How to think about room correction settings like frequency range and gain?

[harkpabst]

As I said I am not going to argue.

This is exactly what you are doing.

If some people believe speaker positioning has no influence on soundstage, than, well, they can believe what they want…

Who said that speaker positioning had no influence on soundstage?

I just asked anybody (not you specifically) to provide some reasoning why placing speakers far away from the front wall would be the only way to maintain a good soundstage. Nobody could, because it's not the case.

"Early reflections" is just a buzz word until you describe in detail what's actually happening. Forget about "early reflections" below ~200 Hz. No such thing, this is still slow bass. Forget about "early reflections" above ~4000 Hz. The diffuse sound field will dominated all acoustic effects. This all has to be seen in correlation with speaker directivity.

@dominikz took the time to explain this over and over and over again. If you don't want to listen, you don't have to. But someone may stand up and contradict whenever myths are reproduced with nothing to back them up.

PS:
I don't want to argue, of course.

 

[Adam205]

It looks OK, though I'd suggest to use the "Individual Channel Room Correction" when the option is presented. That will give you a bit more bass.

Judging from the massive peak at 8kHz I assume you used the smartphone built-in mic to measure, right?
You would get a more reliable measurement and correction with an external calibrated measurement microphone; however the practical difference might or might not be significant. No way to know.

Most importantly, if you are happy with the result, it is good.

What do you think about these measured results? Yes, it is measured via mobile.

 

[Musician]

This is exactly what you are doing.


Who said that speaker positioning had no influence on soundstage?

I just asked anybody (not you specifically) to provide some reasoning why placing speakers far away from the front wall would be the only way to maintain a good soundstage. Nobody could, because it's not the case.

"Early reflections" is just a buzz word until you describe in detail what's actually happening. Forget about "early reflections" below ~200 Hz. No such thing, this is still slow bass. Forget about "early reflections" above ~4000 Hz. The diffuse sound field will dominated all acoustic effects.

This is true if you ignore the precedence effect. Ufortunately , it kicks in if you like it or not, and the first arriving sound ( the direct sound from the speaker ) is a much more important selection for the brain than the diffuse sound, which are attenuated by the brain about -10 dB .

This can not be seen in a microphone measurement.

Regardless of the speaker having good directivity or bad, the direct sound is much more important for the brain than the diffuse sound .

This is something than ASR dont talk much about probably because they ignore the precedence effect and its much easier to just show a klippel measurement. But the reality is different.

I see perfect directivity loudspeakers mostly as a salespoint from the Harman group, promoting Revel speakers.

Perfect directivity dont hurt, ofcourse, but its not nearly as important as people think.

And yes, I have DIY built many loudspeakers, including the good directivity Hybriddist with waveguide , owning expensive SAM monitors, so I have tried and verified everything.

 

[dominikz]

Regardless of the speaker having good directivity or bad, the direct sound is much more important for the brain than the diffuse sound .

This is something than ASR dont talk much about probably because they ignore the precedence effect and its much easier to just show a klippel measurement. But the reality is different.

I see perfect directivity loudspeakers mostly as a salespoint from the Harman group, promoting Revel speakers.

Dr. Toole, Harman and ASR are all actually very vocal about the relevance of loudspeaker direct sound for sound quality, and that in-room steady-state loudspeaker response should not be seen as representative of how humans perceive sound. There are countless threads and posts about this on ASR.

There are also quotes from dr. Toole specifically warning about automated room correction systems that might degrade otherwise good loudspeakers because they can degrade their on-axis response (i.e. direct sound) at higher freuencies.

Good loudspeaker directivity is beneficial for two main reasons:

• Early reflections from such speakers will have spectrum similar to the direct sound so will not degrade perceived sound quality.
• Speakers with good directivity but poor on-axis (direct) response can be effectively fixed by EQ based on anechoic data. Speakers with poor on axis response and poor directivity cannot be helped as easily.

Have you read "Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms, Third Edition" by dr. Toole? If not, it is very much worth the time!

On the topic of precedence effect, chapters 7.6.4 and 7.6.5 discuss it specifically in the context of room reflections in detail, citing several studies. Chapter 10.3 discusses direct sound vs early reflections and how they are perceived. Chapter 7.4 discussed subjective evaluation of various parameters, including side wall treatment, loudspeaker directivity, early reflections, diffuse vs direct field, etc. The box on page 182 discussing in-head localization that can happen in very dead (reflection-free) environments is IMHO especially interesting.

