Room Mode + Speaker Anechoic Measurement Correction

[lempogi]

Hi all. I'm a silent reader in this forum and I'm learning a lot especially from @dominikz and @harkpabst.
I just bought a Wiim Amp Ultra and I would like to use Room Fit for my budget Polk XT20 speaker and SVS SB-1000 Pro subwoofer.
I would like to have your advice if my plan is in the right track or not. Or if you have an additional or better idea or tips I'll appreciate it so much.

1. Do a Room Fit correction for 20-400 Hz.
2. Pick the most needed frequencies to be corrected for 5 slots of PEQ. (If you can teach me how to compute or pick from the 10 suggested correction by Room Fit)
3. Add the other 5 from corrected frequencies in Spinorama AutomaticEQ (714 Hz to 11882 Hz).

Also, there's a positive gain adjustment for 714 Hz and 8694 Hz. I'm thinking that it will messed up the recommended gain adjustment from Spinorama if I will mix the Room Fit correction. What I also learned here is that I can use REW to manually input those mixed 10 frequency corrections and it will show the new computed gain adjustment. Am I correct? I'm sorry this is my first time for room correction and I have also no experience on REW software.

Thanks in advance!

 

[Burnside]

Re REW and computed headroom, see this and the following post
https://forum.wiimhome.com/threads/peq-talk.1500/post-27059

 

[lempogi]

Re REW and computed headroom, see this and the following post
https://forum.wiimhome.com/threads/peq-talk.1500/post-27059

I'll read that one, I appreciate this much!

 

[dominikz]

1. Do a Room Fit correction for 20-400 Hz.

Sounds good! Make sure to optimize loudspeaker placement and subwoofer integration before running RoomFit. IMHO REW is really helpful for this!

2. Pick the most needed frequencies to be corrected for 5 slots of PEQ. (If you can teach me how to compute or pick from the 10 suggested correction by Room Fit)
3. Add the other 5 from corrected frequencies in Spinorama AutomaticEQ (714 Hz to 11882 Hz).

There's no need to do this anymore, since recently you can use both EQ and RoomFit at the same time (so you get 10 bands of PEQ per channel in "RoomFit" plus an extra 10 bands of PEQ per channel in "EQ").
You can therefore use RoomFit to correct the room resonances, and in addition to that you can configure the Spinorama AutomaticEQ filter parameters into EQ (but only configure filters above 400Hz where RoomFit no longer applies) - when enabled both sets of filters will be applied at the same time!

Also, there's a positive gain adjustment for 714 Hz and 8694 Hz. I'm thinking that it will messed up the recommended gain adjustment from Spinorama if I will mix the Room Fit correction. What I also learned here is that I can use REW to manually input those mixed 10 frequency corrections and it will show the new computed gain adjustment. Am I correct? I'm sorry this is my first time for room correction and I have also no experience on REW software.

That is correct, but note that if you use RoomFit in the new "No-Boost Mode" (meaning no pregain adjustment needed) you could then just use the preamp gain value from Spinorama AutomaticEQ directly in WiiM Input pregain view (which is already given in dB), without additional calculations in REW.

But to be honest I wouldn't worry about pregain compensation with WiiM Amp Ultra unless if you plan to run it at full volume (i.e. output volume set to 100). At lower output volume levels you are anyway not reaching 0 dBFS so there's no fear of digital clipping.

BTW that is a nice system you built - good luck and enjoy!

 

[lempogi]

Sounds good! Make sure to optimize loudspeaker placement and subwoofer integration before running RoomFit. IMHO REW is really helpful for this!

Yes sir! 5cm/2in from the wall for speakers and corner placement for the subwoofer. My wife is laughing at me when I'm doing the measurement but I said to her that it's scientifically proven.

There's no need to do this anymore, since recently you can use both EQ and RoomFit at the same time (so you get 10 bands of PEQ per channel in "RoomFit" plus an extra 10 bands of PEQ per channel in "EQ").
You can therefore use RoomFit to correct the room resonances, and in addition to that you can configure the Spinorama AutomaticEQ filter parameters into EQ (but only configure filters above 400Hz where RoomFit no longer applies) - when enabled both sets of filters will be applied at the same time!

Really? That's a good news to me then. I thought what Wiim did is you can combine only the RoomFit plus GEQ and not the EQ. If only Wiim can have a better documentation for this so that beginners like me can easily understand the new updates.

That is correct, but note that if you use RoomFit in the new "No-Boost Mode" (meaning no pregain adjustment needed) you could then just use the preamp gain value from Spinorama AutomaticEQ directly in WiiM Input pregain view (which is already given in dB), without additional calculations in REW.

But to be honest I wouldn't worry about pregain compensation with WiiM Amp Ultra unless if you plan to run it at full volume (i.e. output volume set to 100). At lower output volume levels you are anyway not reaching 0 dBFS so there's no fear of digital clipping.

Good point. If I'm using the HDMI input for TV, the max volume we only use is up to 45. For Wifi/BT/USB is around 35 max.

BTW that is a nice system you built - good luck and enjoy!

