WiiM Amp Ultra: Performance Meets Precision

Wiimer · 2025-06-25T19:14:02-0700

I don't know if this helps, but I plugged an IMM-6C mic into my phone and used the noise level meter app to check roughly.

Sound source.

Speaker: KEF R3 META
Sensitivity 87dB (2.83V/1m)

Amp Ultra volume 100%
Pre-gain ±0dB

At 75% volume limit the sound level meter showed 89dB (1m).

If the volume is increased further, the sound gets louder, but meter value does not increase any further.
#This meter app had a limit of 89 dB...

canard · 2025-06-25T23:34:24-0700

android
"sound analyser app"
very complete... even too complete ;-)

Wiimer · 2025-06-25T23:49:53-0700

canard said:
android
"sound analyser app"
very complete... even too complete ;-)
(if possible to use after a calibration... (but in my case with an old Samsung surprisingly correct))

Is it this app I always use?

Or other one?

canard · 2025-06-25T23:54:35-0700

Wiimer said:
Is it this app I always use?
View attachment 23221
Or other one?

(I am lucky enough to have access to professional calibrated sound level meters to carry out their calibration correctly enough for our modest uses...practical this app...)

canard · 2025-06-26T00:18:27-0700

canard said:
(I am lucky enough to have access to professional calibrated sound level meters to carry out their calibration correctly enough for our modest uses...practical this app...)

(((Not the topic of the thread... but there's a really great little "simple" Android app, even if it's old, unfortunately not "supported" (no CAL file integration, etc.). Spectroid... Just, have any of you had the opportunity to see how it supports the little Dayton USB? If relevant, etc. Thanks in advance ;-) )))

Wiimer · 2025-06-26T00:18:34-0700

canard said:
(I am lucky enough to have access to professional calibrated sound level meters to carry out their calibration correctly enough for our modest uses...practical this app...)

I don't know how to do calibration, do I just play the pink noise on WiiM and start the calibration in app?

I've got 107 dB at 75% volume...

canard · 2025-06-26T00:20:53-0700

Wiimer said:
I don't know how to do calibration, do I just play the pink noise on WiiM and start the calibration in app?

I've got 107 dB at 75% volume...
View attachment 23227

problem....forgot it in your case ..hihi
calibration is done by comparison with another calibrated machine..
(with or without dayton?)
;-)

"help" is important ;-)
(but in any case these approaches are not suitable for very high level measurements)

dominikz · 2025-06-26T00:41:28-0700

harkpabst said:
So, it's certainly not impossible that 0.5 dB can make a perceivable difference. But generally speaking, no, our hearing is not "infinitely sensitive", neither to deviations in the frequency response, nor to effects in the time domain. 0.2 dB at 250 Hz is more or less nothing.

Let me be clear: I agree in principle, and my comment was not meant to imply otherwise. Our hearing is absolutely not "infinitely sensitive", and when talking about 0,5dB deviations we're absolutely talking about subtle changes.

However, people tend to compare and discuss various audio equipment specs which are almost certainly beyond human audibility limits, so I felt it was important to note that in this case we're discussing something that *can* be audible.

Secondly, please note that in the statement in my previous post I noted two important qualifiers:

Resonance Q-factor: wide-bandwidth (low-Q) resonances are much more easily audible than sharp, low-bandwidth (high-Q) resonances.
Playback content: pink noise will make it much easier to hear resonances than e.g. audiophile jazz.

This is well illustrated in dr. Toole's amazing book:

The above illustration is adapted from an AES article by Toole, F.E., and Olive, S.E. (1988). “The Modification of Timbre by Resonances: Perception and Measurement,” J. Audio Eng. Soc., 36, pp. 122–142

harkpabst said:
Few manufacturers of speaker drivers will provide you with drive units (woofers, tweeters, no difference) matched to tolerances any tighter than ±0.5 dB unless you spend crazy money for matched pairs. And it's sooo easy to lose one dB of efficiency, even over a broad range.

Even more so - NONE of the >1300 loudspeaker model measured and collected on spinorama.org reach even the 0,5dB flatness between 300-5000Hz (reference from spinorama.com). The really good ones are usually around 1-2dB.

And then we have the room, which will wreak havoc on the bass response - often causing deviation of more than 10dB.

However, all this in itself does not mean that we shouldn't strive for better performance of other components. FR deviations are linear and they add-up.

It always seems strange to me when people focus on THD figures at 15kHz and step response (and we know that people are very bad at detecting non-linear distortion and differences in time-domain waveform shape), but then disregard frequency response deviations caused by high output impedance (while we know people are pretty good at hearing wideband FR deviations).