So I respectfully disagree with your statement above; importance of direct sound on perception is very often discussed in ASR.
I also disagree with your previous statement that early reflections are detrimental to sound quality - how beneficial or detrimental early reflections are depends on context as well as personal preference.

What is certain is that many people like (and even prefer) the effect of side-wall reflections with wide-directivity loudspeakers.

The optimal acoustic delays from the sidewalls for two channel music is between 20-30ms , this makes the sound more ambient and one doesn’t recognize the stereo system flaws that much .

Can you provide a reference for this claim, please?
To my knowledge of the precedence effect, this amount of delay between direct sound and sidewall reflection would be bordering with being perceived as a distinct echo - I sincerely doubt that would help with clarity. Not to mention that the room would need to be over 15m wide to accommodate.

 

[dominikz]

What do you think about these measured results? Yes, it is measured via mobile.

The vertical sale is a bit compressed so it is difficult to see details, and the resonance of the mobile phone mic (around 8kHz) is skewing the results a bit because it is causing the RC to set the target level too high - that might potentially result in some mid-bass and upper-bass excess.

Looking at the bass extension I assume you are not using a subwoofer?

Also, I see you have a strong dip at ~60Hz in both channels, is the front face of your loudspeakers (where the drivers are) by any chance ~1.3m from any of the walls?

Ultimately, are you happy with how it sounds?

 

[slartibartfast]

The vertical sale is a bit compressed so it is difficult to see details, and the resonance of the mobile phone mic (around 8kHz) is skewing the results a bit because it is causing the RC to set the target level too high - that might potentially result is some mid-bass and upper-bass excess.

Looking at the bass extension I assume you are not using a subwoofer?

Also, I see you have a strong dip at ~60Hz in both channels, is the front face of your loudspeakers (where the drivers are) by any chance ~1.3m from any of the walls?

Ultimately, are you happy with how it sounds?

Don't forget phone mics roll off at low frequencies as well.

 

[Adam205]

The vertical sale is a bit compressed so it is difficult to see details, and the resonance of the mobile phone mic (around 8kHz) is skewing the results a bit because it is causing the RC to set the target level too high - that might potentially result in some mid-bass and upper-bass excess.

Looking at the bass extension I assume you are not using a subwoofer?

Also, I see you have a strong dip at ~60Hz in both channels, is the front face of your loudspeakers (where the drivers are) by any chance ~1.3m from any of the walls?

Ultimately, are you happy with how it sounds?

My speakers are only a few inches from the wall. I'm using a subwoofer. Should I change any other parameters before I take another room correction measurement?

 

[dominikz]

My speakers are only a few inches from the wall. I'm using a subwoofer. Should I change any other parameters before I take another room correction measurement?

First may I ask where is your subwoofer placed - how far from all surrounding walls?
How is the sub connected?
Which crossover frequency are you using?

 

[Adam205]

First may I ask where is your subwoofer placed - how far from all surrounding walls?
How is the sub connected?
Which crossover frequency are you using?

The subwoofer is also a few inches from the back wall and about 1.5m from the wall to the right. The cutoff frequency is about 80Hz and it is an active subwoofer connected to an Onkyo amplifier.

 

[dominikz]

The subwoofer is also a few inches from the back wall and about 1.5m from the wall to the right. The cutoff frequency is about 80Hz and it is an active subwoofer connected to an Onkyo amplifier.

Nice looking setup!
If you can, I'd suggest to move the sub closer to the corner on the right (I'd try to put it right in the corner), and to incease subwoofer level significantly.
At the moment your sub seems to be set relatively lower than the speakers and you don't seem to get a lot of bass below about 70Hz.
After this I'd suggest to re-run RC so we can better estimate the optimal RC config details for subsequent attempts.

 

[slartibartfast]

The subwoofer is also a few inches from the back wall and about 1.5m from the wall to the right. The cutoff frequency is about 80Hz and it is an active subwoofer connected to an Onkyo amplifier.

Isn't the subwoofer connected to the Ultra sub out?

 

[Wiimer]

Sorry to interrupt, but I just wanted to say thank you to @dominikz.

I took your excellent advices and re-did speaker setup in my room. (I moved the speakers and Sub closer to the wall.)