I'm not a rich guy and also it is very hard to request my wife to have an expensive speakers . So I have to research for a good budget one, and Erin recommends the Polk XT20, just have to correct the high frequencies of it. For the subwoofer, I'm initially targeting Speedwoofer 10E but it's out of stock for a long time. I found an open box SB-1000 Pro so I just bought that one. I'm happy for what I have so far!

 

[dominikz]

Yes sir! 5cm/2in from the wall for speakers and corner placement for the subwoofer. My wife is laughing at me when I'm doing the measurement but I said to her that it's scientifically proven.

Really? That's a good news to me then. I thought what Wiim did is you can combine only the RoomFit plus GEQ and not the EQ. If only Wiim can have a better documentation for this so that beginners like me can easily understand the new updates.

Both GEQ and PEQ variants work, but it is true that WiiM documentation is often not very clear about such things

Good point. If I'm using the HDMI input for TV, the max volume we only use is up to 45. For Wifi/BT/USB is around 35 max.

In that case I wouldn't bother with pregain compensation for EQ presets at all, if I were in your shoes.

I'm not a rich guy and also it is very hard to request my wife to have an expensive speakers . So I have to research for a good budget one, and Erin recommends the Polk XT20, just have to correct the high frequencies of it. For the subwoofer, I'm initially targeting Speedwoofer 10E but it's out of stock for a long time. I found an open box SB-100 Pro so I just bought that one. I'm happy for what I have so far!

IMHO you made really smart choices, all really good performing gear in general - but especially so for the price. I'm quite sure it will sound wonderful once you finish tuning the system!

 

[lempogi]

Both GEQ and PEQ variants work, but it is true that WiiM documentation is often not very clear about such things

In that case I wouldn't bother with pregain compensation for EQ presets at all, if I were in your shoes.

IMHO you made really smart choices, all really good performing gear in general - but especially so for the price. I'm quite sure it will sound wonderful once you finish tuning the system!

Well, thank you so much for guiding me on this one. I'll update you what's the result of the corrections. Thanks also @Burnside for that thread link. I'll still read it for additional learnings.

 

[Kidprofessor]

I run both Spinorama and correction from REW for frequencies below the Schroeder frequency (if you don't know how to calculate this I can provide links). It has to be done in a particular order and a particular way for it to work properly with the Wiim.

The first thing to do is to enter the values you have for Auto EQ from the Spinorama data that you have pictured. You may want to ignore the two lowest frequencies, because they will be lower than the Schroeder frequency in your room and they will be corrected anyway. Also I don't enter Spinorama EQ below the crossover to my sub, because it obviously doesn't apply to my sub. I use the EQ section for this.

Now comes the hard part, if you use Room Fit it will immediately turn off your EQ, so instead of Room Fit working with the Spinorama data, it will just ignore it, so you will end up with Spinorama over Room Fit, instead of the other way around, which is usually not ideal. Room fit for bass correction also has worse results than recommended REW figures.

What I do is use REW instead. First thing that is needed to do is to run Room Fit so that you have a space to put the recommended EQ's. However, Room Fit messes with the sub levels in a way that you won't see in the PEQs if you use Subwoofer Calibration. So I recommend turning that off. Once you run Room Fit you have a space to put your REW corrections for bass.

Then you turn on your Spinorama EQ in the EQ section with the Room Fit off. I use the MMM method on REW to measure bass. Here is a link on how to do this:

The moving microphone method (MMM) for Dummies using REW

After getting a couple of weird results on my first MMM measurements, I went searching around the forums & YouTube for what went wrong. To save others from gathering various bits and pieces of info from a dozen sources, I decided to publish this quick write-up here. You can use it as a rough...

www.audiosciencereview.com

The measurement I get with the Spinorama on but Room Fit off looks like this:



Then I generate filters for knocking down the modes using the EQ tool in REW. I set the frequency range below 200HZ, because the Schroeder frequency in my room is about 150HZ. I set it to create negative filters only. This is part of the reason REW is better than Room Fit, which will add positive filters no matter what. I then enter the negative filters into the Room Fit profile I created before, turning off all the generated ones from Room fit first and substituting the REW filters. I also added a tilt to lower the high end. The filter is -4db at .7Q at 1.5khz High Shelf filter. This allows it to match the house curve I was using, which was the Trained Listeners curve from Floyd Toole's research.

The result I get is here:


This sounds amazing in my room, with very tight and accurate bass. It helps that the Speakers I am using are the Affordable Accuracy Monitors by Dennis Murphy, but your Polk XT and sub setup should be able to get this. It also helps that I have good placement of my sub and speakers to avoid any wideband nulls in the bass. You can also use a Harman curve or any other curve that you prefer to knock down the bass peaks. I picked this curve because it was already close to what I measured.

The very narrow band nulls that you can see in the second measurement cannot be heard. Floyd Toole's research has shown that listeners are very sensitive to narrow peaks in bass frequencies, but narrow nulls are ignored. Humans can fill in nulls in our brains because of the harmonics created by low frequency sounds. The peaks are perceived as "boomy" and we can't filter them out with our brains, we need EQ.

