KEF LSX Wireless & DAC/AMP

[roryoconnor]

I don't set it to line out. I adjust the volume to match the speaker amp, until it sounds ' just right' .

Odd that there are many LSX ( and LS50w) owners who report the same effect.

Maybe we have all taken leave of our senses ..

 

[Frank Dernie]

I don't set it to line out. I adjust the volume to match the speaker amp, until it sounds ' just right' .

Odd that there are many LSX ( and LS50w) owners who report the same effect.

Maybe we have all taken leave of our senses ..

PMFJI.
I have no dog in this fight but:
The analogue input of the LSX will go through an ADC, whatever DSP is included in the design then its own internal DACs and amps.
The most direct input would be to a digital input, bypassing the ADC. This will cause the least manipulation of the sound.
Using a separate DAC to convert the signal to analogue first, only for the speaker internals to have to convert it back to digital in order for the speaker to function, makes no sense, logically or technically.
If the external DAC was not a particularly good one and altered the frequency response it may sound "different" and may be preferred but it most definitely is not "better" in any technical sense.
Besides, the Mojo is not a coloured DAC so in this context that is a moot point.
In summary, there is no techical explanation that supports your premise. At best, with good kit, which it is, it would cause no degradation. It is an expensive way to achieve nothing though.
I don't doubt you hear a difference but the only scientific explanation for that is the placebo effect.

 

[VintageFlanker]

Odd that there are many LSX ( and LS50w) owners who report the same effect.

There are many people who "hear' night and day differences by burn-in their audiophile power cable, if we go there...
The fact than few has reported this or that only creates some kind of confirmation bias for others, IMHO. I remember the Qutest's thread on HF being full of people hearing better sound by swapping PSU for their Qutest... Even when Rob Watts himself came to tell that it's impossible.

Again, two DACs may sound different, even through an ADC. Two good DACs will sound the same tho. Mojo and E30 are most probably (99.99%) indistinguishable by ear. Still,

I don't set it to line out. I adjust the volume to match the speaker amp, until it sounds ' just right' .

You adjusted the Mojo manually until it sounds just right... Well, that explains a lot. So, what we are talking about there, is volume ajustement relative to the LSX. Two DACs with different output voltage will sound different, but only because one plays louder than the other. There is no such thing as sound quality involved to begin with.
(In fact: Your solution is objectively worse from a technical standpoint)
We're talking about audible differences only because of the volume control.

 

[roryoconnor]

Fair enough. I agree what you both say makes sense. However, it's not the complete picture. I know what I hear....and there is a difference in sound quality.

The E30 and Mojo have minor differences in presentation, Mojo is warmer with female vocals and handles sibilance better, with greater separation between instruments.

There is a clear difference between the cheaper DACs in the Chromecast Audio and Hifime S2. They manifest this difference with clunkier dynamics and duller imaging.

I wouldn't want to claim immunity from the placebo effect, but I don't think it can account for the entire experience. Nor does collective hysteria amongst KEF active speaker owners make any sense either. I'm sure, like me, they are for the most part reasonably sober and well adjusted individuals with a good grasp of their own faculties

 

[mel]

Fair enough. I agree what you both say makes sense. However, it's not the complete picture. I know what I hear....and there is a difference in sound quality.

The E30 and Mojo have minor differences in presentation, Mojo is warmer with female vocals and handles sibilance better, with greater separation between instruments.

There is a clear difference between the cheaper DACs in the Chromecast Audio and Hifime S2. They manifest this difference with clunkier dynamics and duller imaging.

I wouldn't want to claim immunity from the placebo effect, but I don't think it can account for the entire experience. Nor does collective hysteria amongst KEF active speaker owners make any sense either. I'm sure, like me, they are for the most part reasonably sober and well adjusted individuals with a good grasp of their own faculties

Hi,
I am a software developer. I studied some audio in college physics classes. I have spent hours trying to pinpoint the effect you describe in my small bedroom.

I suspect pre-amp processing or your room might explain the re-digitizing effect, rather than the DACs. Could you describe your room in more detail?

• Have you tried moving your equipment to a different room?
• Have you tried adjusting the sound quality with the DSP equalizer frequency controls from the LSX app? These pre-amp functions are part of the LSX DSP processing. Look at the LSX block diagram for more detail.

I attribute the effect to my small bedroom (roughly 12'x12') being connected to two larger rooms through two doors that are 32" wide. Aligning any of my speakers, especially a decent soundbar, with the narrow door openings results in the best sound quality. The best results are achieved by placing the soundbar at the foot of my bed, so they align with the doorway.

• I generally play jazz or classical music at low volume. I occasionally listen to rock, like U2.
• I source from Tidal Masters using my MacBook Pro M1.
• I connect to the receiver via USB-C using a HDMI adapter to attach to the AV receiver.

