Room Mode + Speaker Anechoic Measurement Correction

[dominikz]

Sorry,
The 30hz boost was normal.
I missed @cristoo saying he turned off non-boost mode in #53.

I see, thanks for clarifying!

The 50-90Hz dip after correction could definitely be avoided with better allocation of PEQs by RoomFit.
It seems the algorithm does a poor job at identifying problem areas in the response, as well as in optimizing usage and distribution of available PEQ bands.
E.g. it makes absolutely no sense to spend PEQ bands boosting the bass droop at 20Hz at the cost of the huge dip at 70Hz.

So IMHO that is definitely worth reporting to WiiM as an issue.

 

[Wiimer]

I see, thanks for clarifying!

The 50-90Hz dip after correction could definitely be avoided with better allocation of PEQs by RoomFit.
It seems the algorithm does a poor job at identifying problem areas in the response and, as well as in optimizing usage and distribution of available PEQ bands.
E.g. it makes absolutely no sense to spend PEQ bands boosting the bass droop at 20Hz at the cost of the huge dip at 70Hz.

So IMHO that is definitely worth reporting to WiiM as an issue.

Ticket (541062) sent

 

[cristoo]

I reearanged my stuff and pushed sub directly in the corner. It is like 5 cm from wall now. Did measurements but they are not promising…

Phase on sub 0, WHA 0, crossover 250 in WHA, gain on sub reduced a bit to 10 o’clock.

 

[cristoo]

I did also full measurements
Crossover reduced to 65 hz , speakers pushed a bit to rear wall (circa 5-10 cm space), BK curve, 12-12 min-max gain, no boosts mode on. Range 20-400 hz.
Here what i got from RF plus test sweep from HousebCurve

 

[dominikz]

I did also full measurements
Crossover reduced to 65 hz , speakers pushed a bit to rear wall (circa 5-10 cm space), BK curve, 12-12 min-max gain, no boosts mode on. Range 20-400 hz.
Here what i got from RF plus test sweep from HousebCurve
View attachment 26098View attachment 26099View attachment 26100View attachment 26101

The new Elac loudspeaker position (close to wall behind) is IMO quite a bit better - the dip around 150Hz is gone, as I hoped. So I'd leave them at the new position and wouldn't touch them anymore. Nice!

On the other hand the new sub position and the 65Hz crossover is quite a bit worse than before. :/

Before making any new changes, can you tell me:

1. The exact dimensions of your room? Width, height and length.
2. Exactly how far is the center of the subwoofer's driver from all walls, ceiling and floor in your room?
3. Are any boundaries around 1,3m away from the sub?

 

[cristoo]

I did some schemes and found out that I am very poor at it. But here we I have, hope it helps.
BTW I have carpets on the floor on both sides of bed, if it matters.

 

[dominikz]

I did some schemes and found out that I am very poor at it. But here we I have, hope it helps.
BTW I have carpets on the floor on both sides of bed, if it matters.

View attachment 26107View attachment 26108

Thanks for the illustration! It is perfectly adequate!

This makes me suspect the 60-70Hz dip in sub response might be due to where your listening position is (LBIR), rather than where the sub is placed (SBIR).

Can you remeasure just the sub response from a different point in the room - e.g. closer to the middle of the room? If the dip moves or goes away at this alternate measurement location we will know the dip was mainly caused by the location of your MLP.

 

[Wiimer]

@cristoo,

As an aside,
It is faster to use the HouseCurve app to measure the sub.

After running RoomFit, open the Now Playing Screen, and save Room Correction Audio to My Favourites.


Saved tracks can be played back from My Tracks.


You can move the sub around and see the response changes while repeat playing a track.

 

[cristoo]

Thanks for the illustration! It is perfectly adequate!

This makes me suspect the 60-70Hz dip in sub response might be due to where your listening position is (LBIR), rather than where the sub is placed (SBIR).

Can you remeasure just the sub response from a different point in the room - e.g. closer to the middle of the room? If the dip moves or goes away at this alternate measurement location we will know the dip was mainly caused by the location of your MLP.

I tired to do extra subs test with original @Wiimer idea, so cross at 250, range around 6500-8000 and mains off.

Sub in the corner and layout as on the scheme.

