• 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!

Help with REW SPL Graphs - First time user

pimogo

Member
Joined
Mar 4, 2022
Messages
18
Likes
4
Hello All -

Was looking for help in understanding what I'm seeing after following Julian Krause's excellent guide on YT.

I ran 9 measurements in and around my listening height. With those 9 measurements, i created an average w/ VAR smoothing as per the guide. The results came in (seen below).

My question relates to what is happening below 40hz. My speaker (Kanto YU6) has a frequency range of 50 Hz – 20 kHz, yet i'm seeing an odd bump happening before 40hz. What could possibly explain that? All 9 measurements exhibit that. I would expect a sharp drop off in that region.

Kanto Yu6- Average SPL.png
 

raindance

Major Contributor
Joined
Sep 25, 2019
Messages
1,037
Likes
968
Are you using a calibrated mic? It is possibly environmental noise, such as your aircon or construction or some other background rumble. Or it is an artifact of the mic.
 

HarmonicTHD

Major Contributor
Forum Donor
Joined
Mar 18, 2022
Messages
3,326
Likes
4,829
Although not having checked all the details let me try to make a guess.

The manufactures frequency specs are mostly within plus minus 3dB of some reference SPL eg 80 or more dB measured (or simulated) for an anachoic chamber.
That means that the speaker radiates energy, although on lower level (12 to 24dB/Octave less) at frequencies below the here quoted 50Hz.

Now putting that speaker in a room where the reflection and resulting interferences of the sound waves from walls, ceiling, floor lead to cancellations and superpositions can easily lead to a picture like the one you posted. These peaks and nulls (cancellations) are absolutely normal for frequencies below ca 300 Hz (Schroedinger Frequency) and are called room modes.
 
Last edited:
OP
P

pimogo

Member
Joined
Mar 4, 2022
Messages
18
Likes
4
Are you using a calibrated mic? It is possibly environmental noise, such as your aircon or construction or some other background rumble. Or it is an artifact of the mic.
Hello - I'm using a UMIK-1 with calibrated narrow_band_response_0_degree file from cross spectrum. If what you say is true (i'm hearing construction right now), are these results no longer valid? Or can i simply disregard anything below 50hz?

I was hoping to move on to a MMM (moving microphone method) of measuring, but now concerned about any potential noise i may generate.
 

HarmonicTHD

Major Contributor
Forum Donor
Joined
Mar 18, 2022
Messages
3,326
Likes
4,829
Hello - I'm using a UMIK-1 with calibrated narrow_band_response_0_degree file from cross spectrum. If what you say is true (i'm hearing construction right now), are these results no longer valid? Or can i simply disregard anything below 50hz?
See my response above. Your measurements are reasonably valid, except that you should use the 90deg calibration file when pointing your UMIK to the ceiling for the REW measurements.
 
OP
P

pimogo

Member
Joined
Mar 4, 2022
Messages
18
Likes
4
See my response above. Your measurements are reasonably valid, except that you should use the 90deg calibration file when pointing your UMIK to the ceiling for the REW measurements.
I was pointing towards the speaker. I read for 2 channel applications this was the preferred method.

When generating the PEQ, should i crop the frequency at 50hz? Otherwise this tends to happen:

Filter 1: ON PK Fc 20.00 Hz Gain -53.20 dB Q 2.000
Filter 2: ON PK Fc 121.0 Hz Gain -4.00 dB Q 27.246
Filter 3: ON PK Fc 144.5 Hz Gain -7.80 dB Q 4.851
Filter 4: ON PK Fc 205.0 Hz Gain -7.10 dB Q 4.058
Filter 5: ON PK Fc 243.0 Hz Gain 6.00 dB Q 1.052
Filter 6: ON PK Fc 325.0 Hz Gain -5.20 dB Q 3.534

(the house curve was obtained from the guide above)

That 20hz gain of -53 db seems suspect, so too does the Q of 27 at 121 hz.

Sorry still much to learn here.
 