I started with a high quality RME ADI-2 DAC. I got equivalent results from pre-amp output from a Yamaha AV receiver using "Straight", 2.1 channel output. I recall from memory that the DAC did not make the most significant difference.

I have AudioEngine A3 speakers connected to two REL subwoofers. I have also experimented with passive bookshelf speakers.

I found a big difference when first connecting the AV pre-amp output to a subwoofer RCA inputs, then to a soundbar using a RCA splitter to 3.5mm sound jack as input to the sound bar. The big difference came after adjusting the subwoofer crossover and the pre-amp bass and treble db.

Someone posted about clicking and popping sounds. I isolated the issue to an open Safari window on my MacPro M1.

 

[mel]

Hi,
I am a software developer. I studied some audio in college physics classes. I have spent hours trying to pinpoint the effect you describe in my small bedroom.

I suspect pre-amp processing or your room might explain the re-digitizing effect, rather than the DACs. Could you describe your room in more detail?

• Have you tried moving your equipment to a different room?
• Have you tried adjusting the sound quality with the DSP equalizer frequency controls from the LSX app? These pre-amp functions are part of the LSX DSP processing. Look at the LSX block diagram for more detail.

View attachment 135332

I attribute the effect to my small bedroom (roughly 12'x12') being connected to two larger rooms through two doors that are 32" wide. Aligning any of my speakers, especially a decent soundbar, with the narrow door openings results in the best sound quality. The best results are achieved by placing the soundbar at the foot of my bed, so they align with the doorway.

• I generally play jazz or classical music at low volume. I occasionally listen to rock, like U2.
• I source from Tidal Masters using my MacBook Pro M1.
• I connect to the receiver via USB-C using a HDMI adapter to attach to the AV receiver.

I started with a high quality RME ADI-2 DAC. I got equivalent results from pre-amp output from a Yamaha AV receiver using "Straight", 2.1 channel output. I recall from memory that the DAC did not make the most significant difference.

I have AudioEngine A3 speakers connected to two REL subwoofers. I have also experimented with passive bookshelf speakers.

I found a big difference when first connecting the AV pre-amp output to a subwoofer RCA inputs, then to a soundbar using a RCA splitter to 3.5mm sound jack as input to the sound bar. The big difference came after adjusting the subwoofer crossover and the pre-amp bass and treble db.

Someone posted about clicking and popping sounds. I isolated the issue to an open Safari window on my MacPro M1.

I forgot to mention that I found bass ports make a substantial difference in small rooms. The AudioEngine A3s have front bass ports. My passive bookshelf speakers are rear ported. The soundbar has no bass port, but is connected to a nice subwoofer.

 

[mel]

You might find the DSP architecture description useful.

'Music Integrity Engine' is the name KEF gives to its custom DSP implementations. As you'd expect, hardware and software are similar in the LSX and LS50W but, given the price disparity, the LSX uses a less powerful DSP chip. As a result, whereas the LS50W runs at an internal sampling rate of 192kHz at all times, the maximum in the LSX is 96kHz – and this requires a wired connection between the two speakers otherwise it's limited to 48kHz.

The key algorithms in the two speakers perform the same tasks: crossover filtering, bass management and equalisation. Low-pass and high-pass crossover filtering are performed using IIR digital filters, with phase correction using a long FIR filter. While KEF's head of acoustics Jack Oclee-Brown describes the latter as 'relatively textbook', it is tweaked to reduce pre-ringing in the step response to avoid the treble sound becoming glassy. Dynamic bass extension lowers the LF corner frequency at low sound outputs, and adjusts it upwards to protect the driver at higher bass levels. KH

 

[mel]

Here's another take: The important point is the optical and 3.5mm connection is intended for "TVs and portable devices". I get the impression one owner actually wanted the LS-50 Wireless 2 features, rather than LSX, based on that post.

My understanding is the LSX is more appropriate for an iPhone, while the LS-50 is designed for a MacBook Pro connection.

I think the tweeter plays an important role in a small room for certain types of music. I happen to be a jazz fan. My understanding is a piano will sound different, given different tweeters, especially when powered by different types of amps. Something about the speed of the amps. My understanding is this also applies to subwoofers, but to a lesser degree.

Rather than combining a Class A/B circuit (to feed the tweeter) with a Class D module (to power the mid/bass unit) as the LS50 Wireless does, the LSX’s amplification is purely Class D. Such a design is compact as well as being more power efficient. This reduces the need for large, bulky heatsinks and means KEF can keep the plastic cabinets (which used to be MDF on the LS50) as small as possible.

