Espruino JavaScript
Bangle.js 3: GPS? #7834
Replies: 27 comments 66 replies
|
I own both Bangle.js 1 and 2, and to be honest, I've never used the GPS. I've always preferred to use a smartphone for anything related to geolocation. If Bangle.js 3 comes out without GPS, that won't stop me from buying it. |
|
I would remove the (optical) heart rate sensor too. For same reason as GPS - it is not good enough and may bring the cost and size down. However OTOH if the watch will be expensive anyway then the HW may not add that much to it, so each feature included may get more buyers. Or maybe other sensors could be more useful instead - ECG, body temperature. For running if one cannot take phone then maybe some other device could add GPS and/or HR sensor over BLE (ring, chest strap). If you would add GPS to something like Puck.js that could solve the GPS issue too. |
|
Interesting - lots of thumbs ups. Do a lot of people feel the same way about the HR sensor? Honestly it has been a bit of a pain. I like having it and the ability to see my heart rate, but we routinely get people grumbling about the accuracy in different scenarios and over the years I've spent a lot of time on it. |
|
My thumb is more about @fanoush delibering on some different tradeoffs, not in support for or against hrm exclusion. But also the idea to integrate more with eg smart rings/separate small computing device (puck.js advanced/ pebble core concept). Like is already possible with hrm chest straps. Taking the hrm chest strap as an example though, I like it as a complement to the included one and not a replacement. For now at least. |
Those cheap (~US$15) aliexpress Colmi rings should be supported in GadgetBridge since this https://codeberg.org/Freeyourgadget/Gadgetbridge/pulls/3896 Too bad I did not try it as I am not thar much interested in HRM but I have some of those rings so this reminds me I should try it when running. I hope having it tightly fit on finger could help with accuracy. They report HRM rate over BLE so either one could connect it to Bangle or even use it directly from phone in GB instead of Bangle HRM. I take HRM as something that is there because we repurposed cheap fitness smartwatches that have it as marketing bullet point, not that is is that much useful and actually working and being accurate. When running the wrist is probably the worst place to have the sensor. But if people want it and it is cheap and there is nothing better to put there instead then so be it :-) |
I think heart rate sensor is good because it's makes it able to detect if the watch is on better. but yeah the accuracy... |
|
Honestly ive never used the watch's gps as my use for the watch has been mostly as a watch with notifications, sleep tracking and step counting. If the swapable back idea goes through, then it might just be worth making a "fitness" back with both an hrm and a gps module. |
|
I'm glad you suggested this, I was thinking it would be a good idea too! I'm aware it adds a lot of technical complexity, but as someone who doesn't use the GPS or the HRM I could happily live without them. My personal use case for any smart watch is as a watch first, then for notifications, then for helpful apps (timer, counters, step counter, etc.). |
|
Thanks - I like the idea of a 'fitness back', but given the quantities of Bangles I ship and the R&D costs associated with it, I can't imagine it's ever going to make enough sense that I would personally develop it. |
|
That brings an interesting point though. Is it possible to have a customizable option, so when ordering the Bangle.js 3 you can either select the model with GPS or WiFi (as an example), or the base one, with HRM? Would it be worth it? |
|
I think I mentioned elsewhere, but GPS is potentially an issue because for it to work well enough you have to put the aerial somewhere. But you could in theory have a Bangle.js add-in back with WiFi (maybe LoRa) that made the Bangle a bit thicker by adding a second PCB. I'm not sure if that'd be something it was worth me doing personally (it'd depend on demand), but for instance I might supply the CAD files for 3D printing a bigger back, as well as the PCB outlines so anyone could easily make their own versions. |
|
I admittedly don't use GPS on my watch that much nowadays. I can't help feel like it still makes sense though. Having it leaves the door open to more types of apps. Especially if the rest of the hardware is more capable in terms of computing and screen real estate. Personally I'll most likely be buying the Bangle.js 3 regardless of how it shapes up. But I would not have bought my original Bangle.js 2 in the first place if it didn't have GPS. For me the more important tradeoff is probably size vs function, and not money vs function. I prefer a not too bulky smartwatch, but if it would be as thin as or thinner than Bangle.js 2 either way I'd probably say include GPS. As @fanoush says maybe it could be moved to a separate companion puck.js device a la the unreleased Pebble core, https://youtu.be/Pwq89K6RBTI?si=nJqhEjR_lNmBDVWQ&t=171 . I think the new pebbles won't have GPS so it could be a differentiator there as well. Maybe you could offer PebbleOS versions with a GPS app included, making it the first pebbleOS watch with GPS. My thoughts as of now. |