HarmonicTHD

Major Contributor
Forum Donor
Joined
Mar 18, 2022
Messages
3,326
Likes
4,829
I was pointing towards the speaker. I read for 2 channel applications this was the preferred method.

When generating the PEQ, should i crop the frequency at 50hz? Otherwise this tends to happen:

Filter 1: ON PK Fc 20.00 Hz Gain -53.20 dB Q 2.000
Filter 2: ON PK Fc 121.0 Hz Gain -4.00 dB Q 27.246
Filter 3: ON PK Fc 144.5 Hz Gain -7.80 dB Q 4.851
Filter 4: ON PK Fc 205.0 Hz Gain -7.10 dB Q 4.058
Filter 5: ON PK Fc 243.0 Hz Gain 6.00 dB Q 1.052
Filter 6: ON PK Fc 325.0 Hz Gain -5.20 dB Q 3.534

(the house curve was obtained from the guide above)

That 20hz gain of -53 db seems suspect, so too does the Q of 27 at 121 hz.

Sorry still much to learn here.
Hi.
No for in room frequency responses and PEQ derivation, always point to the ceiling. That should be mentioned in most online REW manuals. However I would guess the error for these low frequencies is not too large when not pointing at the ceiling. But just redo it to be safe.

As for the PEQ it depends at what average target SPL you are aiming for. Let‘s take the 63dB from your graph above (although I would recommend around measuring at ca 80dB). You would get a smoother frequency response if you lower the peak at ca 15Hz,68dB to the average 63dB. Especially attenuating these very low frequency room induced peaks often lead to a cleaner, less boomy bass and avoids exciting any vibrations in the room eg furniture etc. So therefore I would not cut off your PEQ filter generation at 50Hz.

To check the plausibility of your particular values. You would need to post the full PEQ config page and resulting graphs.
 
Last edited:
OP
P

pimogo

Member
Joined
Mar 4, 2022
Messages
18
Likes
4
Hi.
No for in room frequency responses and PEQ derivation, always point to the ceiling. That should be mentioned in most online REW manuals. However I would guess the error for these low frequencies is not too large when not pointing at the ceiling. But just redo it to be safe.

As for the PEQ it depends at what average target SPL you are aiming for. Let‘s take the 63dB from your graph above (although I would recommend around measuring at ca 80dB). You would get a smoother frequency response if you lower the peak at ca 15Hz,68dB to the average 63dB. Especially attenuating these very low frequency room induced peaks often lead to a cleaner, less boomy bass and avoids exciting any vibrations in the room eg furniture etc. So therefore I would not cut off your PEQ filter generation at 50Hz.

To check the plausibility of your particular values. You would need to post the full PEQ config page and resulting graphs.
Thank you! Will experiment further with 1) upright placement of mic 2) measuring at ~ 80 dB and 3) not cropping at 50hz.

My concern was that those lower frequencies were being introduced perhaps by the computer fan standing a couple feet from the mic..

I'll report back later with results. thanks!
 

HarmonicTHD

Major Contributor
Forum Donor
Joined
Mar 18, 2022
Messages
3,326
Likes
4,829
Thank you! Will experiment further with 1) upright placement of mic 2) measuring at ~ 80 dB and 3) not cropping at 50hz.

My concern was that those lower frequencies were being introduced perhaps by the computer fan standing a couple feet from the mic..

I'll report back later with results. thanks!
Way to go. Good luck.

By measuring at eg 80dB SPL the signal to noise ratio is most likely high enough to capture only the room and not the computer fan (unless it is very loud ). Sure, if you can, switch the fan control to Silent mode in your Bios for the time you do the measurements.
 

raindance

Major Contributor
Joined
Sep 25, 2019
Messages
1,037
Likes
968
Don't waste your time EQing below what is physically possible, either add a high pass at, say 30Hz, fix the peaks add a slope and enjoy. Fixing the valleys will waste power and limit volume.
 
OP
P

pimogo

Member
Joined
Mar 4, 2022
Messages
18
Likes
4
Way to go. Good luck.