The LSX system loses a couple of its bigger sibling’s connections: USB Type B (for computer and laptop hook-ups) and RCAs for connecting legacy hi-fi kit. But KEF has evolved its stereo speaker system concept on the LSX. Since the LS50 Wireless was launched last year, the company has developed its proprietary, wireless inter-speaker connection so, unlike that set-up, the LSX system doesn’t require an ethernet cable to connect the two speakers.

The system can play hi-res files all the way up to 24-bit/192kHz, but will downsample to either 24-bit/48kHz if the speakers are connected to each other wirelessly or 24-bit/96kHz if you do decide to run the supplied ethernet cable between them.

While the LSX’s master speaker overlooks its sibling’s legacy RCA and USB type-B connections, its optical and 3.5mm aux inputs provide means to connect TVs and portable devices, while a subwoofer output offers the opportunity to add more bass to the system.

You can always add a bit more bass by tweaking the balance in the system's EQ settings, which can be found in KEF’s Control app. The app is also used to get the LSX on your network, apply room settings, perform firmware updates and switch sources. The latter can also be performed via the supplied remote – as can volume adjustment – but the LSX lacks the on-speaker touch controls of its sibling.

 

[mel]

LSX Spec sheet
Uni-Q driver array:
HF: 19mm (0.75in.) aluminium dome
LF/MF: 115mm (4.5in.) magnesium/aluminium alloy cone

LF: 70W HF: 30W

LS-50 Spec sheet

Titanium Grey Mineral White Crimson Red Special Edition
LS50 Wireless II
Uni-Q Driver Array:
LF: 130 mm (5.25 in.) aluminium cone
HF: 25 mm (1 in.) vented aluminium dome with Metamaterial Absorption Technology

LF: 280W, HF: 100W
LF: Class D, HF: Class AB
108 dB

 

[mel]

Here's another piece of the puzzle:

"The big difference is that with the DSP preamp we have a lot more control over the crossover, so the response is smoother on the LS50W than the LS50. Inside the DSP we monitor the power and excursion of the driver and step in if it is likely to distort."

 

[mel]

Here's another comment that supports re-digitizing.

I still return to the same conclusion after hours of reading and experimenting in a small bedroom. I do not believe the DAC or double DACing plays the predominant role. My impression is the DSP equalization, crossover and amp control have a more noticeable effect. The type of amp probably is significant, too.

 

[mel]

My guess is the larger LS-50 Wireless 2 is counter-intuitively a better choice for my small bedroom, than the LSX. The larger tweeter is one reason. Jazz piano, drums, bass and brass instruments are important instruments to me. Piano and/or brass instruments usually carries the melody. Wind instruments are a more general term for brass instruments. I also like clarinets, flutes, etc..., but they are rare. Drums and bass carry the rhythm. The dynamics and/or harmony can be adjusted with equalization. Stringed instruments, other than bass, are not high on my list, along with vocals.

DSP processing in the following block diagram will probably have the largest apparent difference. DSP control over the amp is probably important, too. How amplification is achieved is too invisible to me.



I got the best sound quality in a small bedroom by only using the AV receiver pre-amp functions, rather than the amplifier/passive bookshelf speakers.

I passed the output signal via RCA line out from the pre-amp output to the subwoofer. I adjusted the crossover frequency from the subwoofer. I passed the remaining signal via RCA Y line to a 3.5mm input to a sound bar.

The tone controls on the receiver served as an adequate approximation to determine the operative factors in my small bedroom sound dynamics. Bypassing the receiver amplifier for the passive bookshelf speakers seems like an important finding.

I could have achieved more precise results with my RME studio DAC and all of its sophisticated features, but I spent so much time just trying to understand how sound fills my small bedroom, that I lost patience. The DAC seems to be the least important factor.

I connect to Tidal Masters as the source with a MacPro M1. The LS-50 has better laptop connectivity than the LSX. Tidal is not very reliable, so I would connect with USB, not the KEF control app directly to Tidal.

I played quite a variety of music. I did find the RME DAC made a huge different on Coltrane's sax in this recording:

https://tidal.com/browse/track/77621945

Streaming Functionality
Once you've actually got things set up, you'll gain access to a number of different streaming services. The LSX not only offer Spotify Connect, Tidal, and Bluetooth audio, but also give you access to Apple AirPlay 2. That means that, for the most part, listening to music is a breeze, and can be done within moments of activating the speakers. We also love the fact that they worked directly from the Spotify app, bypassing KEF's apps entirely. However, we did have some issues streaming from Tidal; despite selecting gapless playback, we found that there were huge gaps between songs, and that our play/pause commands were laggy as hell. We had the same issue across multiple Wi-Fi networks, and it was pretty frustrating. If you are a Tidal listener, you may want to look for another pair of speakers, or a single system like the $1,090 Devialet Reactor. The sound quality won't be quite as good, but the ease-of-use is top-notch

 

[mel]

Another thread on this topic about the DAC in the overall scheme of sound.

