Room Mode + Speaker Anechoic Measurement Correction
- Thread starter lempogi
- Start date
lempogi
Member
Ok I'll try and test this tonight, thank you! I admire your dedication about it. Wiim should hire you like a test engineer for their products. Mutually they are also benefiting from your efforts.
Did you change any settings on the sub compared to your initial attempts? The sub level in this latest attempt seems about 10dB lower than it was e.g. here:
I'd suggest to use the higher level for the sub because that helps with the dip.
Thanks! My motivation is mainly to learn more myself, but it is also a pleasure to share knowledge and help fellow members along the way.Ok I'll try and test this tonight, thank you! I admire your dedication about it. Wiim should hire you like a test engineer for their products. Mutually they are also benefiting from your efforts.
We can only hope that WiiM will use some of our collective findings and suggestions to further improve their products!
Did you change any settings on the sub compared to your initial attempts? The sub level in this latest attempt seems about 10dB lower than it was e.g. here:
I'd suggest to use the higher level for the sub because that helps with the dip.
Initially like 2-3 days ago when we started all this attempts sub level on unit was on 11 o’clock, than like yesterday after suggestion of one of users I lowered it to 10 o’clock and left so several latest trials are with these seeting. Today I only played with phase 0/180 and moved the sub. Should I increase sub level on unit or in WHA app? Or it doesn’t matter ?
One more thing - when moved the sub I changed the plug i used for mono connection form Ultra (from
right to left just for curiosity). It should not matter but just for info. I can replug like it was before.
I highly appreciate your and all community help. Thanks a lot.
Ok I found one change, I had after switching off for a while my amp, smsl sdb eq setting on instead of direct. I switch to direct and increased the sub level. And we are back to what I had. I think I will give up coz I am losing this me vs acoustic battle even armed with Wiim equipment….
This looks like a much better baseline for RoomFit compared to before - good work!
Now your response above 100Hz is already pretty good, and there's a healthy level of bass below it without severe dips.
Now we just need to fine-tune RoomFit parameters a bit. I'd suggest to try these:
Freq range = 25Hz-200Hz,
"No Boost Mode" = disabled,
Min Gain = -12dB,
Max Gain = +6dB,
Max Q = 10,
Individual Channel RoomFit
So we’re keep on fighting. Me or room acoustics. @dominikz lead towards victory.
Did RF as instructed, rest of setting including sub the same
That's what I had guessed back here. The dip is mainly caused by the listening position, so very hard to fix.
In this latest iteration the dip is a lot deeper again than it was before. Given that RoomFit is working hard to reduce the level from below 30 Hz up to above 100 Hz (up to -25 dB!) I'm still not convinced that a higher subwoofer level is the way to go in this very particular case. I'd go back to the setting where the 70 Hz dip is as narrow as as possible and as small in level as possible.
I was a bit surprised how much lowering the crossover frequency did improve the frequency response further up. This is definitely something you'll want to keep.
Maybe (and just maybe) this could be one of the rare cases where experimenting with overlapping crossover frequencies and/or asymmetric filter slopes could potentially improve the response in one exact listening position. But I don't think I'd want to follow this line of attack. Not without the help or REW and maybe not at all.
In this latest iteration the dip is a lot deeper again than it was before. Given that RoomFit is working hard to reduce the level from below 30 Hz up to above 100 Hz (up to -25 dB!) I'm still not convinced that a higher subwoofer level is the way to go in this very particular case. I'd go back to the setting where the 70 Hz dip is as narrow as as possible and as small in level as possible.
I was a bit surprised how much lowering the crossover frequency did improve the frequency response further up. This is definitely something you'll want to keep.
Maybe (and just maybe) this could be one of the rare cases where experimenting with overlapping crossover frequencies and/or asymmetric filter slopes could potentially improve the response in one exact listening position. But I don't think I'd want to follow this line of attack. Not without the help or REW and maybe not at all.
lempogi
Member
- Joined
- Oct 19, 2022
- Messages
- 976
Hi @cristoo ,
Please help us investigate this issue by sending feedback through the WiiM Home app: go to the More tab and tap Feedback. Our engineers will look into it promptly once received.
You could definitely try, but there's a "but" (or two).A simple question: is it a bad idea to manually edit the profile and adjust the PEQ values in this way?