By measuring at eg 80dB SPL the signal to noise ratio is most likely high enough to capture only the room and not the computer fan (unless it is very loud ). Sure, if you can, switch the fan control to Silent mode in your Bios for the time you do the measurements.
So i tried measuring using the 90 degree file with upright microphone, but the results appear strange. I tried numerous times using both the minidsp provided file as well as the cross spectrum file: both yielded the below. Aside from providing a new calibration file, are there any settings i have to change in REW to accommodate the upright mic?

90 degree - upright mic

90degree.jpg


I then switched over to the 0 degree cal file with mic pointed to speaker--that yielded a much more pleasant image.

flat mic.jpg


I consulted Minidsp and they suggest the following:

Which calibration file should I use and where to point the UMIK-1?​

We provide two calibration files to be used depending on your application.

- For stereo system (e.g. 2ch dirac live, single speaker measurement), use the 0deg file and point the UMIK-1 at the speakers
- For multichannel system (E.g. 5.1/7.1) or a surround application where multiple speakers are spreadout around the room, use the 90deg file and point the UMIK-1 at the ceiling.


Regardless, even if that were the case, why would my 90 degree measurement yield such wildly different results??
 

HarmonicTHD

Major Contributor
Forum Donor
Joined
Mar 18, 2022
Messages
3,326
Likes
4,829
So i tried measuring using the 90 degree file with upright microphone, but the results appear strange. I tried numerous times using both the minidsp provided file as well as the cross spectrum file: both yielded the below. Aside from providing a new calibration file, are there any settings i have to change in REW to accommodate the upright mic?

90 degree - upright mic

View attachment 204273

I then switched over to the 0 degree cal file with mic pointed to speaker--that yielded a much more pleasant image.

View attachment 204274

I consulted Minidsp and they suggest the following:

Which calibration file should I use and where to point the UMIK-1?​

We provide two calibration files to be used depending on your application.

- For stereo system (e.g. 2ch dirac live, single speaker measurement), use the 0deg file and point the UMIK-1 at the speakers
- For multichannel system (E.g. 5.1/7.1) or a surround application where multiple speakers are spreadout around the room, use the 90deg file and point the UMIK-1 at the ceiling.


Regardless, even if that were the case, why would my 90 degree measurement yield such wildly different results??
As rednaxela said. Also my guess. Something is wrong. If you like, sent a screenshot of your REW preference page, so people can have a look and help.

But the second graph looks really good already. Well done. How does it sound?
 
OP
P

pimogo

Member
Joined
Mar 4, 2022
Messages
18
Likes
4
Yeah i must have done something incorrectly. I reviewed all and reran the 90 degree measurements. I get something substantially better. Now remaining questions i have before i can properly generate the PEQ filters.
  1. Is there anything in my settings below that ought to be altered?
    • My speakers are bookshelves, hence why i selected bass limited speaker.
    • I kept the defaults for bass management at 80 hz. -- seems to follow the curve ok.
    • I picked 24 db/octave because i have a rear port.
  2. I included a house curve in preferences so i wasn't sure if should click on 'add room curve'. I did as per the guide i followed and i adopted his settings. Any opportunities there.
  3. I let rew calculate the target level -- it settled on 80.4 dB.
  4. Filter tasks.
    • I followed the guide here but opted for a revised range of 30-20kHz

1651615902993.png


If i generate the EQ filters, i get the following:
  1. 17 filters seems an awful lot. Any opportunities here to simplify?

1651616200418.png



My next step is to plug these into Roon. Shoud i plug these as they are. Its not immediately clear what my headroom should be, but my guess would be at least 6dB to account for the largest gain. Is this correct?

