There’s a common misconception that music streamers are simply about the qualities of their digital chipsets alone. That is, the computer-like streaming module that pulls digital music data from the cloud (from Tidal’s server, say) and packages it into a smooth constant data stream, and the digital-to-analogue converter (DAC) that then turns that data stream into an analogue waveform the rest of your hi-fi system can understand.
It somewhat stems from the broader fallacy that digital audio’s inherent ‘ones and zeros’ makeup means that there can be limited differentiation in their delivery.
But as eloquently illustrated by Cambridge Audio, such chips are comparable to how an orchestra performs: “If the conductor isn’t very good, the orchestra won’t perform as well as they can. However, if the conductor is surrounded by poor musicians or the wrong instruments, again the performance won’t be as good as it could be.”
Indeed, while chipsets do matter, and it is true that all streamers perform one stage of the process with parity (which we’ll get to), their audio performance is predominantly down to how the chips are implemented.
As Paul Neville from Harman International told us, “Engineering a high-quality music streamer requires a holistic approach to the signal path, from the network input to the final analogue output.”
So what exactly matters in music streamer design? Here’s what experts over at Audiolab, Cambridge Audio, Chord Electronics, Cyrus, dCS, Fell Audio, WiiM, Lenbrook (NAD, Bluesound) and Harman International (Arcam, Mark Levinson) had to say about what makes streamers truly sing…
Chips do matter
The streaming module is a crucial music streamer component. It communicates with the source to get your music before packaging it up in a digital language the DAC chip can understand. Up until this point, the digital audio data is more or less handled as it would be had it come from a CD transport.
As dCS’s James Cook tells us, this first process performed by the streaming module guarantees bit-perfect data from a lossless source.
“This means two comparable streamers that support the same version of a streaming service will output an identical data stream,” he says. “This holds true largely regardless of network conditions, the equipment upstream of the network streamer, and to a degree the streamer itself.”
That’s the parity – and to some extent why some people believe that music streamers (those that are DAC-less, anyway) should all sound the same.
What about your home network conditions; does that affect the obtaining of networked music?
“While network stability is important, the reality is that most modern connections, whether wi-fi or wired, offer ample bandwidth for even the highest-resolution audio,” says Cambridge Audio’s engineering team.
“Streaming 4K video with multi-channel audio is a far greater data demand than high-resolution audio. If your network can handle that, it can almost certainly handle your music.”
How the data stream is rendered before the digital-to-analogue conversion stage, however, does make a difference. “The way a streamer unpacks and processes a digital file has already shaped the final sound,” says Fell Audio’s Luke Tyson.
“Its handling and transmission of ones and zeroes within the [circuit] board affect timing, accuracy and ultimately musicality; garbage in, garbage out!”
Tyson says that in this world of increasingly affordable music streamers, “it’s easy to assume that a great music streamer is simply defined by the DAC chipset or streaming module it uses”.
The quality of the streaming and DAC chips themselves does matter – specifically in the case of the latter.
“The first step is selecting the right chip,” says Harman International’s Neville, while Cambridge Audio and Audiolab also stress the importance of using a good-quality DAC chip. But achieving good performance goes far beyond picking expensive chipsets.
“Often you see ‘insert latest DAC name here’ as a marketing bullet point as if just having it in that product guarantees a good sound,” says Cyrus’s Nick Clarke.
But as Audiolab and Cambridge Audio reiterate, the truth is that just as crucial is how it is implemented within and complemented by the surrounding components and circuitry. And often, Clarke continues, “As the performance and functionality of these devices [chips] increase, the scope for not handling them correctly also increases.”
Negating the noise
Several design factors need careful consideration in order to prevent a plague affecting all hi-fi: noise. “Even the best DACs and processors won’t perform optimally if the power supply is noisy or the circuit design introduces unwanted electrical interference,” says WiiM’s Dr. Lifeng Zhao.