The result is very good

Here are the measurements of the 2 Sub connected to the Pro, and main speakers connected to the Ultra.

Sub has been manually corrected with PEQ. (crossed at 120Hz)


To compensate for the speakers, the Sub volume was turned down by 20% before running RC.





Next, after restoring the Sub volume to its original level, I connected the Mini to Ultra's optical input and ran RC on the Mini. (The gray line is corrected by Ultra's RC.)


The final result.


I could copy the RC results from the Mini and overlay the PEQ on the Ultra, but I haven't done that. I probably don't need to.

I have tried several RC's on the main speakers, including 120Hz-500Hz and flat, but I like the sound best with R&B corrected down to 4kHz.

I will try to adjust the Sub a bit more next week, but I feel it sounds good as is. Most of all, what made me most happy was that my room was a bit larger.

I know there are still some things that need to be resolved, but thanks to you I have made some progress.

Please keep up the good work
Thank you so much.

 

[Musician]

Dr. Toole, Harman and ASR are all actually very vocal about the relevance of loudspeaker direct sound for sound quality, and that in-room steady-state loudspeaker response should not be seen as representative of how humans perceive sound. There are countless threads and posts about this on ASR.

There are also quotes from dr. Toole specifically warning about automated room correction systems that might degrade otherwise good loudspeakers because they can degrade their on-axis response (i.e. direct sound) at higher freuencies.

Good loudspeaker directivity is beneficial for two main reasons:

• Early reflections from such speakers will have spectrum similar to the direct sound so will not degrade perceived sound quality.
• Speakers with good directivity but poor on-axis (direct) response can be effectively fixed by EQ based on anechoic data. Speakers with poor on axis response and poor directivity cannot be helped as easily.

Have you read "Sound Reproduction: The Acoustics and Psychoacoustics of Loudspeakers and Rooms, Third Edition" by dr. Toole? If not, it is very much worth the time!

On the topic of precedence effect, chapters 7.6.4 and 7.6.5 discuss it specifically in the context of room reflections in detail, citing several studies. Chapter 10.3 discusses direct sound vs early reflections and how they are perceived. Chapter 7.4 discussed subjective evaluation of various parameters, including side wall treatment, loudspeaker directivity, early reflections, diffuse vs direct field, etc. The box on page 182 discussing in-head localization that can happen in very dead (reflection-free) environments is IMHO especially interesting.

So I respectfully disagree with your statement above; importance of direct sound on perception is very often discussed in ASR.
I also disagree with your previous statement that early reflections are detrimental to sound quality - how beneficial or detrimental early reflections are depends on context as well as personal preference.

What is certain is that many people like (and even prefer) the effect of side-wall reflections with wide-directivity loudspeakers.


Can you provide a reference for this claim, please?
To my knowledge of the precedence effect, this amount of delay between direct sound and sidewall reflection would be bordering with being perceived as a distinct echo - I sincerely doubt that would help with clarity. Not to mention that the room would need to be over 15m wide to accommodate.

I found my notes from testing this and I was wrong . The optimal delay time for reflected sidewall sound for two channel playback is about 15-20ms.

As a amateur sound engineer , the optimal timedelay when recording acoustic instruments and have a nice reverbant sound are also between 15-20 ms . You can also play around using digital reverb with a 2 channel mix , where good results can be found at about the same time span .

30-35 ms delay are often considered as the time span where distinct eco occur.

 

[dominikz]

As a amateur sound engineer , the optimal timedelay when recording acoustic instruments are between 20-30 ms . You can also play around using digital reverb with a 2 channel mix , where good results can be found at about the same time span .

50 ms delay are often considered as the time span where distinct eco occur.

Please note that live performance, sound recording/production, and sound reproduction are three different processes which also have different requirements on equipment and acoustics. They should not be conflated.

A loudspeaker is a sound reproduction device and not a musical instrument. The recordings a loudspeaker is meant to reproduce already contain some spatial cues, and sound signature of the 'venue' in the recording is overlaid on top of the acoustics of our listening room. This in itself is fundamentally different to an acoustic instrument - but it is also not the only difference.
The recordings also vary a lot in the amount and type of spatial cues, as well as in their overall sound signature.

