2.1 Stereo Calibration - what am I missing?

Dave the Rave

Member
Thread Starter
Joined
Jan 3, 2023
Posts
31
Location
Australia
More  
Main Amp
Yamaha AS2100
DAC
Schitt Yggdrasil
Computer Audio
Daphile (USB) or Oppo BD103 (Optical or COAX)
Front Speakers
Yamaha NS-2000
Subwoofers
JBL LS120
Hi Folks,
First off, this tool is really amazing as I immediately got a good sound change on my system. Attached are my measurements, room and right speaker filters as an example.
The measurements were on 70db reference. Speakers are approx 35cm from back, sub i front of right speaker.
Wanted to clarify the followings:

1. Obviously my right speaker is shy on the mid range (600-4k hz) compared to the right. What should I do to address that: Decrease the mid of my left speaker (My speaker has a mid and high volume level) / Bring the right speaker closer to the main hearing area / leave speakers as is and use REW EQ (which I did - see the EQ filters picture)?

2. Would you recommend sending the sub behind the right speaker - currently set at 80 hz with the volume at 2/3 max (sub is 25-150 hz)?

3. REW does not correct the 70-140 hz trough with the EQ. Should I Add a manual filter to boost that range or Re-calculate the EQ filters by setting the Overall Max Boost to 1 and Individual Max Boost to 9 or 12 or Reduce the Target Level to say 67 so that the House Curve touches those dips (but more than 75% need to be equalised?
Note: My amp plays really high (hot) so not worries if I need to crank up the volume during playback

4. Side question, is there a way to send the REW signal via a DLNA player i.e without a soundcard? I wanted to measure the response on my Daphile LMS player which is DLNA compliant.

Thanks you and looking forward to any advice



1672800548656.png



1672800641356.jpeg



1672800921241.png
 
As a starting point I would get a tape measure and get that entire system speakers and tv + stand semetrical... Maybe its just the photo and I am thinking a perfect match L and R speakers from back wall and side walls should match down to the quarter inch or better... Same with toe-in... You want a perfect semetrical triangle between speakers and your Main Listening Position (MLP)... Next maybe get those L and R speakers out into the room a bit more, leaving room behind for the sub... If you can add some/more base and broad-band absorbers... That hard wood floor needs more carpet... Next, you might look at the specs for the mains and see where they go down to and adjust the crossovers in your sub to integrate with your mains before you do EQ... Lastly, you might have a look at Serkan Gur's method to do EQ... https://www.youtube.com/channel/UCLePp5EV2RR5_Xo7rZnnpnQ

P.S. https://www.roomeqwizard.com/help/help_en-GB/html/siggen.html to create tones you can send with your player...
 
Last edited:
That isn't really a trough, it is a region between two peaks. REW is configured for no overall gain ("Overall max boost" is set to 0 dB) so it won't try and lift anything, but your room curve and target level don't match well with the response. If you raise the sub level it might come closer, or you could have a target without the low end boost.
 
Is the graph left and right only, or with the sub?

I’m surprised that symmetrically-placed speakers in a symmetrical room gets measurements this different.

What equalizer will you be using? If it’s a device where you’ll be manually applying filters, then you probably won’t want to mis-match filtering above the Schroeder frequency. I’ve found that it does weird things to the soundstage. Of course, feel free to experiment, but pay attention and see if you notice any changes in imaging.

Regards,

Wayne
 
Thanks heaps for all the replies and pointers. Just to clarify:
- Noted on asymmetrical measuring - I do have constraints though as its a living room setup. Will try sub back, main closer to MLP and increase the sub volume to see effects
- Graph is L and R with subs. My right speakers does have a lower volume than left for some weird reason. Will reduce left speaker mid to see effect.
- Eq is generic which I export as L & R mono 192 filter wav file - I import those into my Daphile system through its BruteFir convolver.
 
What speakers do you have? (for some reason your speakers are listed as your integrated AS2100
Speakers are Yamaha NS-2000 - thanks for pointing the correction.
Before trying the speaker position changes, I tried the EQ as attached to address the dips better
1672917357224.png
 
A quick look at the specs reveal:
"The Squaker and Tweeter are equipped with continuously variable level control.
The attenuator used has a sufficiently large allowable input with little degradation in sound quality."
Frequency: "28 Hz to 20 kHz"
3-Way with Crossovers at: "500 Hz, 6 kHz (-12dB/oct.)"

You might pull the grill off the speakers and see if there are two variable level controls... You might want to match the left and right speaker drivers SPL with those controls...
 
Last edited:
(My speaker has a mid and high volume level)


You might pull the grill off the speakers and see if there are two variable level controls... You might want to match the left and right speaker drivers SPL with those controls...

@Dave the Rave , I was thinking the same as @ddude003 . It's why I asked what the speakers are. The the speakers should be a lot more closely matched before eq if possible and as @Wayne A. Pflughaupt advised ...

then you probably won’t want to mis-match filtering above the Schroeder frequency. I’ve found that it does weird things to the soundstage. Of course, feel free to experiment, but pay attention and see if you notice any changes in imaging.
 
