• Welcome to ASR. 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!

REW Newbie with Salon2

1742442928569.png


If you want to see whether you can hear those peaks and dips, try ERB smoothing. Blue: R channel with 1/6 smoothing, Yellow = R channel with ERB smoothing.

You can see that those peaks and dips mostly disappear. The real problem is bass roll-off below 100Hz. It's about 5dB down at 30Hz. So let us compare it with the PIR from spinorama.org that @staticV3 posted earlier. Once again, I copied the image from that link, then used an image manipulation program to copy the graph. I scaled it horizontally and vertically to match the scale of your measurement then pasted it on top so you can directly compare:

1742445377515.png


We can now easily see where your measurements deviate from the "official" Spinorama measurement. You can see that there is roll-off at both frequency extremes. Bass rolls off from 100Hz down, and treble rolls off from 13kHz up. There is also a low Q suckout between 100Hz and 300Hz. The bass roll-off is likely due your room / position of the speakers / position of listener. You can try moving your speakers and see if the situation improves, what you are looking for is a +5dB increase in bass. Otherwise, consider a subwoofer + DSP.

Treble roll-off may be due to improper microphone calibration or using your mic in the wrong orientation. Please check that you are pointing your microphone at the tweeter (i.e. horizontal) with the correct 0deg calibration file loaded. If the measurement was properly taken, the treble roll-off may be due to measurement distance. To confirm that your speakers are working properly, measure from 1m (3ft) away with your mic on-axis to the tweeter.
 
Last edited:
Had a look through your latest .mdat file. I would start by using EQ to reduce the peaks around 350, 640, 1300Hz, and 6000Hz. Those problems occur above the Schroeder frequency of your room so fixing them should improve overall tonality and stored energy without harming off-axis response.

For example:

https://prnt.sc/nMz7lZP_U9Wf - right channel EQ

https://prnt.sc/ZlaoKGaI_voo - left channel EQ

The bass is tricky because any corrections will only be useful for a narrow listening window. Before using EQ below 300Hz I would add 1-2 good quality subwoofers. This will fix the bass deficit below 80Hz and reduce the peaks and dips caused by room modes.

Low-pass the subs at 80Hz, high-pass the speakers at 80Hz. I assume your Krell processor supports this, if not you can add that functionality with an external DSP such as a miniDSP Flex.

As for the room acoustics, your decay times and clarity values (ratios of direct vs reflected sound) reach their highest point in the mid-bass region, and gradually worsen at higher frequencies. The bad news is you are losing a significant amount of vocal and instrument detail. The good news is that simple/cheap 2-4" absorption should correct this problem nicely, most residential rooms have the opposite problem and need much thicker absorption.
Thank you very much. I will try that and report back.
 
View attachment 437573

If you want to see whether you can hear those peaks and dips, try ERB smoothing. Blue: R channel with 1/6 smoothing, Yellow = R channel with ERB smoothing.

You can see that those peaks and dips mostly disappear. The real problem is bass roll-off below 100Hz. It's about 5dB down at 30Hz. So let us compare it with the PIR from spinorama.org that @staticV3 posted earlier. Once again, I copied the image from that link, then used an image manipulation program to copy the graph. I scaled it horizontally and vertically to match the scale of your measurement then pasted it on top so you can directly compare:

View attachment 437584

We can now easily see where your measurements deviate from the "official" Spinorama measurement. You can see that there is roll-off at both frequency extremes. Bass rolls off from 100Hz down, and treble rolls off from 13kHz up. There is also a low Q suckout between 100Hz and 300Hz. The bass roll-off is likely due your room / position of the speakers / position of listener. You can try moving your speakers and see if the situation improves, what you are looking for is a +5dB increase in bass. Otherwise, consider a subwoofer + DSP.

Treble roll-off may be due to improper microphone calibration or using your mic in the wrong orientation. Please check that you are pointing your microphone at the tweeter (i.e. horizontal) with the correct 0deg calibration file loaded. If the measurement was properly taken, the treble roll-off may be due to measurement distance. To confirm that your speakers are working properly, measure from 1m (3ft) away with your mic on-axis to the tweeter.
Thank you very much for your extensive response.