Again, <1dB deviations are subtle - but at least they have a chance of being audible.

harkpabst said:
There is a great interview with legendary mastering engineer Bog Ludwig published by "Produce Like A Pro" on YouTube. Everyone should watch at least a couple of minutes of this interview, starting at timestamp 22.10. He mentions how incredibly superior digital mastering is regarding EQ, being able to adjust things down to 1/10 of a dB (sic!). But he also explicitly explains how amazed he was that few exceptional artists (presenting the Indigo Girls as an example) could reliably tell a 0.2 to 0.6 dB EQ setting apart. And that's the artists themselves, in a studio environment, listening to their very own stuff. I honestly believe that it takes extremely well trained ears to be capable of this.

As mentioned above - we really can't be only talking about gain when speaking of resonance audibility thresholds. A high-shelf EQ at e.g. 2kHz with gain +0,5dB applied to pink noise would be relatively easy to ABX for most people.
On the other hand a peaking EQ with Q=10 and gain of -5dB at 16kHz applied to an acoustic jazz recording would most likely be impossible to ABX for most people.

If I manage to find the time, I might prepare a simple ABX test so that people can test this for themselves.

dominikz said:

Lastly, let me comment a bit more on the above graph.

The maximum deviation (difference between maximum and minimum value) in the above diagram is:

WiiM Vibelink Amp: 0,24dB (so I estimate the damping factor of around 90)
WiiM Amp Ultra: 0,33dB (so I estimate the damping factor of around 65)
- Note: WiiM Amp Ultra product page states a damping factor of 94 - @RyanWithWiiM Could you double check this with the product manager? In general it would be really great to see output impedance (magnitude and phase) vs frequency measurements for Amp, Amp Pro, Amp Ultra and the Vibelink Amp. Those measurements would allow people to calculate exactly the FR deviations they could expect with any known load.

So as you can see, the Vibelink Amp only does 0,09dB better than the Amp Ultra in this test! This is almost certainly an inaudible difference between the two amps, even in the most revealing and academic of use-cases - and especially so with real media (like music or movies).

However, while both amps are pretty good as far as load (in)sensitivity goes, neither are really great.
You need a damping factor above 200 to achieve <0,1dB frequency response variation with most loads (assuming zero-impedance for the loudspeaker cable - which is impossible). To achieve the same response variation with a 6 feet piece of 12AWG loudspeaker cable between the amp and the load you need a damping factor above 360.
(You can play with values and calculate yourself the deviation for various combinations in the spreadsheet linked in this great article by Benchmark. Both thickness and length of a loudspeaker cable influence its electrical resistance.)

As the previous sentence hinted, *any* real-world loudspeaker cable will have some electrical resistance which will add to the source impedance seen by the loudspeaker, and therefore increase the frequency response variation. This is one example where a cable does make a difference (but that still doesn't mean that a cable needs to be expensive to be good

).

For example, adding a 6 feet piece of 12AWG loudspeaker cable between the Vibelink Amp and Erin's complex load would increase the maximum response deviation at the loudspeaker terminals from 0,24dB to 0,29dB.

Again, WiiM Amp Ultra seems to measure reasonably well as far as output impedance goes and the resulting (small) response deviation should not be a limiting factor for most people. There is IMHO much more sound quality to be gained by focusing on optimizing loudspeaker placement, subwoofer integration and combating room resonances by careful application of PEQ; rather than worrying about a fraction of a difference in FR caused by damping factor.

Still, it is not bad to try and understand it all an put it into proper context!

Anyway, that's just my 2 cents! Sorry for another (very) long post!

canard · 2025-06-26T00:46:31-0700

dominikz said:
Let me be clear: I agree in principle, and my comment was not meant to imply otherwise. Our hearing is absolutely not "infinitely sensitive", and when talking about 0,5dB deviations we're absolutely talking about subtle changes.

However, people tend to compare and discuss various audio equipment specs which are almost certainly beyond human audibility limits, so I felt it was important to note that in this case we're discussing something that *can* be audible.

Secondly, please note that in the statement in my previous post I noted two important qualifiers:

Resonance Q-factor: wide-bandwidth (low-Q) resonances are much more easily audible than sharp, low-bandwidth (high-Q) resonances.

Playback content: pink noise will make it much easier to hear resonances than e.g. audiophile jazz.