Thanks. I will try matching the speaker levels first before EQing this time.
Please note my subs are connected to the pre-out of my Integrated Amp - i.e I do not have an active crossover controller (like a receiver)
Should I measure L and R speakers with the sub on always?
Or should I measure Sub first then L and R without sub => analyse and apply EQ for sub first (say up to 50hz where sub response is better than the speakers) then EQ the L and R from 51 hz to 20k?
 
I would get your L&R speakers drivers SPL matched first... Then a Left only and Right only measurement(s) and the Sub only... Once you get your L&Rs worked out you can then integrate your Sub... Blend... Low pass/crossover and volume/gain to match your mains... Then some EQ... Try and post the .mdat file(s) not screen shorts...
 
Last edited:
I would get your L&R speakers drivers SPL matched first... Then a Left only and Right only measurement(s) and the Sub only... Once you get your L&Rs worked out you can then integrate your Sub... Low pass/crossover and volume/gain to match your mains... Then some EQ... Try and post the .mdat file(s) not screen shorts...

I was just gonna say the same, lol.
Getting those measurements using an acoustic timing reference in REW may also prove useful.
 
Getting those measurements using an acoustic timing reference in REW may also prove useful.

LOL... And you can get the timing reference with the Generator as noted above...
 
Again appreciate all the guidance. So as advised,
- I moved the speakers more towards the LP
- Dialed down the mid and high by 2 on the left speaker and by 1 on the right speaker (to rectify my previous measurement observation)
- Level matched the front/left and subs except let sub run a bit higher
- Took 2 measurements without subs (see attached)
- Played around with the sub volume and crossover. After analyzing the FR, I set the sub at 60hzand volume at 2PM.
- Then took speakers+sub measurement. Also took L+R+Sub measurements (see attached).
- Applied EQ using the Harman House Curve then played it into my system.
=> Sound is way better than previous attempts. I am thinking of applying 0.5 on the target LF and HF slopes of the House Curve rather than leaving them as 0.
Please let me know if I am on the right track.
Thank you


1673100216155.png
 
Hopefully zipped DAT file get attached this time
 

Attachments

There's some kind of something I don't understand going on with at least some of these measurements...

This looks pretty normal. (Even though it shows a couple of very strong reflections)

Left Impulse.png



Here is the next measurement overlaid on the first....

DAC2  L NoSub2.png


And the (presumably) first measurement of the right speaker...

DAC2 R NoSub1.png


Any idea what happened here?
If not, perhaps @John Mulcahy , might have a suggestion.
 
There's some kind of something I don't understand going on with at least some of these measurements...

This looks pretty normal. (Even though it shows a couple of very strong reflections)

View attachment 58288


Here is the next measurement overlaid on the first....

View attachment 58289

And the (presumably) first measurement of the right speaker...

View attachment 58290

Any idea what happened here?
If not, perhaps @John Mulcahy , might have a suggestion.
DAC2 R NoSub1 IR peak is at negative time hence the odd shape. Something is wrong with the measurement timings. Clock ppms are also very high.
 
I did miss an important point. My right speaker definitely has something wrong. I could hear some sort of distortion around 3/4 way of the sweep. Dont know if some driver got screwed....
 
But the analysis shows something wrong on the left speaker (IR)... I recall I got an error saying the Reference Timing could not be completed hence I set the Reference Level Trim to -2 (upon reading on the web). But i did that at the beginning so not sure that level stayed for subsequent measurements.
 
I would be more worried about this one. The phase rotations (even after IR shift) between 300Hz and 3000Hz and the odd dip at 3kHz might point to a faulty midwoofer but this is beyond me, sorry.

DAC2 R NoSub1.jpg
 
The measurements have very large clock adjustments and the adjustments are wrong, so there is some issue with the setup to cause that. It doesn't affect the frequency responses, but it would be better to measure with the Analysis option to "Adjust clock with acoustic ref" turned off to get more useable results.
 
I would be more worried about this one. The phase rotations (even after IR shift) between 300Hz and 3000Hz and the odd dip at 3kHz might point to a faulty midwoofer but this is beyond me, sorry.

View attachment 58295
@Dave the Rave , @serko70 has applied a frequency dependent window to your measurement. You could clarify the setting with him, but the effect is look at the measurement with less or possibly no reflections included in the measurement. This is why it looks different from your measurement. Yours includes reflections that arrived at the mic after the direct sound did. You can tell because REW added [FDW] after the measurement name on the graph legend.
 
but it would be better to measure with the Analysis option to "Adjust clock with acoustic ref" turned off to get more useable results.
It seems to me, or did you once recommend keeping this option always on? Correcting the clock if the input and output devices are different is the right idea. But as I noticed, after the correction, the shape of the impulse near the peak changes.
 
More research reveals that these speakers were made from the early 80s until the late 80s, making them 35ish to 40ish years old... Might be time to repair/replace some/all the drivers and have the crossovers checked and brought into spec if need be...
 
Last edited:
Back
Top