Using only negative filters works much, much better than the way Room Fit does it. As Floyd Toole also says, room correction software is created by people who don't quite understand psychoacoustics research. Room Fit is better than most, but the method I am showing here more closely follows the research of how people hear.

Let me know if you have questions!

 

[dominikz]

I run both Spinorama and correction from REW for frequencies below the Schroeder frequency (if you don't know how to calculate this I can provide links). It has to be done in a particular order and a particular way for it to work properly with the Wiim.

The first thing to do is to enter the values you have for Auto EQ from the Spinorama data that you have pictured. You may want to ignore the two lowest frequencies, because they will be lower than the Schroeder frequency in your room and they will be corrected anyway. Also I don't enter Spinorama EQ below the crossover to my sub, because it obviously doesn't apply to my sub. I use the EQ section for this.

Now comes the hard part, if you use Room Fit it will immediately turn off your EQ, so instead of Room Fit working with the Spinorama data, it will just ignore it, so you will end up with Spinorama over Room Fit, instead of the other way around, which is usually not ideal. Room fit for bass correction also has worse results than recommended REW figures.

What I do is use REW instead. First thing that is needed to do is to run Room Fit so that you have a space to put the recommended EQ's. However, Room Fit messes with the sub levels in a way that you won't see in the PEQs if you use Subwoofer Calibration. So I recommend turning that off. Once you run Room Fit you have a space to put your REW corrections for bass.

Then you turn on your Spinorama EQ in the EQ section with the Room Fit off. I use the MMM method on REW to measure bass. Here is a link on how to do this:

The moving microphone method (MMM) for Dummies using REW

After getting a couple of weird results on my first MMM measurements, I went searching around the forums & YouTube for what went wrong. To save others from gathering various bits and pieces of info from a dozen sources, I decided to publish this quick write-up here. You can use it as a rough...

www.audiosciencereview.com

The measurement I get with the Spinorama on but Room Fit off looks like this:

View attachment 25915

Then I generate filters for knocking down the modes using the EQ tool in REW. I set the frequency range below 200HZ, because the Schroeder frequency in my room is about 150HZ. I set it to create negative filters only. This is part of the reason REW is better than Room Fit, which will add positive filters no matter what. I then enter the negative filters into the Room Fit profile I created before, turning off all the generated ones from Room fit first and substituting the REW filters. I also added a tilt to lower the high end. The filter is -4db at .7Q at 1.5khz High Shelf filter. This allows it to match the house curve I was using, which was the Trained Listeners curve from Floyd Toole's research.

The result I get is here:

View attachment 25916
This sounds amazing in my room, with very tight and accurate bass. It helps that the Speakers I am using are the Affordable Accuracy Monitors by Dennis Murphy, but your Polk XT and sub setup should be able to get this. It also helps that I have good placement of my sub and speakers to avoid any wideband nulls in the bass. You can also use a Harman curve or any other curve that you prefer to knock down the bass peaks. I picked this curve because it was already close to what I measured.

The very narrow band nulls that you can see in the second measurement cannot be heard. Floyd Toole's research has shown that listeners are very sensitive to narrow peaks in bass frequencies, but narrow nulls are ignored. Humans can fill in nulls in our brains because of the harmonics created by low frequency sounds. The peaks are perceived as "boomy" and we can't filter them out with our brains, we need EQ.

Using only negative filters works much, much better than the way Room Fit does it. As Floyd Toole also says, room correction software is created by people who don't quite understand psychoacoustics research. Room Fit is better than most, but the method I am showing here more closely follows the research of how people hear.

Let me know if you have questions!

I've used a similar approach for years and I fully agree it is technically the most sound one, and the most flexible.

However, I feel the need to comment that with the recent introduction of RoomFit "No-Boost Mode" one can now get reasonably similar results with RoomFit.

First, you are absolutely correct that even with "No-Boost Mode" enabled RoomFit will still use a few individual filters with positive gain to shape the overall response. But since the total correction filter response with this mode remains at or below 0dB, it won't/can't introduce any new resonances to the system. So IMHO this response shaping strategy can be viewed as an inefficiency of RoomFit filter calculation algorithms rather than something audibly problematic.

IME the differences in corrections between RoomFit and REW are now mainly in flexibility and precision - RoomFit is limited by the fact that the least amount of smoothing you can use is 1/12 (while in REW you can use even less), and the maximum Q used by RoomFit filters is 10 (while in REW it can be higher).

A note on measurement method here: in REW you can use various kinds of measurement methods to measure your system response while in RoomFit you are more limited. With REW I usually use MMM (as it is very repeatable and implicitly adds some desirable spatial smoothing). MMM implementation in RoomFit is IMHO not ideal (see here for more info), and the Multiple Measurement options has a long standing bug (sweep usually doesn't play successfully all 3 times for both channels). So in effect, at the moment IMO it is best to use the basic single-sweep method in RoomFit. Luckily, there shouldn't really be much difference in measured response below 200Hz between single sweep measurement and MMM, so for this purpose it should not be an issue.​

This means that the system response after automatic correction calculated by RoomFit (with "No-Boost Mode" & 1/12 smoothing) will be slightly less smooth and precise than the one from REW (with max individual and total gain set to 0dB and Var smoothing) - assuming we used the same single sweep measurement, the same correction frequency range, and the same target curve in both.