This is well illustrated in dr. Toole's amazing book:

The above illustration is adapted from an AES article by Toole, F.E., and Olive, S.E. (1988). “The Modification of Timbre by Resonances: Perception and Measurement,” J. Audio Eng. Soc., 36, pp. 122–142

Even more so - NONE of the >1300 loudspeaker model measured and collected on spinorama.org reach even the 0,5dB flatness between 300-5000Hz (reference from spinorama.com). The really good ones are usually around 1-2dB.

And then we have the room, which will wreak havoc on the bass response - often causing deviation of more than 10dB.

However, all this in itself does not mean that we shouldn't strive for better performance of other components. FR deviations are linear and they add-up.

It always seems strange to me when people focus on THD figures at 15kHz and step response (and we know that people are very bad at detecting non-linear distortion and differences in time-domain waveform shape), but then disregard frequency response deviations caused by high output impedance (while we know people are pretty good at hearing wideband FR deviations).

Again, <1dB deviations are subtle - but at least they have a chance of being audible.

As mentioned above - we really can't be only talking about gain when speaking of resonance audibility thresholds. A high-shelf EQ at e.g. 2kHz with gain +0,5dB applied to pink noise would be relatively easy to ABX for most people.
On the other hand a peaking EQ with Q=10 and gain of -5dB at 16kHz applied to an acoustic jazz recording would most likely be impossible to ABX for most people.

If I manage to find the time, I might prepare a simple ABX test so that people can test this for themselves.

Lastly, let me comment a bit more on the above graph.

The maximum deviation (difference between maximum and minimum value) in the above diagram is:

WiiM Vibelink Amp: 0,24dB (so I estimate the damping factor of around 90)

WiiM Amp Ultra: 0,33dB (so I estimate the damping factor of around 65)

Note: WiiM Amp Ultra product page states a damping factor of 94 - @RyanWithWiiM Could you double check this with the product manager? In general it would be really great to see output impedance (magnitude and phase) vs frequency measurements for Amp, Amp Pro, Amp Ultra and the Vibelink Amp. Those measurements would allow people to calculate exactly the FR deviations they could expect with any known load.

So as you can see, the Vibelink Amp only does 0,09dB better than the Amp Ultra in this test! This is almost certainly an inaudible difference between the two amps, even in the most revealing and academic of use-cases - and especially so with real media (like music or movies).

However, while both amps are pretty good as far as load (in)sensitivity goes, neither are really great.
You need a damping factor above 200 to achieve <0,1dB frequency response variation with most loads (assuming zero-impedance for the loudspeaker cable - which is impossible). To achieve the same response variation with a 6 feet piece of 12AWG loudspeaker cable between the amp and the load you need a damping factor above 360.
(You can play with values and calculate yourself the deviation for various combinations in the spreadsheet linked in this great article by Benchmark. Both thickness and length of a loudspeaker cable influence its electrical resistance.)

As the previous sentence hinted, *any* real-world loudspeaker cable will have some electrical resistance which will add to the source impedance seen by the loudspeaker, and therefore increase the frequency response variation. This is one example where a cable does make a difference (but that still doesn't mean that a cable needs to be expensive to be good ).

For example, adding a 6 feet piece of 12AWG loudspeaker cable between the Vibelink Amp and Erin's complex load would increase the maximum response deviation at the loudspeaker terminals from 0,24dB to 0,29dB.

Again, WiiM Amp Ultra seems to measure reasonably well as far as output impedance goes and the resulting (small) response deviation should not be a limiting factor for most people. There is IMHO much more sound quality to be gained by focusing on optimizing loudspeaker placement, subwoofer integration and combating room resonances by careful application of PEQ; rather than worrying about a fraction of a difference in FR caused by damping factor.

Still, it is not bad to try and understand it all an put it into proper context!

Anyway, that's just my 2 cents! Sorry for another (very) long post!

thank you
(Post in thread 'WiiM Amp Ultra: Performance Meets Precision' https://forum.wiimhome.com/threads/wiim-amp-ultra-performance-meets-precision.7167/post-133267)

canard · 2025-06-26T00:58:29-0700

Everything becomes "acceptable"... very demagogic...

But yes, for 95% of what is listened to in the world (by 96% of people), which typically supports/supports compression like MP3 or AAC at less than 192kb it's not really important...

;-)
(the real subject...is cognitive sciences..."learning and expertise" ;-) )

dominikz · 2025-06-26T01:13:12-0700

Mr Ee said:
For some recordings, although 100 was loud it wasn't ridiculously loud - the amp ultra ran out of puff / gain. Increasing pre-gain did increase the volume, but it didn't sound as refined.