Thanks for the honestly there! I think that's my concern - that maybe it's not used much, but it is something that makes people interested in it. Having a small, programmable, wrist-mounted device with GPS does feel like something that's not really offered elsewhere. Even if I don't actually use it much I am excited about what I could do with it. Without it and HRM, it's just a touchscreen display on your wrist. |
|
I agree with size; most modern watches are bulky, and not at all something I would wear daily. The bangle.js 2 is a perfect size, so keeping it light and small would be quite valuable for me. |
|
I use the GPS and I think it is one of the features that separates the Bangle from its alternatives like the pinetime. |
|
btw pine time pro will include gps |
|
Just to put this out there (I don't know if it's feasible or not yet). What if there wasn't GPS but there was WiFi instead? It seems a lot of users would value WiFi, and while I don't think an ESP32 makes sense as the main processor due to its power usage, we could use one for WiFi or even potentially share the IO pins for the display so that apps could have the choice of running on the ESP32 or the nRF52840. |
|
If you stay with Nordic there is companion nRF7002 wifi chip from nordic that can do 5ghz and is low power. It also comes bundled on some modules like this one https://www.aliexpress.com/item/1005005542789484.html However some recent esp32 can do 5ghz too(?) and some are finally better at low power https://github.com/espressif/esp-idf/tree/master/examples/bluetooth/nimble/power_save#typical-current-consumption-with-management-enabled (with external 32KHz XTAL added) |
|
My concern with nRF7002 is it'll require Nordic's new SDK I think, so a complete Bluetooth rewrite for me. It's a pretty huge change. Recent ESP32 should be fine for 5ghz and it opens up quite a few new possibilities. There are a ton of toolchains for ESP32 (IDF, Arduino, MicroPython, etc) so it's possible someone could write a watch app in Python for instance and have basically complete control of the ESP32 for their app - which avoids having to worry about abstraction and interaction with Bangle.js itself.
|
That is advantage but also adds complexity and fragmentation. But I am all for it :-)
well sooner or later this should probably happen so that would be good opportunity. I think new port of bluetooth.c from scratch could give a shorter cleaner code and maybe it wouldn't be so much work compared to the rest. This would also allow you to go with new nrf54 family which has lower power consumption. also newer BLE features like LE audio could be useful |
|
I don't really want a wifi on my watch. It consumes a lot of power, only works when wifi is available, and hard to enter passcode for a new wifi, etc. |
|
WiFi isn't really valuable to me, seems like i'd always keep it off. I can get all data that needs wifi from my phone, with minimal software modification. |
|
A comment spotted re no GPS on the new Pebbles on the Verge's article on the new watch design: 
|
|
:) Interesting - thanks! Although it looks like nobody else actually voted that comment up as it's right at the bottom? I totally see Pebble's point of view - they never had GPS (afaik?) so there'd have to be a big engineering effort to add that into Pebble OS - plus the cost increase in each device. It looks like Eric's taking a very sensible view - just putting in what people loved about Pebble, and not much else. Most people wanting GPS are mainly wanting it for running? But then they're also the people that are going to be disappointed that the HRM isn't as good as the one on a Garmin/Apple watch costing 4x as much, and half of them will be running in a city with high buildings around where the GPS struggles anyway. Maybe it's better to look at Pebble, and to stick with what we're good at. |
Not sure about that. GPS can be also used for hiking (or geocaching) where you don't care about HRM at all. also you may use something else for HRM and still appreciate GPS even when running. |
|
... I do use GPS for hiking myself, but if I'm honest I've always got a bag with supplies (and my phone) in, and even now it's possible to configure Gadgetbridge so that it can use the phone's GPS instead of the watch GPS, so it wouldn't really affect my experience. |
|
I like the idea of offloading as much as possible, whenever possible, to the phone, to keep the watch as slim, light, inexpensive, and running as long per charge as possible. |
|
I understood that, if you want good GPS, it needs a lot of money and good hardware. I have bangle 2 and I tried running software, but it never worked as well as my sport watch. So every time I do sports I change to sports watch. So really, no need for gps, because I think you can't get it good without huge effort. |
|