This is all of course assuming optimal RoomFit configuration, use of a calibrated measurement microphone, and a mobile phone that doesn't change the measured response (see here for one example of this strange issue in some phones).

However, human hearing has some limits as well, and I wouldn't bet that these minor differences in correction precision would be very significant to most listeners under blind and level-matched conditions. This is something to consider at least.

Lastly a practical concern: at the moment there is still no import function in WiiM EQ and RoomFit, so any externally calculated EQs unfortunately needs to be input manually. This is quite tedious to do, especially if you use both the individual channel RoomFit (2x10 band of PEQ with 3 attributes each) plus stereo EQ for loudspeaker correction based on spinorama data (up to 1x10 band PEQ with 3 attributes each). That gives a total of 90 numbers to type into the WiiM Home App for a correction you want to apply manually (and you may wish to try different profiles so it can multiply)! Fortunately WiiM has already hinted that import functionality is on their radar (see here) - and I do hope it will come soon! That would really make it much easier to use REW to create custom corrections for WiiM devices.

In summary, while I'd definitely suggest to anyone interested in audio to learn how to use REW, IMHO comparably good results can now be achieved with RoomFit for those unwilling to go quite as deep down the rabbit hole!

 

[cristoo]

Hi,

Apology for stepping into discussion - I don’t want to hijack OP topic - but if I may take opportunity of this interesting discussion.

I have Elac debut 52 speakers and on recent WHA and Ultra firmware I did RoomFit in range 20-400 in No Boost mode with subwoofer calibration. I used BK curve.
Now for me bass is less boomy and more controlled so I am satisfied (at least for now). But having option to add another eq layer with EQ function should I apply another filters like taken from internet measurement - like to compensate Elac’ weakness above 400hz ? I find some high tones like cymbals in rock/metal music bit harsh but maybe because particular song is recorded that way.
I found on Erin’s Corner Elac DB 52 review and if I understood correctly he recommend at 1.6 kHz -2b HS filter. When I apply this in Eq (as single additional filter) sound in less harsh but noticeably darker and less airy so i prefer to turn it off.

Bottom line : after RoomFit application should one add additional filters over RF correction and if so how to calculate them.

Thanks in advance.

 

[Kidprofessor]

I've used a similar approach for years and I fully agree it is technically the most sound one, and the most flexible.

However, I feel the need to comment that with the recent introduction of RoomFit "No-Boost Mode" one can now get reasonably similar results with RoomFit.

First, you are absolutely correct that even with "No-Boost Mode" enabled RoomFit will still use a few individual filters with positive gain to shape the overall response. But since the total correction filter response with this mode remains at or below 0dB, it won't/can't introduce any new resonances to the system. So IMHO this response shaping strategy can be viewed as an inefficiency of RoomFit filter calculation algorithms rather than something audibly problematic.

IME the differences in corrections between RoomFit and REW are now mainly in flexibility and precision - RoomFit is limited by the fact that the least amount of smoothing you can use is 1/12 (while in REW you can use even less), and the maximum Q used by RoomFit filters is 10 (while in REW it can be higher).

A note on measurement method here: in REW you can use various kinds of measurement methods to measure your system response while in RoomFit you are more limited. With REW I usually use MMM (as it is very repeatable and implicitly adds some desirable spatial smoothing). MMM implementation in RoomFit is IMHO not ideal (see here for more info), and the Multiple Measurement options has a long standing bug (sweep usually doesn't play successfully all 3 times for both channels). So in effect, at the moment IMO it is best to use the basic single-sweep method in RoomFit. Luckily, there shouldn't really be much difference in measured response below 200Hz between single sweep measurement and MMM, so for this purpose it should not be an issue.​

This means that the system response after automatic correction calculated by RoomFit (with "No-Boost Mode" & 1/12 smoothing) will be slightly less smooth and precise than the one from REW (with max individual and total gain set to 0dB and Var smoothing) - assuming we used the same single sweep measurement, the same correction frequency range, and the same target curve in both.

This is all of course assuming optimal RoomFit configuration, use of a calibrated measurement microphone, and a mobile phone that doesn't change the measured response (see here for one example of this strange issue in some phones).

However, human hearing has some limits as well, and I wouldn't bet that these minor differences in correction precision would be very significant to most listeners under blind and level-matched conditions. This is something to consider at least.

Lastly a practical concern: at the moment there is still no import function in WiiM EQ and RoomFit, so any externally calculated EQs unfortunately needs to be input manually. This is quite tedious to do, especially if you use both the individual channel RoomFit (2x10 band of PEQ with 3 attributes each) plus stereo EQ for loudspeaker correction based on spinorama data (up to 1x10 band PEQ with 3 attributes each). That gives a total of 90 numbers to type into the WiiM Home App for a correction you want to apply manually (and you may wish to try different profiles so it can multiply)! Fortunately WiiM has already hinted that import functionality is on their radar (see here) - and I do hope it will come soon! That would really make it much easier to use REW to create custom corrections for WiiM devices.