I have a difficult room and an even more difficult listening position based on the room and the usage of the room.
For purposes of measurement, I used a 9' equilateral triangle. That position is not usable as a MLP. I did this just to get a baseline of what I was facing.
I will re measure based on 1 meter. I'm guessing this will have a significantly different result.

These measurements were based on using bridged Mc7270 amps for the lower terminals and a Yamaha MXD1 on the treble terminals.
I decreased the gain on the Mc7270s down to 1.4V so that the bass would not spike up to 90db.
If I had left the gain at 0.75V, there would be an increase in bass from 100 hz down of 10 to 15 dB.

I will remeasure and see how it works at 1 meter and also using 0.75V gain.

Thank you very much.
 
Had a look through your latest .mdat file. I would start by using EQ to reduce the peaks around 350, 640, 1300Hz, and 6000Hz. Those problems occur above the Schroeder frequency of your room so fixing them should improve overall tonality and stored energy without harming off-axis response.

For example:

https://prnt.sc/nMz7lZP_U9Wf - right channel EQ

https://prnt.sc/ZlaoKGaI_voo - left channel EQ

The bass is tricky because any corrections will only be useful for a narrow listening window. Before using EQ below 300Hz I would add 1-2 good quality subwoofers. This will fix the bass deficit below 80Hz and reduce the peaks and dips caused by room modes.

Low-pass the subs at 80Hz, high-pass the speakers at 80Hz. I assume your Krell processor supports this, if not you can add that functionality with an external DSP such as a miniDSP Flex.

As for the room acoustics, your decay times and clarity values (ratios of direct vs reflected sound) reach their highest point in the mid-bass region, and gradually worsen at higher frequencies. The bad news is you are losing a significant amount of vocal and instrument detail. The good news is that simple/cheap 2-4" absorption should correct this problem nicely, most residential rooms have the opposite problem and need much thicker absorption.
Thank you so much for your extensive response.

Greatly appreciated.
 
Thank you.

I have the microphone on a boom stand at tweeter level and pointed straight up. I'm using the 90 degree calibration file.

I will try it at 1 meter pointed directly at the tweeter using the non90 degree calibration file.
 
Last edited:
As @alex-z said, this graph looks pretty normal.

My next step (in your shoes) would be to:

1) Download and configure SoundSource for Mac ($40)

2) Use REW to create a set of EQ filters to correct the bass (lots of tutorials on this, it's mostly automatic)

3) Load those into the headphone EQ slot in Soundsource (I know it's not headphones but it works anyway)

4) Enjoy your dramatically smoother bass.

BTW - when you do the graphs, it's important to select the right scale so you can tell what's going on. Normally it's a 50dB scale, in the most recent set it's zoomed out to show 400dB which is more than enough to show the full waveform of a nuclear bomb going off 8 feet from the mic... it tends to make your speaker / room issues look small in comparison. ;)
@kemmler3D Finally figured out how to use the Soundscape + REW. Wonderful. Thank you for the suggestion.
Another dumb question - I've been doing measurements 1 meter away pointing at tweeter for each individual speaker.
My Left speaker clearly shows some severe room modes. My right speaker is showing pretty good.
How do I integrate the REW settings into 1 Soundscape Headphone EQ file? Just put both setting together? That doesn't seem correct.

Thanks
 
@kemmler3D Finally figured out how to use the Soundscape + REW. Wonderful. Thank you for the suggestion.
Another dumb question - I've been doing measurements 1 meter away pointing at tweeter for each individual speaker.
My Left speaker clearly shows some severe room modes. My right speaker is showing pretty good.
How do I integrate the REW settings into 1 Soundscape Headphone EQ file? Just put both setting together? That doesn't seem correct.

Thanks
I actually just use one MMM (moving mic measurement) in REW and only touch the bass, so that works reasonably well as a mono correction.

I'm not sure but it might be possible to specify whether each filter is for left or right and combine them into one file... This is a guess, however.
 
AFAIK, SoundSource does not natively support separate EQ for left and right.

I'd follow @kemmler3D's advice and do a single MMM measurement capturing the combined response of L+R, then correct <~300Hz only.
 
Back
Top Bottom