Thanks - yes, I think it's safe to say that whatever I do for HRM and GPS it's not going to be as good as a £400 watch by a company with a huge R&D budget. It's a hard one because I think there are a lot of quiet people on the forum that do use and value GPS and HRM, but at the same most of what I get as feedback is complaints when they don't work perfectly. ... but if I took the approach of not shipping things people complained about, I'd never ship anything :) |
|
Yeah, that's true. You would make bias conclusion, if you only read complaints. On the other hand, you can get some idea, what are most wanted sensors with what are most used/downloaded apps. Also, if something gets a lot of complaints, it is also widely used. |
|
To me the value of Bangle JS is its open source hackability. The fact that I can write my own app or watch face and share it with a larger community is the killer app. In this it is somewhat like LEGO: the value is in what you can make with it . That's the domain in which I think Bangle can compete in and win. |
|
It seems many people don't use GPS on Bangle.js, but I use it a lot for golfing. I really hope the next version to have the GPS. I don't really care about HRM though. |
|
Thanks! And yes, I totally agree about the buttons and ability to make sound - both something I'll be trying to include. |
|
I use GPS a lot but for the BangleJS2 I carry my phone as a backup on races as I can't rely on the Bangle getting a signal. I'd love to not have to carry my phone on runs or rides! I use the GPS daily too, it's the main feature after BLE I make use of. HRM I would use, but I found that the one on the Bangle wasn't accurate, or I had it setup wrong, so I don't use it at the moment. |
|
For me, the lack of GPS wouldn't be a problem. I used it a little at the beginning, but mainly due to battery issues, I always used my smartphone when I needed it and I don't miss it. |
|
+1 for waterproofing |
|
Don't worry, waterproofing is essential for me too! |
|
While I could never get the GPS working correctly on my first BangleJS, I've found the GPS on the replacement I bought to be a lot more reliable and have been using it a lot lately mounted to my bike handlebars with Gipy running to follow gpx routes. Especially if the BangleJS 3 is going to have more colours on screen which could display more detailed maps I'd love if it could still have GPS, but understand also that for example my phone can supply coords as a substitute |
|
I would definitly want gps. The more accurate, the better :) |
|
My Bangle2 just died so the reason I saw this was looking for any updates on Bangle3. I wouldn't have bought a Bangle2 without GPS and HRM, and I wouldn't buy a Bangle3 without them. I bought the watch for running in the first place. The main thing I would wish with the GPS is faster time to lock - even with aGPS I sometimes give up. I don't take my phone with me when I run. I hope having GPS wouldn't make the watch too much bigger than the Bangle2 - it's pretty much the perfect size at the moment. The only features I use more than GPS on Bangle2 are telling the time and the vibration alarm. I have no interest in WiFi in a watch. I'd keep it turned off. If we are comparing things to the new pebbles, I also have no interest in a microphone. In fact, I think it's an antifeature. There are enough microphones around me all day. |
|
Exactly my view. A faster fix would make GPS more useful. I give up on it too if it doesn't fix within half a minute |
|
The GPS lock can be quite fast now if you've uploaded AGPS in the last 12 hours or so, but the AGPS data on Bangle.js 2 only lasts 24 hours. A different radio could improve on that - the ublox ones were good for 7 days |
|
GPS can be flaky, on a watch more so. Not interested. Also not interested in Pebble. Bad management killed it in 2016, and once is enough for me. |
|
Honestly, I haven't used the GPS at all since I got my watch (3 months ago). I just don't see the need of it if it's paired to my phone, and can get the location from there. In my mind, I prefer a long battery life over GPS.... That being said, if we do add the GPS, we might also be able to squeeze a bigger battery in there as well. Even though the watch advertises up to a month battery life, realistically, it's just not possible, and if you do manage it, you can't use it like it's the smartwatch you expect. Currently I charge my Bangle every 5 days or so. It would be a quality-of-life upgrade, that also would contribute to longer life with GPS on. Nice that you're designing the 3 from scratch! It's pretty cool when it's not just open-source software, but right now open-source hardware, in a sense, where the whole community can come together and discuss about what all should be in the next generation of the Bangle, creating something amazing in the process! |
|