I tried different place in my bed as this is generally the place I will listen music to. In real world to feel comfortable to only place is the one I marked on a scheme (so top of bed) but i checked other to see results. I some case i received errors messages in RF (poor quality of measurement).

But to the point

Approx center of bed



End of bed so approx 1-1,5 m to tv stand and speakers. This time results are different



I can’t really move my listening place, in practice form left (current) to right side and that’s it.

 

[lempogi]

Good job!

Out of these I'd choose +10ms, since that gives more energy around the crossover (prior to correction). Try using 10ms, apply the corresponding RoomFit profile and do some listening to see if it sounds OK to you.

Thanks! I'll choose 10ms profile then.

The only 'issue' now is the too high target level set by RoomFit, which results in a few dB extra bass. It is not necessarily a problem, however - many people prefer a little extra bass.

Yes I like it also with a little bit of extra bass. But if in case I will reduce the volume/gain of the sub a little bit, do I need to rerun RoomFit again for a different profile? Or my existing profile will still just be good to use?

Hope you're satisfied with the results!

Super satisfied! Thank you so much for guiding us on making our setup sound so good in the best practical way!

 

[dominikz]

I tired to do extra subs test with original @Wiimer idea, so cross at 250, range around 6500-8000 and mains off.

Sub in the corner and layout as on the scheme.

I tried different place in my bed as this is generally the place I will listen music to. In real world to feel comfortable to only place is the one I marked on a scheme (so top of bed) but i checked other to see results. I some case i received errors messages in RF (poor quality of measurement).

But to the point

Approx center of bed

View attachment 26119View attachment 26120

End of bed so approx 1-1,5 m to tv stand and speakers. This time results are different

View attachment 26122View attachment 26123

I can’t really move my listening place, in practice form left (current) to right side and that’s it.

Given that the 60-70Hz dip is gone in the last measurement position, I'd say this demonstrates that it is mainly caused by your listening position. This means we can't really fix it by moving the sub around.

One more thing you could try is to set the crossover frequency to 50Hz (i.e. below the dip) and remeasure the whole system response at your MLP. I suspect we will again see the same dip, which would be another confirmation that it is caused by LBIR.

 

[cristoo]

Given that the 60-70Hz dip is gone in the last measurement position, I'd say this demonstrates that it is mainly caused by your listening position. This means we can't really fix it by moving the sub around.

One more thing you could try is to set the crossover frequency to 50Hz (i.e. below the dip) and remeasure the whole system response at your MLP. I suspect we will again see the same dip, which would be another confirmation that it is caused by LBIR.

Today’s measurements were thought, RF reported background noise and notice about error (probably because audiable resonance when sub was on high volume). Nevertheless I managed to do several successful attempts. First shot gave very similar graph but what surprised me was the rest were much different although I moved like about 1 m range on the bed. Don’t know why second graph is totally different.

To sum it up

@dominikz is this dip unavoidable ? Can I do anything else ? Try with crossover, phase, change latency (as I moved sub) ? Or It have to as it now and I have to deal with it and just listen to music.

Btw just for info I have class d amp which volume is set and fixed at 50% as I use only volume on Ultra. I haven’t change it since Ultra setup coz I found it well balanced but maybe it can be also tweaks. Don’t know.

Anyway thanks for all help and advices.

 

[dominikz]

Yes I like it also with a little bit of extra bass. But if in case I will reduce the volume/gain of the sub a little bit, do I need to rerun RoomFit again for a different profile? Or my existing profile will still just be good to use?

If you change sub level I'd suggest to re-run RoomFit.

Altenatively you can just control the level of bass with a single low shelf EQ - in that case no need to run RoomFit again.

Super satisfied! Thank you so much for guiding us on making our setup sound so good in the best practical way!

I was happy to help!

I hope the process illustrated the importance of speaker/sub placement and integration for sound quality - even with room correction available!

 

[dominikz]

First shot gave very similar graph but what surprised me was the rest were much different although I moved like about 1 m range on the bed. Don’t know why second graph is totally different.

The second to last graph is some kind of error. The others look fine, and they seem to support my conjecture that your 60-70Hz dip was caused by where your main listening position is. It is expected that changes in response would be pretty significant for different listening spots.