In summary, while I'd definitely suggest to anyone interested in audio to learn how to use REW, IMHO comparably good results can now be achieved with RoomFit for those unwilling to go quite as deep down the rabbit hole!

I was wondering if I would be able to blind test identify the REW eq vs the Room Fit one, but I don't have a good way to do that at the moment. My "much, much better" assessment is something I should know better not to trust without blind testing. .

As a general principle, fewer eqs and fewer boosts are better, especially if starting with speakers that have a smooth response. You're almost certainly correct that with no-boost mode activated in Room Fit that no new resonances are created, but the measurements and my ears tell me that the low end isn't quite as smooth with Room Fit. This is probably due to differences in the algorithms used to generate the filters. The Room Fit filters always fill up all 10 available bands for some reason, and I would guess that's an inherent problem, but can't be sure. But yeah, blind testing would be needed.

Room Fit below the Schroeder Frequency using no-boost mode with a calibrated mic would produce results that most people will be very happy with.

I forgot to mention I used variable smoothing in REW for the graphs shown, but originally used 1/48 for the MMM measurement. Was trying not to overload the OP.

Part of the reason I like the REW method is that I can use any curve I want, and the Trained Listeners curve just happens to sound good on almost everything I listen to. The Affordable Accuracy Monitors have a wonderful midrange that is highlighted by this curve. I also did a Harman Curve using the same method and felt like it was a bit too "boom and tizz" for my taste. And I love bassy music. I probably like a "darker" and "warmer" sound than most people though.

What I would really like to see from Wiim would be 20 bands of PEQ in Room Fit (purely to add Spinorama EQ, few filters are needed below the Schroeder frequency), and simple bass and treble controls in the EQ section. Just two sliders with bass and treble to be able to adjust for old recordings that tend to have little bass and too much treble.

 

[Kidprofessor]

Hi,

Apology for stepping into discussion - I don’t want to hijack OP topic - but if I may take opportunity of this interesting discussion.

I have Elac debut 52 speakers and on recent WHA and Ultra firmware I did RoomFit in range 20-400 in No Boost mode with subwoofer calibration. I used BK curve.
Now for me bass is less boomy and more controlled so I am satisfied (at least for now). But having option to add another eq layer with EQ function should I apply another filters like taken from internet measurement - like to compensate Elac’ weakness above 400hz ? I find some high tones like cymbals in rock/metal music bit harsh but maybe because particular song is recorded that way.
I found on Erin’s Corner Elac DB 52 review and if I understood correctly he recommend at 1.6 kHz -2b HS filter. When I apply this in Eq (as single additional filter) sound in less harsh but noticeably darker and less airy so i prefer to turn it off.

Bottom line : after RoomFit application should one add additional filters over RF correction and if so how to calculate them.

Thanks in advance.

The Q value on your HS filter can make a big difference. Something around .7-.8 will get you a filter that is similar to standard bass and treble controls, but an even lower Q may work for you. I would recommend playing with the the frequency controls, maybe higher or lower than 1.6 will work. You could set the gain by ear. I came to my treble tilt by setting the gain by ear so I could hear the cymbals but it wasn't harsh. I'm 52 though, so my high end hearing is typically rolled off.

BK curve is pretty flat and low level through the bass for most people. Having less bass will automatically create the perception that there is more high midrange and treble, so you may want to add a low shelf filter at about 100HZ, Q of .7 or so, and up the gain slightly and see if that helps with the perception of harsh highs. I wish Wiim would offer multiple Harman curves with different bass levels like Dirac live.

Can we have simple bass and treble controls please Wiim?

 

[slartibartfast]

I was wondering if I would be able to blind test identify the REW eq vs the Room Fit one, but I don't have a good way to do that at the moment. My "much, much better" assessment is something I should know better not to trust without blind testing. .

As a general principle, fewer eqs and fewer boosts are better, especially if starting with speakers that have a smooth response. You're almost certainly correct that with no-boost mode activated in Room Fit that no new resonances are created, but the measurements and my ears tell me that the low end isn't quite as smooth with Room Fit. This is probably due to differences in the algorithms used to generate the filters. The Room Fit filters always fill up all 10 available bands for some reason, and I would guess that's an inherent problem, but can't be sure. But yeah, blind testing would be needed.

Room Fit below the Schroeder Frequency using no-boost mode with a calibrated mic would produce results that most people will be very happy with.

I forgot to mention I used variable smoothing in REW for the graphs shown, but originally used 1/48 for the MMM measurement. Was trying not to overload the OP.

Part of the reason I like the REW method is that I can use any curve I want, and the Trained Listeners curve just happens to sound good on almost everything I listen to. The Affordable Accuracy Monitors have a wonderful midrange that is highlighted by this curve. I also did a Harman Curve using the same method and felt like it was a bit too "boom and tizz" for my taste. And I love bassy music. I probably like a "darker" and "warmer" sound than most people though.