Let’s take the power supply. Ideally you want separate supplies feeding the digital (DAC) and analogue (post-DAC) stages with, as Audiolab’s Jan Ertner describes, “clean, consistent power”.
Ertner says that the DAC’s supply in Audiolab’s flagship 9000N (our current Product of the Year streamer) having multiple discrete ultra-low-noise regulators providing power separately to the left and right channels for each conversion stage is a “vital contributor” to its performance.
As an aside, Neville at Harman believes that linear power supplies are preferable to the smaller and arguably more efficient switch-mode supplies here, as they produce less noise and provide “cleaner and more stable power for optimal sound quality”.
As he rightly points out, you need these separate analogue and digital supplies to be isolated from one another, too. And ‘away’ from sensitive circuitry, full stop.
But you can go one better than that. Chord Electronics’ Matt Bartlett says that using a streamer with an external DAC is “one of the best ways to isolate the audio signals, as this allows you to keep all the analogue audio signals you actually listen to, away from any of the noisier streaming functions”.
As we mentioned earlier, the majority of music streamers (including those in our best music streamers buying guide) do, however, have DACs built-in.
Those without one do the pre-DAC music pulling and processing only, are often aspirational designs and are referred to as ‘network transports’ or ‘network bridges’; the recently reviewed Lumin U2 Mini is an example of such.
You then have noise creeping into the audio path or supply from the streamer processing or conversion. Bartlett likens streamers to computers (they are “computers optimised to run dedicated audio/streaming software”), which are inherently noisy.
“They have a lot of clocks and signals that, if not handled correctly, can induce noise,” he says. “This is exacerbated by the need to make the software run as fast as possible to give a positive user experience; users will not appreciate lags and delays in playback if the software cannot keep up. To do this effectively, larger processing and memory is required which, more often than not, also requires faster clocks and more demand on the power supply, and so more noise.”
Cyrus’s Clarke explains that the way streamers receive and decode data – by receiving it in blocks (sometimes referred to as ‘packets’) and placing it in a buffer for decoding in blocks – can also produce noise.
“Specifically when dealing with compressed audio, a poor [decoding] implementation will have measurable artefacts at the block-processing frequency,” he says.
In summary, “A low-noise environment is crucial to unlocking the full potential of high-resolution audio,” says Cambridge Audio.
A question of timing, among other aspects
One of the most important aspects of digital audio delivery is trying to prevent (or minimise) ‘jitter’.
In the case of music streamers, this is the distortion artefact that occurs from slight timing deviations in the clocking signal during the digital-to-analogue conversion stage. These small errors mean the reproduced analogue waveform isn’t exactly the same as the original.
Unsurprisingly then, every manufacturer we spoke to here mentioned the importance of clock design in a music streamer.
The most important one is the ‘master’ that regulates the conversion stage; however Cyrus’s Clarke also explains that there is scope for large amounts of jitter in the sending of data to the DAC (as there is with any digital source pre-DAC).
He says that while modern DACs are then “very good at handling this ‘automatically’, it would be much better if the data was correctly timed in the first place”.
This somewhat goes back to Clarke’s earlier comments that what happens before the DAC stage has an impact, as well as Tyson’s claim that “the way the renderer buffers and reconstructs the digital audio signal before passing it to the DAC can make or break the overall performance”.
For this reason, dCS’s Cook says that for high-performance streamers it is vital that the streaming solution can accept an external clock signal as this helps to ensure that the DAC-bound audio signal leaving the streamer has a low level of jitter.
“A key feature of this is having high-quality clock circuitry inside the product running at integer multiples of the two base audio sample rates (44.1kHz and 48kHz) which can be fed into the streamer,” he says.
Clocks and timing aside, Greg Stidsen of Lenbrook says that the circuit board layout is important too: “How the traces are laid, where the components are placed – all of this affects the sounds to varying degrees.”
Stidsen keenly reminds us that these kind of changes almost exclusively evolved the Bluesound Node 2 (2015) to the Node 2i (2018) on the engineering side of things, and that we noted in our review just how much the performance had improved in the newer model as a result.
