Music streaming: Google’s next failure

I am using streaming services for over 10 years. Well, most of us do, considering that YouTube was founded already back in 2005. While many of you just looked at some funny video clips in the beginning, I’ve put on streaming very seriously. It all started with last.fm, a music recommender system called “Audioscrobbler”, which allowed me to explore new music based on my own taste. Music is my passion and I already owned a collection of over 30’000 MP3 songs. But in spite of this great treasure of music, I realized that I was always listening to the same songs up and down… last.fm opened up for me a universe of exciting new stuff! Music I have never heard, of artists I would never choose myself.

But last.fm also had its weaknesses. It was only the algorithm, that decided what to listen, without the possibility to select a specific song. That’s why I switched to other on-demand services later. From Simfy, Spotify, Google Play Music (GPM), Qobuz and Tidal – I tried them all.

Google came to market very late (November 2011) with their own service: Google Play Music (GPM). At that time, competition was far ahead of them. Spotify launched in October 2008, giving it a full 3-year lead over Google. GPM startet as an online music locker only, that allowed uploads of personally owned or licensed music to the cloud for listening on multiple devices. It took until May 2013, until GPM finally became a full unlimited streaming service for music. That means almost 5 years after Spotify! How can a world-leading technology company miss a trend like this?

Despite all the failure, GPM also had some unique features: Upload of your own music and Search. There are some songs or artists that are not available online. Uploading your own songs to your personal library solves this problem. And we do not have to talk about Google’s search capabilities. Unlike Spotify, the search on GPM works like a charm.

You can upload your own music on GPM

Unfortunately, GPM does not yet offer a lossless option to store your own songs. All your music will be converted and stored in lossy MP3 format, what disqualifies the service for the audiophile community. So for the guys for whom special bootlegs and quality have such a high priority, this service cannot be used. What a pity, Google missed another chance – again.

If you lag so much behind the competition in time to market, understanding the music business could help to close the gap. Quality could be an aspect. While Spotify and Google still do not provide lossless music services, other providers (Tidal, Deezer, Qobuz etc.) have advanced into this area.

Unfortunately, Google one more time fails to get the audio fans on board: People who are willing to spend more than $9.99 per month, for good quality music.

For example: Tidal charges about $25 per month and Qobuz has offerings that go up to $400 per year (about $33 per month). Another missed opportunity by not offering a HIGHRES subscription option similar to Qobuz or Tidal.

Qobuz even has a team that takes care of music editorial, which focusses on good quality recordings. A real added value for music lovers. Google also seems to miss this train.

Qobuz editorial services

Instead, history is repeating. Google once again, has randomly thrown a new overlapping product on the market. In May 2018, they launched its new music service called “YouTube Music“. The service should replace Google Play Music on the long run. As usual with Google, the service was once again not available worldwide at the launch date. Of course in all the countries where GPM was already available? Wrong! And of course all albums and playlists are automatically taken over from GPM. Wrong again! Guess what: You start all over again! How cool is that? 🙂 (Irony)

According to a tweet (see article from xdadevelopers) from May 2018 by T. Jay Fowler (Head of YouTube Music), Google is working on it. However, six months after its initial release (in Switzerland, we got finally access to YouTube Music by mid of November), YouTube Music still does not migrate uploaded music and saved playlists from GPM.

YouTube Music: Libraries from GPM are not being migrated

OK, but at least Google knows already your music taste, not? Mehhh! They have no clue. You start all over teaching Google which artists you like (Update May 2020: Two years after launching YouTube Music, the service finally migrates the GPM audio library and personal taste preferences).

YoutTube Music: All of your music taste learned in GPM is lost

But that’s not all of streaming, right? They have at least some great streaming devices, not? What about Chromecast? Good that you ask! Chromecast was introduced in July 2013 ($35) and I loved it. It was such a refreshing user experience, that I switched from Spotify to GPM, just because Spotify was refusing to support Chromecast for the first two years. The first hardware generation was HDMI only and for a Gen1 product, I thought it was a great device.