Many thanks for any and all help. Looking forward to auditioning my very own PEQ. :)
 

HarmonicTHD

Major Contributor
Forum Donor
Joined
Mar 18, 2022
Messages
3,326
Likes
4,829
Yeah i must have done something incorrectly. I reviewed all and reran the 90 degree measurements. I get something substantially better. Now remaining questions i have before i can properly generate the PEQ filters.
  1. Is there anything in my settings below that ought to be altered?
    • My speakers are bookshelves, hence why i selected bass limited speaker.
    • I kept the defaults for bass management at 80 hz. -- seems to follow the curve ok.
    • I picked 24 db/octave because i have a rear port.
  2. I included a house curve in preferences so i wasn't sure if should click on 'add room curve'. I did as per the guide i followed and i adopted his settings. Any opportunities there.
  3. I let rew calculate the target level -- it settled on 80.4 dB.
  4. Filter tasks.
    • I followed the guide here but opted for a revised range of 30-20kHz

View attachment 204312

If i generate the EQ filters, i get the following:
  1. 17 filters seems an awful lot. Any opportunities here to simplify?

View attachment 204314


My next step is to plug these into Roon. Shoud i plug these as they are. Its not immediately clear what my headroom should be, but my guess would be at least 6dB to account for the largest gain. Is this correct?

Many thanks for any and all help. Looking forward to auditioning my very own PEQ. :)
Just having a very very quick look at your graph (but not having looked in the details as it gets late here) - I would say give it shot.

However, if you give a Like to JohnPM, the previous poster, he might come back and help you as he is the inventor of REW and I wouldn’t even dare to give you advice from this point onward ;):)
 

Erici

Active Member
Forum Donor
Joined
Jun 17, 2020
Messages
176
Likes
258
Location
Seattle Area
Hey pimago. You're picking things up quickly here! If you change your flatness target from 1 to 3 dB you'll drastically cut back on the number of filters. And any filter with a Q over 10 will not be very effective, so you might eliminate them also.
 

Rednaxela

Major Contributor
Joined
Mar 30, 2022
Messages
2,052
Likes
2,675
Location
NL
I included a house curve in preferences so i wasn't sure if should click on 'add room curve'. I did as per the guide i followed and i adopted his settings. Any opportunities there.
Personally I don’t use the house curve function, just the room curve. Not sure what happens if you use both.

There are arguments against correcting room response full range. Instead you may want to correct your room response only in the bass region, and focus there on the rolloff and the big peaks. This should clean up your bass nicely, but the end result may feel a bit bass light. You can use the LF Rise controls in the room curve section (like you do already), to compensate and fine tune.

Your HF Fall looks rather nice already so I would leave that part uncorrected and let things fall where they fall. At least at first. You may be able to further fine tune the HF section by only correcting your speaker’s anechoic response there. This depends a bit on your speakers’ characteristics and of course the availability of anechoic or quasi-anechoic measurement data.

The dip around 600 Hz could be placement related and hard to EQ. See what you can achieve there but careful not to throw too much power at it.

Overall I’ve had good results with this strategy. FWIW it will probably also address your concern re. the number of filters applied.

HTH!
 
OP
P

pimogo

Member
Joined
Mar 4, 2022
Messages
18
Likes
4
Just having a very very quick look at your graph (but not having looked in the details as it gets late here) - I would say give it shot.

However, if you give a Like to JohnPM, the previous poster, he might come back and help you as he is the inventor of REW and I wouldn’t even dare to give you advice from this point onward ;):)
Ha! Thank you for all your help thus far!
The dip around 600 Hz could be placement related and hard to EQ. See what you can achieve there but careful not to throw too much power at it.
This is interesting and something i wish to understand better. Why is the 600 hz dip placement related?
 

Rednaxela

Major Contributor
Joined
Mar 30, 2022
Messages
2,052
Likes
2,675
Location
NL
This is interesting and something i wish to understand better. Why is the 600 hz dip placement related?
For all clarity: it was only a guess.

Here is a nice resource that explains this much better than I could. You may be able to confirm this yourself with the presented quarter wavelength cancellation frequency formula. With speakers in a corner you sometimes have this effect adding up multiple times, from the back wall, side wall, and floor, obviously depending on their respective distances to the speaker.

HTH!
 
Top Bottom