What I would really like to see from Wiim would be 20 bands of PEQ in Room Fit (purely to add Spinorama EQ, few filters are needed below the Schroeder frequency), and simple bass and treble controls in the EQ section. Just two sliders with bass and treble to be able to adjust for old recordings that tend to have little bass and too much treble.

Are you using stereo or individual channel Room Correction?

 

[cristoo]

The Q value on your HS filter can make a big difference. Something around .7-.8 will get you a filter that is similar to standard bass and treble controls, but an even lower Q may work for you. I would recommend playing with the the frequency controls, maybe higher or lower than 1.6 will work. You could set the gain by ear. I came to my treble tilt by setting the gain by ear so I could hear the cymbals but it wasn't harsh. I'm 52 though, so my high end hearing is typically rolled off.

BK curve is pretty flat and low level through the bass for most people. Having less bass will automatically create the perception that there is more high midrange and treble, so you may want to add a low shelf filter at about 100HZ, Q of .7 or so, and up the gain slightly and see if that helps with the perception of harsh highs. I wish Wiim would offer multiple Harman curves with different bass levels like Dirac live.

Can we have simple bass and treble controls please Wiim?

Thanks for your reply and advices.

As far as I remember (I am out of home right now) the extra filter that I mentioned is HS 1.6kHz, gain -1.5 or - 2db ( I experimented a bit) and Q 2.0. As I wrote I is purely taken from Erin’s review and entered manually into Wiim Eq just for testing and bit blindly.

I didn’t change RF settings just simply provided above eq setting but as trebles become less noticeable and less harsh, i found overall sound quality worse because of less air and overall too much reduction of mids and trebles, on the other hand according to Erin this tweak should provide more balance sound for Elac.

Maybe I am simply stuck in rabbit hole of perusing of perfect sound in budget friendly setup but recent Wiim options (like joint EQ and RF) give another opportunities that I want to utilise.

Thank for your time and all good advices. Highly appreciated.

 

[Kidprofessor]

Are you using stereo or individual channel Room Correction?

Stereo, only because I didn't get good results with individual channel. In theory with my unbalanced room the latter should be better, but stereo produced smoother results to my ear and in measurement. I'm manually entering all the values as well, so less typing is better, haha.

 

[lempogi]

I run both Spinorama and correction from REW for frequencies below the Schroeder frequency (if you don't know how to calculate this I can provide links). It has to be done in a particular order and a particular way for it to work properly with the Wiim.

The first thing to do is to enter the values you have for Auto EQ from the Spinorama data that you have pictured. You may want to ignore the two lowest frequencies, because they will be lower than the Schroeder frequency in your room and they will be corrected anyway. Also I don't enter Spinorama EQ below the crossover to my sub, because it obviously doesn't apply to my sub. I use the EQ section for this.

Now comes the hard part, if you use Room Fit it will immediately turn off your EQ, so instead of Room Fit working with the Spinorama data, it will just ignore it, so you will end up with Spinorama over Room Fit, instead of the other way around, which is usually not ideal. Room fit for bass correction also has worse results than recommended REW figures.

What I do is use REW instead. First thing that is needed to do is to run Room Fit so that you have a space to put the recommended EQ's. However, Room Fit messes with the sub levels in a way that you won't see in the PEQs if you use Subwoofer Calibration. So I recommend turning that off. Once you run Room Fit you have a space to put your REW corrections for bass.

Then you turn on your Spinorama EQ in the EQ section with the Room Fit off. I use the MMM method on REW to measure bass. Here is a link on how to do this:

The moving microphone method (MMM) for Dummies using REW

After getting a couple of weird results on my first MMM measurements, I went searching around the forums & YouTube for what went wrong. To save others from gathering various bits and pieces of info from a dozen sources, I decided to publish this quick write-up here. You can use it as a rough...

www.audiosciencereview.com

The measurement I get with the Spinorama on but Room Fit off looks like this:

View attachment 25915

Then I generate filters for knocking down the modes using the EQ tool in REW. I set the frequency range below 200HZ, because the Schroeder frequency in my room is about 150HZ. I set it to create negative filters only. This is part of the reason REW is better than Room Fit, which will add positive filters no matter what. I then enter the negative filters into the Room Fit profile I created before, turning off all the generated ones from Room fit first and substituting the REW filters. I also added a tilt to lower the high end. The filter is -4db at .7Q at 1.5khz High Shelf filter. This allows it to match the house curve I was using, which was the Trained Listeners curve from Floyd Toole's research.

The result I get is here:

View attachment 25916
This sounds amazing in my room, with very tight and accurate bass. It helps that the Speakers I am using are the Affordable Accuracy Monitors by Dennis Murphy, but your Polk XT and sub setup should be able to get this. It also helps that I have good placement of my sub and speakers to avoid any wideband nulls in the bass. You can also use a Harman curve or any other curve that you prefer to knock down the bass peaks. I picked this curve because it was already close to what I measured.

The very narrow band nulls that you can see in the second measurement cannot be heard. Floyd Toole's research has shown that listeners are very sensitive to narrow peaks in bass frequencies, but narrow nulls are ignored. Humans can fill in nulls in our brains because of the harmonics created by low frequency sounds. The peaks are perceived as "boomy" and we can't filter them out with our brains, we need EQ.