The first Chromecast device released July 2013

However, some hardware generations of Chromecast devices later, the picture has changed. Why? I don’t understand what is Google’s strategy here. On video devices, Google introduced the third generation of Chromecast in October 2018. A device which supports max. 1080p @ 60fps, two years after releasing the Chromecast Ultra ($70) that supports 4K Ultra HD and High Dynamic Range (HDR10 & Dolby Vision). In 2018, 4K has become the defacto standard and a simple Full HD (1080p) device looks a bit dusty. Also, still using a micro USB port to power the device, instead of USB Type-C surprises me a bit. Why still having two separate video Chromecast devices? What is the advantage of the 3rd Gen device compared to the 2nd Gen device (besides that it is slightly faster)? 3rd Gen should have been 4K at $35. Another opportunity missed. Your competition is not sleeping. Hello Amazon!

But this blog is about music. Let’s talk about the devices specialized for music. Of course you can use the HDMI based Chromecast devices for listening music. However, beside the fact that HDMI is not ideal for listening music (Jitter), the Google HDMI devices do have another limitation: They only support sampling rates up to 16-bit/48kHz.

That’s where the Chromecast Audio comes into the game (Update January 2019: Google has discontinued sales of Chromecast Audio). In place of the second-generation model’s flexible HDMI cable is an integrated 3.5 millimeter audio jack/mini-TOSLINK socket, allowing the Chromecast Audio to be attached to speakers and home audio systems.

Chromecast Audio with optical TOSLINK signal

Using the optical TOSLINK signal, the Chromecast Audio supports sample rates up to 24-bit//96kHz. OK, while this was a good starting point in 2015, the HiFi world of today is all about high quality DAC’s and lossless streaming. The dominant digital interface is not TOSLINK, but USB and I2S. From a resolution perspective, 24-bit/96kHz is nice, but not any longer state of the art. Here I have to mention, that the Google devices use the Google Cast protocol. And there is another problem: The Google Cast protocol does not support important high resolution audio formats such as:

• FLAC >16-bit/96kHz (for example: FLAC 24-bit/192kHz)
• DSD
• MQA

But what do you expect from a company which does obviously not love music? The first generation of Chromecast was a great idea, but since than (2013), Google has failed to deliver what audio people wanted. They created an audio device and a streaming protocol, but forgot to invite the audio fan base to the party. Once again, other and more agile companies have filled the gap. HiFi specialists such as SONORE and SOtM-Audio have created specialized audio renderer – audiophile micro computers with Ethernet input and USB Audio output, that connects to almost every USB DAC. A superb Chromecast device, if you will.

SONORE ultraRendu

As you can see, the market of music streaming is quite complex. There are a variety of devices that serve different purposes. In this article I differ between:

• Streaming Service: On-Demand “all you can eat” service to consume music online and offline. Streaming services do mainly focus on their own music portfolio as a service. Examples: Spotify, Tidal, Qobuz, Deezer, Napster, Google Play Music (GPM), YouTubeMusic etc.
• Streaming Client: User Interface (UI) to enable music streaming. Every streaming service comes with its own UI as well as all network streaming players. As an alternative streaming client, there is also a software only approach that promises more hardware independence. Example for the last category: Roon Labs
• Network Streaming Player: Hardware device that includes a home-grown streaming client and a DAC. In many cases it also includes an amplifier and sometimes even a loudspeaker. Examples: Bluesound, Cocktail Audio, Sonos etc.
• Network Bridge: An enhanced DAC that includes an Ethernet interface and allows music streaming with an existing 3rd party streaming client. Examples: PS-Audio Direct Stream, Mytek Brooklyn Bridge etc.
• DAC (Digital-Analog-Converter): Converts the digital signal to analog. Typical interfaces are USB (type B) or I2S (Inter-IC Sound). The market of DAC’s is huge and ranges from a few bucks to multiple thousand dollars.
• Audio Renderer (often also referred as “Networked Audio Adaptor”): Mini network computers that accept a network data stream from another computer or NAS resulting in less noise ultimately being injected into your DAC. Examples: SONORE microRendu, SOtM sMS-200 Ultra etc.