Harman International’s Neville agrees that “adherence to high-speed circuit board design principles” is crucial for signal integrity.
Neville and Clarke (Cyrus) discuss how in such hybrid digital and analogue products, careful management of signal grounding (the reference point at which voltages are measured) is also important, with particular focus on where the analogue and digital ground domains intersect to minimise noise and interference.
Of course, Neville adds, on the whole, carefully tuned analogue stages are essential too!
Software and sound
Where digital is the topic of conversation, software isn't far behind.
When you think of music streamer software, you likely think of the app – the thing that can turn your phone or tablet into a practical and intuitive remote control for the platform your streamer runs.
And indeed it’s that operation, alongside the behind-the-scenes design, that impacts not only a product’s user experience but also its audio performance.
Stidsen of Lenbrook – who knows a thing or two about developing streaming systems, with the company’s Bluesound brand having developed its own system (BluOS) based on a Linux kernel – goes as far as saying that “audio quality is defined just as much by software as it is by hardware in a streamer”.
“A well-designed streaming app directly impacts how music sounds, from how the signal is routed to how it interacts with the hardware,” says Audiolab’s Jamie O’Callaghan, tying in with an earlier point that the way in which a streaming platform processes digital data is primarily down to software.
“The various codec implementations and audio chain are all areas where the sound quality can be greatly diminished if the software engineer does not have audio expertise,” adds Stidsen.
WiiM’s Dr. Lifeng Zhao also shares the opinion that hardware alone doesn’t guarantee a great streamer, and it is unfortunate but true that our reviews team has come across great-sounding streamers that we haven’t been able to wholly recommend due to their sub-par user experience.
Nothing is worse than buggy app control prone to drop-outs or lag. Actually, there is one thing: music playback glitches.
“A platform must have a proven track record of stability, ongoing updates and innovations,” says O’Callaghan at Audiolab, while his colleague (and others) stress the importance of platform relevance and longevity, considering the never-ending road of streaming service updates and protocols that lay ahead of music streamer development.
The big picture
Going back to our earlier point, despite those who, from the dawn of digital audio, have insisted that all digital audio kit must sound the same, Audiolab’s Ertner says that “this is demonstrably not the case, and that includes music streamers – even those that use the same streaming platform or chips”.
As several streamer manufacturers have examined here, everything from power supply and analogue stage design to software processing and platform handling impacts streamer sound performance – not just chipset selection.
To that end, Fell Audio’s Tyson proffers a fitting denouement: “A more entry-level ESS Sabre DAC chip, when given a carefully considered discrete power supply feed, better component layout, and efforts to keep signal integrity, can yield superior results and will most likely provide a more complete musical performance than simply opting for the more expensive DAC model.”
So the next time you’re shopping for and comparing music streamer specifications, it would be wise to remember the bigger picture.
MORE:
Best music streamers: top network audio players tested by our experts
Hi-res music streaming services compared: is Tidal, Qobuz, Amazon or Apple Music better?
Hi-fi and music streaming live harmoniously, but two Big Tech brands hold the relationship back
Vinyl records and music streaming both hit landmark highs – but the one growing faster in revenue may surprise you
Qobuz reveals average payout per stream; claims it is higher than rivals
Industry sources:
- Audiolab – Jamie O’Callaghan, Global Marketing Director; Jan Ertner, Head of Electronic Design
- Cambridge Audio – engineering team combined
- Chord Electronics – Matthew Bartlett, MD
- Cyrus – Nick Clarke, MD
- dCS – James Cook, Product Marketing Manager
- Fell Audio – Luke Tyson, Design Lead
- Harman International – Paul Neville, Senior Director of Global Audio, Luxury Audio
- Lenbrook Group – Greg Stidsen, CTO
- WiiM – Dr. Lifeng Zhao, CEO
You must confirm your public display name before commenting
Please logout and then login again, you will then be prompted to enter your display name.
