Newshot Title: Bringing old speakers back to life: Home Assistant Connect AUX-2 first-look experience
Body: Over the years of playing with smart homes, lighting, sensors and switches have basically been connected through Matter and Home Assistant, but multi-room audio is still stuck in a stage where each device fights for itself. If you want the living room HomePod, the study Sonos and the traditional powered speaker on the desk to play the same song at the same time, you are often blocked at the door by the ecosystem walls between different brands.
To fill this gap, the Open Home Foundation (OHF) team developed a brand-new open-source audio protocol called Sendspin. As the Sendspin 1.0 standard nears release, OHF's commercial partner Nabu Casa today launched brand-new official hardware to fully support this new protocol: Home Assistant Connect AUX-2.
You can think of it as an audio receiver box built specifically for Home Assistant and Music Assistant. Its use is very pure: connect the traditional speakers or old amplifiers in your home to Home Assistant, and work with Music Assistant to achieve whole-home music synchronization. Beyond basic wireless streaming, Nabu Casa has also packed in PoE power, a combined analog / optical output and a bunch of maker expansion ports, catering both to ordinary users who want to buy it and use it right away and to geeks who like tinkering with ESPHome, leaving them plenty of room.
After using it for a few weeks, I want to talk about what kind of experience this little box can bring to users in China, and what problem the new protocol called Sendspin behind it actually solves.
Why do we still need Sendspin when we already have AirPlay?
Before talking about the hardware, we cannot avoid Sendspin, the core technical foundation of Connect AUX-2. This is an open-source multi-room audio protocol led by OHF (the nonprofit foundation behind Home Assistant and ESPHome), and it has now advanced to the 1.0 RC1 stage.
The reason we are not using the ready-made AirPlay, Google Cast and Sonos and instead "reinventing the wheel" is that existing commercial protocols are too closed. When I previously talked with Paulus Schoutsen, chairman of OHF and founder of Home Assistant, about this project, he mentioned that if you just want to write a program to receive audio streams, make a set of music spectrum lighting effects, or add a custom track-skipping knob, you can hardly move within closed commercial protocols. A year ago, after the core team finished a closed-door Matter meeting in Chicago, they made the decision in a pizza shop to step in themselves, and spent a year writing this royalty-free open-source protocol from scratch.
Unlike simple phone screen casting, Sendspin is more like an audio collaboration system running on the local LAN. By stamping high-precision timestamps into audio packets and calculating clock drift in real time, it lets devices in different corners of the home buffer first and then produce sound precisely on the same beat, and it can automatically negotiate codecs according to device performance. More practically, relying on Music Assistant's bridging function, you do not need to replace all the old speakers in your home. Existing AirPlay speakers, HomePods, Google Cast and even ordinary Bluetooth speakers costing a few dozen yuan can all be pulled directly into the same Sendspin group and play songs in sync with AUX-2.
At the same time, Sendspin splits sound output, audio source input, track-skipping control, cover display and spectrum visualization into independent roles at the underlying level. This means you can let AUX-2 focus on producing sound on the TV cabinet, let the wall-mounted control panel handle track skipping and album cover display, and let the light strip under the sofa read the audio rhythm through LedFx and pulse along with the music. Although it does not yet support home theater surround sound standards such as Dolby, for whole-home music synchronization and sound-light linkage alone, this open-source solution is already quite capable.
Refined and easy on the eyes, but hard offline
Back to Connect AUX-2 itself. The first impression after opening the box is that it is finally no longer the bare board I experienced with similar solutions before, but has truly achieved the refinement of a consumer digital product. The pure white, rounded body perfectly continues the Nabu Casa family design language.
However, its body layout leaves one minor but real complaint. Because the ports are distributed on the left and right sides of the body, no matter whether you place it on a desk or on a speaker rack, the network cable and audio cable have to split out to the left and right, which is somewhat maddening for neat freaks when managing cables day to day. Compared with the previous Home Assistant Green, which can elegantly hide cables behind it, or the Connect ZBT-2, which sits on a desk like an art piece, AUX-2 is best hidden behind the speaker so you do not see it.
Fortunately, the port specifications are generous enough, especially that combined analog and digital optical 3.5mm output. You do not need to go into the backend to change any switch; as long as you plug in a 3.5mm to square-port optical cable, the device will automatically recognize it and switch to digital optical output. For the past few weeks I have been using it with IKEA's new 45 cm SOLSKYDD speaker, and in actual listening the optical connection is also what I recommend most. Not only is the sound foundation cleaner, it also completely avoids the poor contact and common-ground current noise floor common with analog audio cables; if you have a higher-end standalone DAC or traditional amplifier, you can also connect it losslessly through this digital port.
The gigabit Ethernet port on the body also supports PoE power, so one network cable can handle both power and audio transmission, which is very friendly for users who want to install speakers in ceilings or concealed in weak-current boxes. Looking further to the side, the USB port, serial port and the SEEED Studio Grove expansion port familiar to makers in China are all there, showing that the official team never intended from the start to treat it as just a simple audio receiver.
Half-minute effortless network setup, leaving the tinkering to those who need it
If you have a PoE switch at home, the onboarding process of AUX-2 can even be described as "effortless." After plugging in the network cable and powering it on, Home Assistant almost instantly pops up a new device discovery prompt; just keep clicking confirm, and the whole initialization process can be completed in about half a minute. Even if you only use Wi-Fi, opening the Home Assistant companion app on your phone and tapping a few more steps will get it done easily.