So generalizing the principle that might work when adding reverb to a recording of an acoustic instrument to loudspeaker placement is IMHO unjustified. Even if it may work in some cases, it is absolutely not guaranteed to work every time.

A lot of research was done in this field and it is nicely summarized in dr. Toole's book I referenced above. It is IMHO a book anyone interested in audio should read, regardless whether they agree with some of its conclusions or not.

 

[dominikz]

Sorry to interrupt, but I just wanted to say thank you to @dominikz.

I took your excellent advices and re-did speaker setup in my room. (I moved the speakers and Sub closer to the wall.)

The result is very good

Here are the measurements of the 2 Sub connected to the Pro, and main speakers connected to the Ultra.

Sub has been manually corrected with PEQ. (crossed at 120Hz)
View attachment 20122

To compensate for the speakers, the Sub volume was turned down by 20% before running RC.
View attachment 20125

View attachment 20123


Next, after restoring the Sub volume to its original level, I connected the Mini to Ultra's optical input and ran RC on the Mini. (The gray line is corrected by Ultra's RC.)
View attachment 20126

The final result.
View attachment 20127

I could copy the RC results from the Mini and overlay the PEQ on the Ultra, but I haven't done that. I probably don't need to.

I have tried several RC's on the main speakers, including 120Hz-500Hz and flat, but I like the sound best with R&B corrected down to 4kHz.

I will try to adjust the Sub a bit more next week, but I feel it sounds good as is. Most of all, what made me most happy was that my room was a bit larger.

I know there are still some things that need to be resolved, but thanks to you I have made some progress.

Please keep up the good work
Thank you so much.

I'm really happy to hear you found some of the information and advice helpful!
I see you are using cascaded WiiM devices for multi-level correction - an interesting approach!

Have you perhaps considered trying to apply EQ based on anechoic data above say 400Hz and RC only below 400Hz? I find that usually works really well. You can find EQ profiles based on anechoic data for many loudspeaker models at spinorama.org.
Anyway, just something to consider if you want to continue tweaking. Enjoy!

 

[Wiimer]

Have you perhaps considered trying to apply EQ based on anechoic data above say 400Hz and RC only below 400Hz? I find that usually works really well. You can find EQ profiles based on anechoic data for many loudspeaker models at spinorama.org.
Anyway, just something to consider if you want to continue tweaking. Enjoy!

I'll try that next week.
My KEF has already been measured by Erin

Thanks for the sound advice

 

[canard]

I'll try that next week.
My KEF has already been measured by Erin

Thanks for the sound advice

Wiim could easily allow a capture in the axis of one louddpesker at 70cm to help... it would not be complicated for them

but somewhere joins the possibility of raw measurements, as for effective control of correction etc...

 

[dominikz]

What seems to work fairly well in many cases (and is relatively physically inobtrusive) is to push the speakers close to the wall behind them (but keep about 5cm of gap to the wall if speakers have a bass-reflex port in the back), and to put the subwoofer in a room corner. Of course it may not work in every room and every layout, but it is a reasonable place to start with.

If you are interested to know more about why these are solid choices I suggest to read this article by Genelec.

In short, pushing speakers close to the wall behind them and subwoofer in a corner will push the main SBIR cancellation to a much higher frequency where it is audibly less problematic. Additionally, putting both speakers and sub close to boundaries will increase bass output, which you will anyway tame/knock-back later by EQ (i.e. room correction) - so it gives you more headroom for EQ.

Note that EQ is very good for reducing resonances/peaks in the response, but is not ideal for boosting dips in the response - this is what is driving placement suggestions mentioned above.

After you have optimized placement you should set subwoofer crossover frequency (typically at 80Hz) and match its level to your loudspeaker. Don't be afraid to have the sub slightly louder at this point - that will anyway be fixed by RC later.
If you can measure the in-room response with REW to fine tune the crossover that is even better, but if not you should still be able to achieve solid results by ear.

Once this is done use the automatic WiiM Subwoofer & Speaker Sync function in the WHA to tune the relative delays of the loudspeakers and subwoofer.