Thanks! I didn't think I'd really be in a position to do something entirely custom, but it looks possible. It's still a pretty huge expense, and I'd have to commit to an initial order of 10,000 watches which is obviously a lot of money. Realistically what's going to happen I think is I'll pay for the R&D out of my own pocket, but to actually get the 10,000pcs order done I'll have to do preorders again, so if not enough people want one then it may not happen. I think realistically we will have an HRM - because all the smartwatches have them, you can just get an HRM on a flex PCB, with the correct rear glass, for not that much money. It seems GPS is another matter though because fitting the aerial in the case and tuning it correctly can be very tricky, especially in a round watch. I'm still on the fence about it - it really shines if you're a runner, but if you have your phone with you everywhere (and it's Android, not iPhone) then as you say GPS seems a bit pointless. I am hopeful we'll be able to get a bigger battery as at minimum there is newer technology and the round case should give us some extra volume for it, but also with the 'bigger back' option that's suggested it should be possible to put a massively bigger battery in there if needed. |
|
Alright, thanks! Having a round watch would be nice - are you planning on the screen being round as well? Also, just an fyi, technically you can setup an iPhone to push the location to the bangle using a location shortcut as well. |
|
Oh yes - I posted up in another thread but it's a round screen like you'd find one a Garmin watch. I wouldn't try a round watch with a square screen :) I think I probably would go the KickStarter route - and when I do that I have to push quite hard with advertising (getting reviews/etc) - so it's definitely something I'd do. If I could find a company that'd help with that without being extortionate or spamming everyone I'd be very interested though :) |
|
so its going to be round or square? |
|
almost certainly round |
|
Hi, Thanks for your work on the bangle! I have made a kid happy with a pokedex and pokemon watch on this device |
|
Thanks! Bangle.js can connect to other Bluetooth LE devices right now, but not classic Bluetooth without having some drastic hardware changes. I'm afraid Bluetooth Audio is probably not going to happen any time soon :( More RAM could be an issue as it's in the chop but I think sticking more flash memory on should be easy and is something I really want to do. |
|
I've had a BangleJS2 since the original Kickstarter (and am now on my second!). I have used the GPS once or twice in all that time, and really only out of curiosity and once to set location for sunrise/sunset times when on holiday. So GPS not a priority for me, nor heart rate (also used once, out of curiosity). Not have I ever wished it had WiFi. |
|
My idea for why I'd love a microphone: I made a routine/task app for my Bangle, and I keep wishing there were a way to just tap my watch and record a quick voice note to add a task. Right now it means pulling out my phone, opening the app, recording, and saving, which is a lot of friction. The idea would be: tap the watch, talk, and the app automatically sends the audio to the phone, transcribes it, and adds it as a task or note. That happens multiple times a day, and those are exactly the things that tend to get forgotten. I wonder how practical is a mic in a watch like bangle. |
|
There are waterproof mics available - I think it would be pretty neat. However, I use iOS, and it's pretty painful to set a transcription service up for the bangle, so it wouldn't get any use from me unless somebody or myself develops an app for the bangle in the future... |
|
It's definitely something I'm looking at - I really don't want to risk the waterproofing, but if I can do it safely it'd make a lot of sense. Not just for storing/transcribing voice notes, but it should be possible to do a minimal voice recognition that'll tell between a few keywords on-watch so you could for example use it to load apps or trigger events (eg home assistant) |
|
Well, have you ever thought of making maybe a pro version and a base version? |
|
Yes, there were some discussions on the "I demand a Bangle.js 3" thread - the problem really is I don't sell enough devices to make it worth me having two different designs. It's twice the testing, certification, etc costs (not to mention keeping stock) so it'd just make both watches more expensive. I'd likely have a replacable back with extra things users could add to the watch (LoRa, WiFi, etc) but I expect those to be pretty low-volume and maybe something provided by the community rather than me |
|
I’m really happy to be able to run sometimes without bringing my phone. The watch is small and lightweight. And with apps like Run+ and Recorder, I can directly get a GPX file with my position and heart rate. I’d use it even more often if the battery didn’t go down so quickly! |
|
I'd love to see the Bangle.js 3 pivot toward high-end hardware. While its openness and hackability are unrivaled in the watch market, the current budget specs make it hard to compete with mainstream watches. If we aim for the best possible hardware that remains accessible (without being 'crazy expensive'), we’d finally have a premium foundation to build apps that don't just work, but actually rival other watches out there. That said, I would include gps also as many applications knowing the location without the need of a phone is a must. |
|