 

[mel]

I like this subwoofer design:

https://us.kef.com/kc62-subwoofer.html

 

[mel]

Two important distinctions:

Nice diagrams about passive vs. active architectures.

1. Whereas all active speakers are considered powered, not all powered speakers should be called active.

The difference is that a powered speaker has the same signal path as a passive set-up, but the amplification (often both pre and power amp) is hidden in one of the speaker enclosures. There will also be a speaker cable link between the master speaker (that has the amplification) and the slave (passive) partner.

Most affordable powered models tend to use such a configuration. Such an approach has some of the neatness of truly active operation but few of the technical benefits.

....

1. Some active speakers with streaming like the KEF LSX, KEF LS50 II and B&W Formation Duo are fine examples of such all-in-one stereo speaker systems, whereas others like the Dali Rubicon 2 C or the new Beolab 28 come with replaceable connectivity modules (or 'hubs') that handle the source side of things, meaning they can be easily upgraded. After all, unlike speaker technology, streaming technology is anything but timeless.

 

[mel]

One the best recordings I have found to judge music is this antiphonal recording. A separate orchestra is recorded into each of the two channels. The recording is in a church, so the reverberation is an important aspect of the overall acoustic sound. Gabrielli is the composer.

My RME ADI DAC makes the instruments sound like they occupy 1/1,000 or 1/1,000,000 of the acoustic space. I like music that has great space to allow the instruments to fill that vast space with reverberation or musical effects. The organ is an instrument that epitomizes the filling of space, like an echo in a canyon. I am starting to come around to the realization that the active vs. passive sequence of crossover and amplification stages (explained in previous post) has a significant impact on how this type of music is rendered. Moreover, the effects of bi-amplificaiton should be considered. Whether the tweeter is amplified by a class AB or class D amplifier may be significant, due to the speed of tweeters.

I did a simple test to illustrate the difference. I only used the preamp output from an AV receiver with DSP. In the two comparisons, I connected the subwoofer and soundbar differently.

In the first test, I connected the RCA line preamp output to the subwoofer line inputs. I connected the subwoofer output to a small soundbar on my desk.

In the second comparison, I connected the subwoofer from the LFE RCA preamp subwoofer socket. I connected the soundbar directly from the preamp RCA line sockets.

I also tested the "Pure Direct" switch, which minimizes preamp (DSP) processing.

The difference was profound. This experiment gave me clear insight into how DSP and active hardware can be used to advantage.

 

[mel]

A few more pieces of the active versus powered speaker architecture puzzle are contained in the following articles about a related product, the Vanatoo Transparent One.

My takeaway is the internal DAC and DSP processing constraints govern the performance of a true "active" speaker, as opposed to a pure analog "powered" speaker.

My impression is the DAC has the primary influence on the performance of active or powered speakers. The second and third order downstream effects of amplification and crossover are magnified by the DAC output.

I am veering towards the ELAC Navis, because my RME studio DAC performs so well.

• https://audioxpress.com/article/fresh-from-the-bench-vanatoo-transparent-one-encore-powered-speakers
• https://www.stereophile.com/content...ncore-powered-loudspeaker-system-measurements

I hooked my RME studio DAC to my soundbar. The sound of acoustic instruments on Tidal Masters is great, when played at low volume and arm length distance! I don't expect a soundbar to perform so well.

 

[Katji]

Two important distinctions:
Nice diagrams about passive vs. active architectures.

Disappointing. (Probably expectation bias here.)
I imagined something illustrating signal path like...

[ USB ][ XMOS ]--->[ DSP (including frequency division) ]--->[ DAC ]--->[ analog-input amplifier/s ]--->drivers

+ other input sources, analog.

 

[Katji]

I don't set it to line out. I adjust the volume to match the speaker amp, until it sounds ' just right' .

Odd that there are many LSX ( and LS50w) owners who report the same effect.

Maybe we have all taken leave of our senses ..

It's not odd, it's explained by basic psychology. (Explanations readily available on internet.) "Taken leave of our senses" is another story.

 

[mel]

Suppose you want to connect the LS50W in a traditional way like RCA line in (without network connection). Is the intent to connect via HDMI ARC or the 3.5mm jack?

I am getting the impression that the LS50W is not for me. I can compromise on the lack of XLR balanced input.

I just want the active DSP processing to improve audio. Networking is convenient, but not a big feature.

I am getting the impression that the LS50W is some sort of a AV preprocessor with amplified speakers? The active DSP functionality is probably more for AV reasons, than pure audio. Like not blowing out the speakers during sound effects like explosions.

Why not use USB-C, rather than HDMI?

The HDMI ARC supports reads like soundbar functionality. I don't even own a TV. I understand ARC as being incompatible with computer monitors, much less older ones.