You can adjust the speaker distances by simply moving your speakers or adjusting their delays accordingly in Roon. For SPL differences Roon is the only choice unless your amp is capable of adjusting left-right balance in dB terms precisely, analog amps aren’t.
I’m not sure what exactly you mean by distortion. Long sweeps will typically require more memory but if the sweep sound is distorted this could be clipping due to too high mic or playback volume. I always select the maximum memory allocation to REW (2GB) and also allow GPU processing by default.
ok, so the measurements for SPL and time alignment you adjust in the video is just for manual use and have to be done manually, and the convolution filter, that I get at the end does only effect the frequency response, correct?
Yes there’s definitely something very wrong with the left measurement!
yes, frequency and phase.
Thinking about giving this a try.it It is a clever concept. Besides Magnus and the op who has tried it? Also I have 1 sub, I was thinking of keeping it on for the measurement.
I tried it, don’t have a sub. It definitely smoothed out the response so that it’s closer to the Harman curve.
Listened with it activated for an afternoon, then the next day switched it back off for comparison. Decided to keep it deactivated. My room is a mess don’t get me wrong. L vs R differences in bass modes (R speaker is close to a bookcase, L is out in the open), high ceilings, etc etc. Something about the DSP removed the dynamics* from the sound. It sounded less exciting.
In the past I’ve tried the EQ method as well and same thing. Maybe sounded ‘cleaner’ (maybe) but less exciting. YMMV
System = Cambridge CXA81 & CXNV2 + Linn Ninka passive (sealed) speakers
- I’m calling it ‘dynamics’ because I don’t have a better word for it.
thanks for that. I have tried different RC, but always found having it off to sound better to me.
Curious to have a go with this one. Definitely adds to the hobby, pursuing the best sounding playback for music.
Bee carefully all when using convolution in roon by this metod. You must using headrom parameter -10dB ! You are loosing amplifire power cca 20% but this is tax for this metod. 8% is +5 dB on L and +5 dB on R
Result is fantastic. Thanks lot
2 Likes
I know what you are talking about when you say dynamics. Actually I believe I can see if a correction will deteriorate the sound by comparing the height of the “unnormalized” impulse responses before and after correction. This correction method might have failed if your left and right speakers have extremely different FR. The curve generated over RMS average might have caused too much correction for a speaker (or both). You might try generating two different curves for each speaker and then add speaker gains to match them after correction if at least their LF cutoff and slopes are similar. Don’t give up that easy 
Thanks for this Serkan, you’re right, one shouldn’t give up that easily 
Perhaps when I find a minute I will try again with a curve for each speaker.
1 Like
Is it possible for the right speaker to be at time 0 when viewing the impulse response?
The left shows a minus value, but the right is at 0. Just wondering if i did not have something set correctly when taking the measurements.
Yes and it’s quite common.
I also have one sub only connected with an Hi-level connection and I am curious to know what are your findings
thanks
Hi Serkan,
sorry for the dumb question but what values I need to put in the roon speaker setup filter?
I used to put the distance from my seat to the left and right speaker, so for example the right speaker is at 2m and the left is at 2,10m but I would like to try using more accurate measurements as you do in your video. But once I have the ms delay from both the curves from 0 I am not sure what value I should use in Roon. Could you clarify please?
thanks a lot
Marco
You put the measurements calculated from the impulse response for distance.
And SPL for gain.
To answer your earlier question, I used this technique and found much it removed too much bass.
I probably need to try it again with a different LF slope and lower limit frequency.
The calculation of the impulse response and setting the distance in speaker setup made a huge difference, and I will keep that even if I don’t use the filters.
but what about the negative one? I cannot input negative values. Do I subtract the negative from the positive?
Interesting comment about the bass. I wonder if it would be better runt he whole process without the bass. Have you tried that?
You have to enter “10” to the box in the upper left corner (left speaker distance difference in cm). This will add the necessary delay to the right speaker which you will see in milliseconds at the bottom left corner (under right speaker distance box)
I was about to do the measurements today but I couldn’t get the 48khz sample rate to work with asio4all. For some reson it outputs heavily distorted test signal through my system. 44,1khz works fine and 48khz works through java drivers and I can play yotube and windows sounds with that setting but REW + asio4all distorts the test signal.
My setup is Asus Zenbook laptop with Win11 and the newest REW linked in the tutorial. I use hdmi output straight into my Linn ADSM/3. Mic is UMIK-1.
I troubleshooted for three hours but couldn’t get it to work. I tried all the sample rate settings in Windows drivers (it should work with 24/48 but no), disabled all the irrelevant audio drivers in Windows, updated the hdmi output drivers and fiddled with asio4all control panel settings. I have no idea why it doesn’t work. My previous laptop with Win10 worked fine in this same system.
Do I lose much if I just use 44,1khz?
@OCA
I noticed that in the video you suggest to align the SPL using 500Hz and 3 octaves but in the pdf you say to use 315Hz and 4 octaves.
Which one would you say to use?
Thanks