PEQ Beta - Testing Experience

[Brantome]

Please use this thread only to discuss your personal experience with the testing of the PEQ feature and the beta firmware you requested. This thread should not replace submitting tickets as that route provides WiiM support with the necessary device and app logs to diagnose and hopefully fix any issues.

From what WiiM have already posted:

“For those testing please see the tutorial below

• PEQ

We appreciate your ongoing participation as a Beta tester.

The PEQ firmware will be released early next week.

If you encounter any issues during the update process or while using this function, please do not hesitate to contact us. Your feedback is invaluable to us.

Thank you for your continued support and contribution to our Beta testing program!”

 

[Steve Woodhouse]

Hi Steve,

If the Beta testing proceeds smoothly, we aim to officially launch this version by the end of August. We will not stop there, though; our ongoing improvements will include expanding the number of bands and enabling independent channel Parametric Equalization (PEQ). Thank you for your support!

Right, I’m going to hold you to that.

JOKING! Thanks for replying so promptly, let’s hope everything runs smoothly and you can manage that.

Good luck with everything.

 

[qwertzuiop]

Hi Team,

first of all, thanks for finally bringing us the PEQ and letting us test it.

Tester here from Germany and wanted to also provide feedback.
First of all, the PEQ implementation is well done with its 4 Bands and sounds as I would expect a good PEQ shoudl sound. It works flawlessly with all sources/services and is convenient to use. The possibility to save 4 Presets and easily switch is awesome (you should also make more than one preset for GEQ available in the future).

Some bugs I found in the implementation:
- For the two Shelf filters, the Q factor seems to have no influence/does not work. Changing Q does neither change EQ curve nor the sound, it seems to me like it stays at a fixed Q=1 although you can adjust it (for the two peaks Q works fine)
- As someone already mentioned, switching from PEQ to GEQ and back stops playing completely (Tidal Connect). No real issue for me, but maybe good for you to know. Turning PEQ on/off and changing profiles works flawless.
- The idea of adjusting frequency and gain with the sliders is nice and convenient, but please add a manual entry like for Q. On a phone screen being a very sensitive two way slider, it is impossible to set exactly values (I guess many people using PEQ have their exact filter values defined and setting exactly those is near impossible).

Some ideas/wishes for the future (you already mentioned you plan to do that, please do that and it would be awesome)
- Let us choose the filter type or at least set the Shelf filter to peak as well. I do use Low Shelf often, but never use High Shelf (just not practical for EQing headphones to Harman), so for me this practically becomes a 2 or 3 band PEQ only which is just not enough to be of real use
- And of course even better would be: do more bands, 5 free choosable bands or even better 10 would make it perfect. Adding separate L/R would be even greater.

For now, well done, keep those awesome new features on, looking forward to future improvements!

 

[D1N0]

- The idea of adjusting frequency and gain with the sliders is nice and convenient, but please add a manual entry like for Q. On a phone screen being a very sensitive two way slider, it is impossible to set exactly values (I guess many people using PEQ have their exact filter values defined and setting exactly those is near impossible).

I second this. manual entry or a arrow control or a landscape mode.

Also when I touch the bottom scale of the graph my android text selection kicks in.

 

[Wiimer]

I second this. manual entry or a arrow control or a landscape mode.

Also when I touch the bottom scale of the graph my android text selection kicks in.
View attachment 1972

Well, I thought too.

 

[joffy1780]

I know it's been part of the roadmap for some time, but landscape mode would be a welcome addition in general.

 

[AnyNameIWish]

GUI: Need manual input on values
PEQ: Need to be able to change LS and HS to PK
PEQ: Would like to be able to either have separate settings for left and right or a setting L,R,L+R on each filter

Now to the my extremely weird and very subjective observation: If I have GEQ OFF and switch to PEQ ON and then PEQ OFF, it seems like the sound is much clearer!!???

I don’t know if I can trust my ears but please read again: I actually think that that the PEQ sounds clearer than the GEQ even if both are set to OFF! (I guess I have to join the tin foil hat club…)

You need to toggle the ON/OFF switch to swap between the two eqs…

 

[pvog]

- For the two Shelf filters, the Q factor seems to have no influence/does not work. Changing Q does neither change EQ curve nor the sound, it seems to me like it stays at a fixed Q=1 although you can adjust it (for the two peaks Q works fine)

Yeah I don’t see any change in my measurements when I adjust the q for a shelf filter.
So far the pk filters work as expected.

I agree more bands and especially separated for each channel would be great. And the ability to choose between pk, ls and hs for each of them.

Anyway great first implementation.

 

[onlyoneme]

As I got my PEQ FW - 4.8.525960, time for some testing. And first note - PEQ affects the SQ even if with 0 dB gain. A small comparison for 1000 sine signal, 0 dBFS:

 

[onlyoneme]

What happens when we limit the source signal level? The same 0 dB PEQ gain:



Noise floor stays the same, signal resolution is limited to 16 bit depth. Distortions eliminated, - 0.5 dB signal attenuation was enough.

 

[merifon]

resolution is limited to 16 bit depth.

Isn't it a little more than 16bit?

 

[onlyoneme]

Isn't it a little more than 16bit?

Let's say more or less 16 bit

 

[onlyoneme]

I will look at the filters behavior later today. I will probably take ADI-2 behavior as the reference one.

 

[onlyoneme]

At WiiM's lab, they could sure do better...
Is the reported SNR referred to worst noise level at low frequencies? In order to listening results, could it be "weigthed", since level at more hearable frequencies, is consistently lower?

It's in the span range I use, starts at 10 Hz.

 

[lassehoej]

I have done some more testing, and found that all four filters cannot use a Q of more than 1.0. Lower Qs have an effect/change, but above 1.0, all stays the same (effectively Q 1.0).
Also, large adjustments of positive gain alters the dynamics when listening to music. Some kind of soft limiter so that the overall volume goes down in passages of music with heavy bass beat, if the bass i adjusted to a high gain, and then the volume goes up again when the beat stops.

 

[Smartplug]

I will look at the filters behavior later today. I will probably take ADI-2 behavior as the reference one.

Thanks for testing the beta using the test bench you have. This will help wiim to fix the bug and speed up its public release. Me and the rest of the members of this forum appreciate what you do.

 

[onlyoneme]

I can fully confirm that effective range of Q factor is 0.01 - 1.00, regardless of the filter type.

 

[onlyoneme]

It wasn't easy at all to set the same PEQ settings on both ADI and WiiM units ...

Negative gain test:

LS 82 -1.5 0.9
PK 340 -6.5 0.5
PK 1980 -0.5 1.0
HS 9900 -2.5 0.5






Very similar results.

 

[onlyoneme]

How these PEQ settings above affect the multitone test signal? Results below:




The elevated noise floor level affects this severely, but otherwise no problem.

 

[onlyoneme]

These noise level issues look similar to the ones observed by ASR user mdsimon2 here:

DSP Measurements and Rising Noise Floor

In several threads related to DSP platforms I've seen requests for measurements with filters engaged. I thought it unlikely that implementing filters in a DSP would have measurable effect on noise / distortion other than the obvious change in the fundamental, however @Sokel recently mentioned...

www.audiosciencereview.com

 

[merifon]

Is the noise floor totally indipendent from signal level?