One thing that felt very comfortable is that ordinary users who buy it for daily listening never need to touch ESPHome Builder at all. Home Assistant will take over its settings and firmware updates directly in the background, just like managing an ordinary smart plug. Compared with the wild solutions in the community where you previously had to find tutorials yourself, modify code, compile and flash firmware, official hardware is indeed on another level in terms of peace of mind. And when one day your hands itch to tinker, you can take over the device in ESPHome at any time, change the codec, add sensors, or connect a small screen; the underlying permissions are fully open to you.
Listening experience and stability: use Ethernet if you can, avoid Wi-Fi
Once music is actually playing, you can clearly feel the traces that the Sendspin protocol is still in its polishing phase, as well as the performance limits of the ESP32-S3 chip inside. The most intuitive experience is that every time you skip a track or just press play, the buffering wait is a little longer than expected; when forming or disbanding a multi-room speaker group, the response speed is also half a beat slower than Apple's own AirPlay. But there is nothing to complain about in sound quality; it reaches the level expected of a front end at the same price.
But as long as buffering ends and the music officially starts, its synchronization performance is very stable. Under the stress test of multiple devices in my home playing at the same time, the sound in each room aligned perfectly, and after several hours of continuous playback there was no channel drift or out-of-sync phenomenon at all.
Here is a pitfall-avoidance tip: if you can plug in a network cable, try not to use Wi-Fi. The 2.4GHz wireless environment in homes in China is very crowded, and streaming lossless FLAC music directly over Wi-Fi may cause stuttering, while the aging wireless module of the ESP32-S3 itself amplifies this shortcoming. I have to say it is indeed a bit regrettable that it did not catch the mass shipping and adaptation of Espressif's new-generation chips.
If there is really no way to run a network cable where you place the speaker, I strongly recommend going into the ESPHome Builder configuration and enabling the lower-bandwidth Opus codec; smoothness in a wireless environment will improve immediately. You only need to add the following lines to the configuration file for it to take effect (note that the Opus patented codec is for personal testing and non-commercial use only):
media_source:
- id: !remove sendspin_source
- platform: sendspin
id: sendspin_source
task_stack_in_psram: true
static_delay_adjustable: true
codecs: [OPUS, FLAC, PCM]
Hidden tricks on the speaker rack: hang a lyrics screen, or use it as a signal springboard
If you only use it for listening to music, the optical port and PoE mentioned earlier are already worth the price; if you have strong hands-on skills, the expansion ports on the side are where this machine is most fun.
Take that USB port. Besides serial control, it can also be used directly for remote mounting of a Zigbee or Z-Wave coordinator. Many people's Home Assistant hosts are stuffed in weak-current boxes or deep inside cabinets, and the wireless signal is poor after passing through walls. In this case, plugging the coordinator into the AUX-2 on the living room TV cabinet can naturally move the signal center to the middle of the room. At present, ZHA's remote mounting function is still experimental, and I am quite looking forward to the official team later fully connecting this玩法 with its own Connect ZBT series hardware, or even using it directly as a Thread border router.
The Grove port next to it directly connects to SEEED Studio's huge modular ecosystem. I had AI generate the driver configuration according to the spec, plugged in a Grove RGB LCD character display, and put on a shell I printed with my Bambu Lab printer, so it can scroll in real time beside the speaker showing the current song title, album and progress bar.
The moment I saw the character display light up, the retro feel came out at once (I have uploaded this configuration code and the 3D printing model to GitHub and MakerWorld; if you want to replicate it, remember to switch to the ESPHome Builder Beta channel to compile the firmware). As for the serial port next to it, the official team has also built in control presets for devices such as LG and Samsung TVs, so it is also convenient to pull the old living room TV into automations.
Is it worth buying? A talk about purchase channels in China and alternatives
Home Assistant Connect AUX-2 is priced at 520 yuan (early-bird price 450 yuan). Although there are small teams like Apollo Automation in overseas communities making similar third-party ESP32 audio boxes, first, they simply cannot be bought in China, and second, once the high transshipment shipping costs are included, there is no cost-effectiveness at all. The good news is that AUX-2 will next be listed directly in SEEED Studio's official stores on major e-commerce platforms in China, so you can buy it directly at the original price without the hassle of overseas shopping.
As for how to choose, it depends entirely on your expectations for sound quality and tinkering. If you have good traditional speakers or amplifiers and want the peace of mind of optical output, PoE power and native firmware maintenance, then just wait for SEEED Studio to list Connect AUX-2.
If you happen to also want to add a local voice assistant in the room and are not fixated on optical output, the Home Assistant Voice Preview Edition (overseas price $59), which is also available in China, is actually more cost-effective. It comes with a 3.5mm audio port and natively supports Sendspin, so it can solve voice control along the way.
And if your budget is very tight and you are familiar with flashing firmware yourself, buy a SEEED Studio reSpeaker Lite development board for just over 200 yuan and flash ESPHome yourself; you can likewise run a stable Sendspin player at low cost. A search on MakerWorld will turn up plenty of matching shells; pick a good-looking one and put it together, and it is not bad at all.
Connect AUX-2 can be regarded as Nabu Casa's most ambitious hardware product to date. It neatly packs the previously high-threshold multi-room audio synchronization technology, together with a series of Home Assistant hardware connection expansion capabilities, into a small box that an ordinary person can set up and use in half a minute.
As maker kits developed for Sendspin come to market one after another, the Sendspin 1.0 standard will also be officially released by the end of the year. I look forward to the strong audio hardware industry in China being able to support this open-source newcomer and unify the fragmented ecosystem of the audio market.
Image source: 南桥北桥不是桥; this article is authorized for exclusive Chinese publication by SSPAI, and reproduction is prohibited without permission.