View attachment 26230
First, it's possible but not very likely that RoomFit got the correction all wrong in this area. Increasing the level might really do nothing good here regarding a flatter response.
Second, there are obviously more than one peak filters in close proximity (to reach this high level of attenuation), but that doesn't mean that the center frequency of any of those filters is right where you would want to place the manual correction. So, you might end up having to do quite some more manual work of shifting filters around, all for a somewhat questionable approach.
It's never wrong to try out something. I wouldn't expect the fix to be that easy, though.
I was thinking of simply setting the left and right sides to +6dB each around 75Hz, but perhaps you are rightYou could definitely try, but there's a "but" (or two).
First, it's possible but not very likely that RoomFit got the correction all wrong in this area. Increasing the level might really do nothing good here regarding a flatter response.
Second, there are obviously more than one peak filters in close proximity (to reach this high level of attenuation), but that doesn't mean that the center frequency of any of those filters is right where you would want to place the manual correction. So, you might end up having to do quite some more manual work of shifting filters around, all for a somewhat questionable approach.
It's never wrong to try out something. I wouldn't expect the fix to be that easy, though.
The good thing is (and I didn't mention it) that at least the second problem does not exist if you use PEQ instead of directly modifying the RoomFit profile.I was thinking of simply setting the left and right sides to +6dB each around 75Hz, but perhaps you are right
Yes, I didn't mention it too, but I was thinking about to do that
IMHO this layout is giving you the best baseline for EQ, since in the "Measured" response there are no severe dips in either channel. Unfortunately, since the <50Hz resonance is wide and very high in level, the way that RoomFit currently matches the target curve recreates the 60Hz dip in the corrected response.
There may be several perfectly valid ways to address the issue:
- @harkpabst suggests to tune subwoofer parameters until the dip is narrower and less audible,
- @Wiimer suggests to manually modify the RoomFit filters to avoid bringing down the dip, and
- @WiiM Team seems willing to look at the issue from software perspective if you open a support ticket.
- It may also be possible to get a different result by further modifying RoomFit parameters, or
- You could measure with REW and use it to generate correction PEQ filters that you load into RoomFit manually.
IMHO RoomFit should still be able to match the target below 50Hz without bringing down the dip as well, but in the current implementation your suggestions are all IMO valid approaches.
I'd say this is absolutely OK to do, and actually it is what I'd expect RoomFit to do automatically.A simple question: is it a bad idea to manually edit the profile and adjust the PEQ values in this way?
View attachment 26230
However matching the PEQ values to the dip manually might not be trivial, especially since the horizontal scale of the plot in RoomFit is really strange (seems like linear scale 20Hz-50Hz, then some kind of transition region without a clear scale from 50Hz to 100Hz, and then log scale from 100Hz to 20kHz). So matching the response exactly would require a bit of fine-tuning that may not be super user-friendly and definitely difficult to explain in a few forum posts.
A simple filter like you suggest may be something to try, sure! In that case I'd probably reassign filter #10 from each RoomFit channel to this task.
Thanks for taking an interest in this issue! It would indeed be ideal if RoomFit algorithm could be improved to better handle these kinds of cases.
Nice, looks this approach worked OK for you as well! I'd probably just disable the #10 PEQ (538Hz) in the left channel in the profile, I doubt such a high-Q correction is needed this high in frequency.
With this new correction the bass level seems lower and better matches the overall full-range target (and between channels), but the bass response is a bit more uneven than before. This might not really be an audible issue, however.
Do you have any preference between this and the previous correction?
I was thinking of stacking L/R PEQs on RoomFit for testing. Would this approach cause other problems?
The team replied to my ticket about this. I told them to contact @cristoo as it is not an issue with my device
I don't think so (I'm assuming internal processing in the digital domain has additional headroom), but it would perhaps be a bit unintuitive to manage for the user. E.g. if you re-do RoomFit at any point you'd need to remember you have this separate PEQ configured elsewhere which may or may not apply anymore.
Agreed. And if I think about it, stacking PEQ is not reflected in RoomFit, so he need to use HouseCurve...
slartibartfast
Major Contributor
- Joined
- Apr 18, 2024
- Messages
- 4,597
When I combine RoomFit and PEQ I give the PEQ the same name as the RoomFit profile so I know they belong together
Similar threads
