks, much apperciated!

My first graph is an average generated from 7 measurements so it might have already been smoothed. The subsequent L and R graphs are adjusted mic positions to the left and right, with the mic pointed towards the individual speaker but the signal output are from both. I will do more individual measurements when I have time this weekend. I'm very curious on the measurements of the speakers themselves so I'll try to do some near field measurement that minimise the room interaction as well as a few more experiments from the listening position.

My living room is small, 4m (deep) x 3.4m (width) x 2.7m (high). The speakers are 1.7m apart measured centre to centre, and my seating position is 2.7m measured from tweeter to my ear. I have a floor plan drawn up but I'll need to go into CAD to get a diagram out. I'll do that later on.

I have not used DSP yet. These measurements is the first step towards my next move, whether I need to to invest in DSP hardware or playing around with the room itself. My setup is currently streaming flac and dsd files from a NAS to the DAC via wifi, so it's not so easy to introduce digital DSP in the signal chain, as it either requires software DSP done on a PC and fed to the DAC via USB, or a hardware DSP device fed from the router via ethernet cable and then into the DAC. I'm not keen to go with a PC / wired solution at the moment.

Currently the only solution I found is a minidsp SHD with xlr input / output that can be inserted between the DAC and amp, but I can't swallow the £1600 price tag (minidsp products are SO expensive here in the UK!) I have read about Dspeaker anti-mode but my hunch is to stick with more open software platform. I need to look into this in more detail.
 
OP
Archsam

Archsam

Senior Member
Joined
Apr 8, 2020
Messages
326
Likes
513
Location
London, UK
Despite the asymmetry, the response above 200 Hz looks excellent. The dip at 130 looks like it might be wide enough to respond to digital room correction.

Yeah i am happily surprised that the little Harbeth is doing pretty well. A few months ago I started playing with treble boost via my Yamaha amp because I had a hunch it could compensate for the relatively early drop off in treble above 10kHz, based on sterophile's measurement from 10 years ago. I turn up the treble boost about 1/4 to 1/3 of the dial and on paper will add about 3dB at 20kHz so quite a mild boost. That was good enough an improvement to stop me from buying a pair of second hand Focal 1008be on ebay!

Really curious how they will sound once I smooth out the big dip around 130hz.
 
Top Bottom