Using only negative filters works much, much better than the way Room Fit does it. As Floyd Toole also says, room correction software is created by people who don't quite understand psychoacoustics research. Room Fit is better than most, but the method I am showing here more closely follows the research of how people hear.

Let me know if you have questions!

Thanks for your detailed steps of measurement using REW. For sure I'll try this in the future. I need to build my confidence first as those are a bit technical for me as of now. Need to learn more.

 

[dominikz]

Bottom line : after RoomFit application should one add additional filters over RF correction and if so how to calculate them.

Have you perhaps tried to use the auto-generated EQ correction for Elac Debut 2.0 B5.2 (link to EQ) from spinorama.org?
That should fix the brightness issue completely and significantly improve the overall anechoic response:

(source)
Just load the above EQ filters to WiiM EQ view, you can use stereo filters since the same correction should apply to both channels. Note that you don't need to configure any filters below 400Hz, if 400Hz is the upper frequency you use in RoomFit.

I was wondering if I would be able to blind test identify the REW eq vs the Room Fit one, but I don't have a good way to do that at the moment. My "much, much better" assessment is something I should know better not to trust without blind testing. .

As a general principle, fewer eqs and fewer boosts are better, especially if starting with speakers that have a smooth response. You're almost certainly correct that with no-boost mode activated in Room Fit that no new resonances are created, but the measurements and my ears tell me that the low end isn't quite as smooth with Room Fit. This is probably due to differences in the algorithms used to generate the filters. The Room Fit filters always fill up all 10 available bands for some reason, and I would guess that's an inherent problem, but can't be sure. But yeah, blind testing would be needed.

Room Fit below the Schroeder Frequency using no-boost mode with a calibrated mic would produce results that most people will be very happy with.

I forgot to mention I used variable smoothing in REW for the graphs shown, but originally used 1/48 for the MMM measurement. Was trying not to overload the OP.

Part of the reason I like the REW method is that I can use any curve I want, and the Trained Listeners curve just happens to sound good on almost everything I listen to. The Affordable Accuracy Monitors have a wonderful midrange that is highlighted by this curve. I also did a Harman Curve using the same method and felt like it was a bit too "boom and tizz" for my taste. And I love bassy music. I probably like a "darker" and "warmer" sound than most people though.

What I would really like to see from Wiim would be 20 bands of PEQ in Room Fit (purely to add Spinorama EQ, few filters are needed below the Schroeder frequency), and simple bass and treble controls in the EQ section. Just two sliders with bass and treble to be able to adjust for old recordings that tend to have little bass and too much treble.

I agree with so many of the points you raise - especially about the need to link both room and loudspeaker correction to outputs, and to have a much more simple tone control available. This is something I also brought up several times - including very recently in this post.

A small tip regarding how to approach blind listening when comparing different EQ strategies - there is actually a relatively simple but very effective way to do it, in case you're interested.

In rough terms the process is:

1. Calculate EQ filters which you want to compare (e.g. with REW and with RoomFit)
2. Select audio files you want to use for your comparison
3. Use audio processing software (e.g. something like Audacity, or foobar2000 with a PEQ plugin, but many others work as well) to process the source audio file with the EQs and create two new files:
   • A) original file processed only with RoomFit-calculated EQ correction,
   • B) original file processed only with REW-calculated EQ correction
4. Disable EQ in your playback chain completely
5. Use foobar2000 ABX comparator plugin to compare the two new files.

By comparing these two files with "embedded" EQ corrections in a system with EQ disabled you are basically comparing the EQ corrections themselves.

I've used this approach in the past to compare results with various room correction tools (e.g. see this example).

BK curve is pretty flat and low level through the bass for most people.

Stereo, only because I didn't get good results with individual channel. In theory with my unbalanced room the latter should be better, but stereo produced smoother results to my ear and in measurement. I'm manually entering all the values as well, so less typing is better, haha.

Note that if "individual channel" RoomFit is used, you get +3dB in the bass when playing both channels (which is a natural consequence of coherent vs incoherent sound source summation, not something specific to WiiM - a bit more about it here, here and here).

So IMHO B&K is a pretty balanced target when using "individual channel" correction, but it is bass deficient if using "stereo" correction - though this will to some extent of course also depend on loudspeaker directivity, and how "live" the listening room is.

When "stereo" correction is used I find that a curve with a more aggressive downward slope is indeed needed compared to "individual channel" correction, so I understand why you'd prefer the "trained listener" curve when using the "stereo" correction.

 

[lempogi]

I've used a similar approach for years and I fully agree it is technically the most sound one, and the most flexible.

However, I feel the need to comment that with the recent introduction of RoomFit "No-Boost Mode" one can now get reasonably similar results with RoomFit.

First, you are absolutely correct that even with "No-Boost Mode" enabled RoomFit will still use a few individual filters with positive gain to shape the overall response. But since the total correction filter response with this mode remains at or below 0dB, it won't/can't introduce any new resonances to the system. So IMHO this response shaping strategy can be viewed as an inefficiency of RoomFit filter calculation algorithms rather than something audibly problematic.