To sum it up

@dominikz is this dip unavoidable ? Can I do anything else ? Try with crossover, phase, change latency (as I moved sub) ? Or It have to as it now and I have to deal with it and just listen to music

It is probably unavoidable at that listening position. But we may still have a few things to try.
First, to be sure I'd suggest to try my suggestion from the previous post with the 50Hz crossover.

 

[dominikz]

Btw just for info I have class d amp which volume is set and fixed at 50% as I use only volume on Ultra. I haven’t change it since Ultra setup coz I found it well balanced but maybe it can be also tweaks. Don’t know.

There's no problem with the amp, IMHO!

 

[cristoo]

The second to last graph is some kind of error. The others look fine, and they seem to support my conjecture that your 60-70Hz dip was caused by where your main listening position is. It is expected that changes in response would be pretty significant for different listening spots.

It is probably unavoidable at that listening position. But we may still have a few things to try.
First, to be sure I'd suggest to try my suggestion from the previous post with the 50Hz crossover.

Ok, another round of testing.

Please note following settings.

Crossover 50hz, speakers latency 9ms, RF 20-400 max min gain 12db, no boost. WHA sub phase 0. Sub WHA 0db, on Sub 10 o’clock

And now for Sub phase 0




But for sub phase 180 (and all the rest parameters the same)



And for the same as take 2 settings HC test sweep



What do you think ?

 

[dominikz]

Thanks! I'll choose 10ms profile then.

Yes I like it also with a little bit of extra bass. But if in case I will reduce the volume/gain of the sub a little bit, do I need to rerun RoomFit again for a different profile? Or my existing profile will still just be good to use?

Super satisfied! Thank you so much for guiding us on making our setup sound so good in the best practical way!

Perhaps you'll be interested to try this - today I wanted to see if there's any way to workaround the current limitation of RoomFit where we target level isn't set optimally due to the fact that the correcting range is also used to calculate the target curve level.

So while normally I'd suggest to set the RoomFit correction range upper frequency to about 400Hz, I wanted to see what happens if I set the upper limit to a much higher frequency - in the below test I chose to use 20Hz to 8000kHz range in RoomFit.

With this setting the target level was correct in both channels (as expected), but I was very pleasantly surprised to see that RoomFit still chose to assign all of its filter bands at frequencies under 500Hz:


Note that this will only work in some cases, mainly those where bass level is significantly higher than the mids and treble (so mids and highs stay below the target), and where a calibrated measurement microphone is used to measure. So it might work in your case, but we can't be sure without testing.

In case you decide to test it out I'd be interested to see your results. If you see RoomFit assigns any filters with significant gains above about 500Hz I'd advise to use the previous calibration instead. At least until WiiM improve their RoomFit algorithms.

In my case this RoomFit profile ("no boost mode") compares very favorably to a correction I created with REW (with only negative gain filters used):

Note that I used a slightly different target for the REW correction, which explains most of the differences.

 

[cristoo]

And 4 phase 180 I did full range RF (20-8000). All settings the same as previous post.

Here are results

And HC

 

[dominikz]

Ok, another round of testing.

Please note following settings.

Crossover 50hz, speakers latency 9ms, RF 20-400 max min gain 12db, no boost. WHA sub phase 0. Sub WHA 0db, on Sub 10 o’clock

And now for Sub phase 0

View attachment 26167View attachment 26168
View attachment 26169

But for sub phase 180 (and all the rest parameters the same)

View attachment 26170View attachment 26171View attachment 26172

And for the same as take 2 settings HC test sweep

View attachment 26166

What do you think ?

I think this shows that indeed the main issue is the location of your MLP, since the 60Hz dip stays regardless whether this part of the spectrum is reproduced by your sub or your speakers.
With this configuration when you invert the subwoofer phase the dip is reduced in the left channel but not the right, which is also not ideal.

All this taken into account, I'd suggest to try the following:
1) Put the subwoofer back to its previous original position
2) Modify the subwoofer crossover frequency to 80Hz, phase to 0°, delay 0ms
3) Modify RoomFit parameters: Freq range 35Hz-500Hz, "No Boost Mode" disabled, Min Gain -12dB, Max Gain +6dB, Max Q 10, Individual Channel RoomFit
4) Re-run RoomFit and post your results.

 

[cristoo]

Did as advised although sub is in the same place. Settings as you suggested. Phase on sub and WHA 180. Latency both for speaker and sub 0 ms. RF seetings as suggested.


Results