On the streaming client side, there are the already mentioned lossless streaming services such as Tidal and Qobuz, which are nicely integrated in most high quality network streaming players (such as Cocktail Audio, Bluesound etc.) and there are totally new client approaches such as Roon Labs, which focuses on music lovers, that have content from many sources, often acquired over years of collecting. Roon identifies all music, then enhances it with the latest metadata. It supports local files and music on cloud streaming services (currently limited on Tidal and Qobuz, which makes sense as both services are focusing on HIGHRES streaming). And of course, Roon is supported by all of the main audio manufacturers. This includes the mentioned audio renderer (SONORE/SOtM-Audio) as well as Ethernet connected DAC’s (network bridges), such as PS-Audio Direct Stream, Mytek Brooklyn Bridge and many more. Eventually the network bridges will displace the audio renderer on the long-term, because they combine both functions (DAC & audio renderer) in one system.

Roon interface with its rich metadata

However, as great as those solutions are, most of them yet still have limitations. Some are only working while connected in the home WiFi and/or to the HiFi system (Roon, Bluesound, Cocktail Audio etc.), others don’t support offline listening or do not allow the combination of cloud based streaming services and your own local file collection. The typical network streaming players (Bluesound, Cocktail Audio, Naim, Auralic, Cambridge Audio etc.) originally came with their own user interface, which replaced the one of the integrated streaming services (Spotify, Tidal, Qobuz etc.). That made this type of solutions open for a multitude of streaming services, including the support for local files. However, the user was then bound to a specific hardware manufacturer. Another typical limitation of network streaming players is, that they are built for home usage (requires dedicated hardware) and are usually not well suited for on-the-road usage. Also, being able to support many streaming services under one user interface comes often with a compromise in usability.

That’s the driving force behind Roon Labs. It is a software developing company, with a focus on creating the best user experience. Even though Roon Labs now produces its own hardware components (roon nucleus), the partnership and integration with other audio manufacturers has always been their key strategy (Roon Labs has never competed in the core area of their partners). Like the traditional network streaming players, Roon also replaces the original user interface of the streaming services, but the execution is nearly excellent. Currently, there is no better user experience than Roon. On top of that, you have a broad choice of hardware suppliers. The disadvantage of Roon is its missing offline and on-the-road support. You need a Roon core (server) connected to a home network in order to use the service. The on-the-road feature is on the roadmap, but until today, no date has been committed.

Lately, the streaming services are trying to fight back, by integrating their own clients natively into the network streaming players (Spotify Connect, Qobuz Connect etc.). Why? Because who controls the user interface, makes it harder to be replaced. But then again, we have the problem when playing local files… Most of the streaming services mainly focus on delivering their own content. And even if the service supports local files (Spotify), they then only work locally, which means no access to those files while you are on the road. Aiiii! I told you, not easy!

Finally, you pretty often end up with more than one music streaming solution: One to use for your high-end HiFi system at home and another one for mobile usage while on the road. This also means managing two independent systems to curate your playlists and so on. That’s what I always liked on GPM and Chromecast: It allowed me to use the native user interface of GPM and play my own files as well as music offered by the streaming service. Because my own files have been uploaded to the music library in the cloud, it worked great in my home system as well as on the road. Unfortunately, it only offers low-level consumer quality. That’s good enough while on the road, but not OK when listening to music on my HiFi system at home.

 

Summary

Google had it all in their hands, but they simply don’t get it. Some music industry knowhow would be helpful here. I don’t think that Google will ever be a great manufacturer of hardware devices. It is simply not their core strategy. Instead of releasing another Chromecast device, they should focus on software development and their value added services. Integration is the key! Get native Google Cast support in every hardware device available on the market.

My advice to Google regarding its streaming music offering:

• Merge all music service offerings into one ASAP (GPM & YouTube Music)
• Enable a smooth transistion for existing GPM customers (migrate library etc.)
• Offer a HIGHRES streaming option for music lovers (similar to Tidal and Qobuz)
• Enhance the Google Cast protocol to support high-resolution music formats such as 24-bit/192 kHz FLAC, DSD and MQA
• Work closer with audio manufacturers to integrate Google Cast into their devices
• Enhance the music locker service, to leave the uploaded files in its original format (without converting them to lossy MP3) and support high-resolution FLAC, DSD and MQA files.
• Have a dedicated editorial team that cares about high quality recordings for your streaming music service

Cheers,
Alex