Could be that by boosting pre-gain some peaks which were close to 0dBFS went into digital clipping.

In general, streaming content is usually mastered to about -14LUFS, which usually means that the average level is around -14dBFS. Peaks (on wide dynamic range content) can still reach up 0dBFS, meaning that boosting pre-gain could push them over 0dBFS and into clipping.

To put the power demands into perspective - with volume set to maximum, using a 4Ohm load, playing a track that has -14 dBFS average level and 0dBFS peaks the Amp Ultra should be using only about 8W on average and the full 200W for the highest peaks.

How loud this would be perceived would mainly depend on personal preference, listening distance, size of the room and speaker sensitivity. I wrote recently about this in a bit more detail in this post.

I believe the main thing to highlight here is that the Vibelink Amp and the Amp Ultra don't have significantly more power than the Amp/Amp Pro. I.e. people who find that 60W is not enough for their circumstances will likely find that 100W is insufficient as well.

canard · 2025-06-26T01:17:21-0700

dominikz said:
Could be that by boosting pre-gain some peaks which were close to 0dBFS went into digital clipping.

In general, streaming content is usually mastered to about -14LUFS, which usually means that the average level is around -14dBFS. Peaks (on wide dynamic range content) can still reach up 0dBFS, meaning that boosting pre-gain could push them over 0dBFS and into clipping.

To put the power demands into perspective - with volume set to maximum, using a 4Ohm load, playing a track that has -14 dBFS average level and 0dBFS peaks the Amp Ultra should be using only about 8W on average and the full 200W for the highest peaks.

How loud this would be perceived would mainly depend on personal preference, listening distance, size of the room and speaker sensitivity. I wrote recently about this in a bit more detail in this post.

I believe the main thing to highlight here is that the Vibelink Amp and the Amp Ultra don't have significantly more power than the Amp/Amp Pro. I.e. people who find that 60W is not enough for their circumstances will likely find that 100W is insufficient as well.

and here we come back to the subject...of behavior when faced with the load...
;-)

harkpabst · 2025-06-26T02:42:47-0700

canard said:
and here we come back to the subject...of behavior when faced with the load...
;-)

No. Not related at all.

Load dependency has nothing to do with gain or power output.

canard · 2025-06-26T02:54:18-0700

harkpabst said:
No. Not related at all.

Load dependency has nothing to do with gain or power output.

That's not what I meant to point out...(and precisely)
(here the interest of the ultra will perhaps be in certain circumstances the fact that it is more comfortable at low impedance, etc.not power...)
no problem
;-)

harkpabst · 2025-06-26T02:55:42-0700

canard said:
That's not what I meant to point out... no problem
;-)

Then try to rephrase your point, please.

canard · 2025-06-26T02:59:25-0700

harkpabst said:
Then try to rephrase your point, please.

not important...let's stay there... ;-)
(maybe you will have understood it....writing is difficult for me even in my own language... ;-)
and then here is nothing very original in my words...no interest)

dominikz · 2025-06-26T04:19:03-0700

canard said:
That's not what I meant to point out...(and precisely)
(here the interest of the ultra will perhaps be in certain circumstances the fact that it is more comfortable at low impedance, etc.not power...)
no problem
;-)

I believe I now see what you are aiming at, and I agree that it seems likely that the Amp Ultra will be able to provide more current (due to dual-TPA3255), so should in theory be better suited to drive low-impedance loudspeakers (compared to Amp or the Amp Pro, each having just a single TPA3255).

However, even with dual-chips, good thermal management is key to drive low impedances with high power for prolonged periods. I hope that this is something we'll be able to read from Amp Ultra measurement reports (once they are published).

WiiM Amp Ultra: Performance Meets Precision

​

Key Features​

Versatile Connectivity & Streaming​

Smart Home Audio Control​

Designed for Cinematic & High-Fidelity Listening​

Future-Proof & Always Improving​

Pricing & Availability​

Attachments

Major Contributor

Major Contributor

Major Contributor

Major Contributor

Attachments

Major Contributor

Major Contributor

Major Contributor

Senior Member

Major Contributor

Major Contributor

Senior Member

Major Contributor

Major Contributor

Major Contributor

Major Contributor

Major Contributor

Senior Member

Similar threads

Key Features

Versatile Connectivity & Streaming

Smart Home Audio Control

Designed for Cinematic & High-Fidelity Listening

Future-Proof & Always Improving

Pricing & Availability