After that use the WiiM Room Correction function - I propose to use the following parameters to start:

• Target curve: B&K
• Correction range: 20-400Hz
  • This type of EQ much above the specified range can no longer be considered "room correction" but rather "loudspeaker correction". This can easily make the sound worse, especially if you have good loudspeakers.
  • Note: If you don't have a sub that goes down to 20Hz use instead the lower-bound frequency from the sub spec sheet.
• Max gain: 12dB
  • You need to have a high value here for the app to be able to bring down severe room resonances. A current limitation in the app is that the same max gain control affects both positive and negative gain filters, but this will apparently improve soon.
  • After the separate positive/negative max gain controls are implemented I'd suggest to keep the max positive gain below 3dB (preferably 0), and keep max negative gain at 12dB.
• Max Q: 5
  • Note that the default value of 10 is IMHO fine in case no EQ boosts/positive filters are applied.
  • If you want to be more conservative, especially until separate positive/negative max gain controls are introduced you can use a lower value like "5", but note that that will also be less precise in knocking down resonance peaks.
• Smoothing: 1/12 octave
  • IMHO 1/12 octave is the best choice to have any precision in addressing resonant peaks.
  • It would be better if WHA offered progressive variable smoothing (like in REW), but this is not implemented.
• Subwoofer Calibration: enabled if you have a sub, disabled otherwise.
• Multiple Measurements: enabled
  • Not critical either way, but I prefer to have it enabled and move the mic a little bit between the attempts to get some spatial smoothing.
• Precision Room Correction: disabled
  • I wrote about this function here, including why I wouldn't recommend it in most cases.
• Import a Calibration File: yes (if you have an external calibrated measurement microphone)

This should already give you solid results - you can share WHA screenshots here and we may be able to help you fine tune if needed.

Hope this is helpful!

After having tested the Min Gain and Max Gain RC controls introduced in WHA 3.1.8 more extensively I have to unfortunately report that the current RC implementation doesn't really handle well low values of Max Gain.

I.e. if you set Max Gain to 0,5 dB (minimum currently allowed) RC will spend most of the PEQ bands stacking such low-gain filters one on top of the other - which is a very inefficient use of the available EQ (example 1, example 2, example 3).

Therefore I have to revise my Min/Max Gain proposed values; these are the values that I'd use with the current RC implementation:

• Min Gain: -12.0dB
• Max Gain: +3.0 dB to start (but even higher values might be needed in case you notice multiple filters stacking at the same frequency )

I've already raised a support ticket to @WiiM Support about this and made a few suggestions for potential improvement:

• Consider supporting Max Gain value of "0 dB" - i.e. with this setting the RC algorithm would only use negative-gain PEQ filters.
• Assume that if a person limits Max Gain to a low value they are OK if some dips don't match the target so the RC algorithm should in general avoid stacking multiple positive-gain filters at the same frequency.
• Always prioritize cutting peaks over boosting dips - i.e. assign more PEQ bands to negative-gain filters over positive-gain filters. People are simply more sensitive to peaks than dips so it makes sense to focus on fixing peaks as well as possible.
• Consider implementing separate controls for Individual Max Gain and Overall Max Gain (like Individual Max Boost and Overall Max Boost controls in REW).
• Consider implementing separate controls for Boost Max Q and Cut Max Q - i.e. to be able to have sharp/precise cut filters (e.g. Q up to 10) to handle severe resonances, but only gradual boost filters (e.g. Q up to 3) to avoid any chance of audible ringing.
• Consider implementing "basic" and "expert" view in the RC settings. "Basic" view might only expose a few parameters (e.g. target curve, frequency range and mic calibration file loader), while "Expert" view would expose all of config parameters for fine-tuning by advanced users.

@WiiM Team and @WiiM Support I hope you will consider at least some of these suggestions, and I'd appreciate if you could provide some thoughts/feedback. Thanks in advance!

 

[Adam205]

Nice looking setup!
If you can, I'd suggest to move the sub closer to the corner on the right (I'd try to put it right in the corner), and to incease subwoofer level significantly.
At the moment your sub seems to be set relatively lower than the speakers and you don't seem to get a lot of bass below about 70Hz.
After this I'd suggest to re-run RC so we can better estimate the optimal RC config details for subsequent attempts.

Thank you! These are Klipsch R-50M speakers.
Unfortunately, I don't have the option to move the subwoofer, let alone put it in a corner. You can't see it in the photo, but I have a wifi router in the corner. Should I have the subwoofer on or off during room correction? So how do I set the parameters for a new room correction measurement in this case?

 

[Adam205]

Isn't the subwoofer connected to the Ultra sub out?

I have a WiiM Mini