IME the differences in corrections between RoomFit and REW are now mainly in flexibility and precision - RoomFit is limited by the fact that the least amount of smoothing you can use is 1/12 (while in REW you can use even less), and the maximum Q used by RoomFit filters is 10 (while in REW it can be higher).

A note on measurement method here: in REW you can use various kinds of measurement methods to measure your system response while in RoomFit you are more limited. With REW I usually use MMM (as it is very repeatable and implicitly adds some desirable spatial smoothing). MMM implementation in RoomFit is IMHO not ideal (see here for more info), and the Multiple Measurement options has a long standing bug (sweep usually doesn't play successfully all 3 times for both channels). So in effect, at the moment IMO it is best to use the basic single-sweep method in RoomFit. Luckily, there shouldn't really be much difference in measured response below 200Hz between single sweep measurement and MMM, so for this purpose it should not be an issue.​

This means that the system response after automatic correction calculated by RoomFit (with "No-Boost Mode" & 1/12 smoothing) will be slightly less smooth and precise than the one from REW (with max individual and total gain set to 0dB and Var smoothing) - assuming we used the same single sweep measurement, the same correction frequency range, and the same target curve in both.

This is all of course assuming optimal RoomFit configuration, use of a calibrated measurement microphone, and a mobile phone that doesn't change the measured response (see here for one example of this strange issue in some phones).

However, human hearing has some limits as well, and I wouldn't bet that these minor differences in correction precision would be very significant to most listeners under blind and level-matched conditions. This is something to consider at least.

Lastly a practical concern: at the moment there is still no import function in WiiM EQ and RoomFit, so any externally calculated EQs unfortunately needs to be input manually. This is quite tedious to do, especially if you use both the individual channel RoomFit (2x10 band of PEQ with 3 attributes each) plus stereo EQ for loudspeaker correction based on spinorama data (up to 1x10 band PEQ with 3 attributes each). That gives a total of 90 numbers to type into the WiiM Home App for a correction you want to apply manually (and you may wish to try different profiles so it can multiply)! Fortunately WiiM has already hinted that import functionality is on their radar (see here) - and I do hope it will come soon! That would really make it much easier to use REW to create custom corrections for WiiM devices.

In summary, while I'd definitely suggest to anyone interested in audio to learn how to use REW, IMHO comparably good results can now be achieved with RoomFit for those unwilling to go quite as deep down the rabbit hole!

Your summary definitely aligns with my situation. I'm a very busy family man and right now I just want to have a good sounding audio system in our living room without spending too much time measuring everything. With the help of Wiim RoomFit, it is easier for me. It's not perfect, but better, than without any room correction. But I still open myself on learning REW and all the technicalities of audio science. It's very interesting and I love the interactions of people here. Maybe some day I'll be able to do those deeper measurements.

 

[dominikz]

Your summary definitely aligns with my situation. I'm a very busy family man and right now I just want to have a good sounding audio system in our living room without spending too much time measuring everything. With the help of Wiim RoomFit, it is easier for me. It's not perfect, but better, than without any room correction. But I still open myself on learning REW and all the technicalities of audio science. It's very interesting and I love the interactions of people here. Maybe some day I'll be able to do those deeper measurements.

That is very reasonable, IMO! And my opinion is that you probably won't be missing too much in the way of sound quality either way, TBH.
But that doesn't mean it isn't fun to learn and experiment with other approaches!

Just be aware that the quality of RoomFit correction is still at least somewhat dependent on the phone you use - a) the phone built-in mics are in general not very reliable (they are not individually calibrated), and b) some phones seem to misrepresent the bass even when using individually calibrated external microphones (see this example).

But many phones work well enough so you may not run into these issues at all.

BTW, do you already own a measurement mic that you could use?

My hope is that in the future WiiM will find some way to make RoomFit less sensitive to this kind of phone-specific issues. One idea tossed around the forum was to use the WiiM Voice Remote microphone in RoomFit, if possible. Presumably this could be pre-calibrated at the factory, it would avoid relying on the phone, and already comes bundled with many devices. But I guess no one knows if WiiM is actually considering ever implementing this or not.

 

[cristoo]

Have you perhaps tried to use the auto-generated EQ correction for Elac Debut 2.0 B5.2 (link to EQ) from spinorama.org?
That should fix the brightness issue completely and significantly improve the overall anechoic response:
View attachment 25923
(source)
Just load the above EQ filters to WiiM EQ view, you can use stereo filters since the same correction should apply to both channels. Note that you don't need to configure any filters below 400Hz, if 400Hz is the upper frequency you use in RoomFit.

Thanks so much for reply and advice. Frankly speaking I was thinking to do as you suggested when found these data at spinorama site but my RoomFit at the moment is done according to BK curve and this settings seems to be for Harman (it is described at spinorama site as “These EQ optimise the Harman score and is a good start for far field listening like in your living room”). So can I use this settings along with BK RoomFit 20-400 hz correction ?