I would clearly prefer HRM to GPS. In my view GPS is easily replaceable by the phone's GPS, while HRM is a distinct stand out feature of a wearable device. I come from a health background, and currently I am exploring possibilities of health tracking using open source devices. This is not so much about running and fitness tracking, but rather collecting overall basic vitals over time like pulse rate and steps for health tracking. Currently there are very few open source devices suitable for that. I would love to see a device with an improved HRM unit that could perhaps also do SpO2 readings, similarly to the upcoming PineTime Pro. That would also be a big improvement for the fitness geeks here, to get more precise HR readings. There is a lot of research on using wearables in the health sector and it will be a big market. I very much dislike seeing public health sector shelling money out to the big IT companies for locked down devices with very little to no control over the data. And I admire the Bangle ecosystem which allows for really easy App development, which would enable easy deployment of custom applications and configurations for e.g. research studies. |
|
Thanks! Yes, the plan is we're going to put a much better HRM on Bangle.js 3, hopefully with SpO2 - and we're adding more memory to help with logging, and Wifi which can help with letting the watch upload that data. We actually already sell a surprising amount of Bangle.js 2 into adademia for research, so it's something that's very much on my mind! |
|
So... I guess I'm a bit late to the party. The GPS is most fun feature of Bangle.js2... and also most problematic one. Reception is poor and it eats battery way too quickly. I guess it depends on what the tradeoff would be. I like how small bangle.js2 is. Wifi... again, it depends on what the tradeoff would be. Wifi is somehow nice, but if it would requite bigger battery, or other compromises... I like small and light. But if I can add a wish to the list... some kind of sound-producing device / beeper would be nice. |
|
That's been my concern really - I could put GPS on, but it's never going to be up to the quality of Garmin, who can throw millions at R&D and have custom aerials included in the bezel. So if users get the Bangle.js 3 and are then unhappy about the GPS it almost feels like it's not worth having it. Wifi shouldn't drain the battery until it's used - so it's up to the apps how it's used, but if it for instance syncs new stories at midnight when the watch is on charge, it's a huge bonus and doesn't affect battery life at all. It's also including a proper speaker. It takes up a lot of space inside but I think it's a very valuable addition. |
|
WIFI: my fear with tiny battery is that if battery is 20% and app uses wifi (and draws 0.6A), battery will die immediately. This is already problem with GPS on b.js2 (and I believe GPS is not nearly as power hungry as WIFI). GPS: one of advantages is that you can use it to get precise time. You don't even need full fix for that. speaker: even tiny piezo beeper would be useful for alarms, etc. Even old watches had that for beeps. But I'll take speaker, too :-). |
|
No GPS wouldn't be a showstopper for me; I rarely use it and I usually carry other devices (tablet, smartphone, bike computer, outdoor navigator..) that can record location. However it is a nice-to-have feature. But leaving out the HRM monitor, as others have suggested, would be an absolute no-go for me and I would definitely not buy it then. In fact the HRM is the main reason I wear a smartwatch anyway, otherwise I'd just use a really cheap wristwatch for checking the time and the smartphone for all other functions. I find the HRM on the Bangle.js 2 actually not too bad; I don't need it to be really precise but I want to wear it all day with regular recordings to have a general idea of my activity levels. For sports, I use a chest strap HRM with the smartphone, but putting this on every day would be a hassle, so I really want the HRM on the bangle.js as it is easy and convenient. |
|
Great! Looking forward to it. Can I take the opportunity to thank you for work? The bangle.js 2 is very impressive. I got one about 3 years ago, but then put it away when I got a Garmin Instinct which was 4 times the price, but I found most of the features quite impractical to use. Recently it lost connection with the Garmin servers and then it became totally useless (and the Garmin support only had standard replies and there was no way to hack it), so I took the bangle out again and really like it and wouldn't want to go back to Garmin. The GPS on the Garmin watch wasn't so useful either as it drained the battery, and the HRM was generally good but also had lots of discrepancies compared to the chest strap HRM, and it didn't always pick up a raised heart rate. So I wouldn't feel too bad about difficulties with the bangle HRM. Actually I think the HRM and also the battery time are much better now than what I remembered them from 3 years ago, perhaps the firmware updates fixed a lot of issues (it was on v15 I believe). So, well done and keep up the good work! |
|
Thanks! That's really nice to hear! Yes, we did make a few changes to the HRM over the years so it should be a lot better. It's interesting to hear your take on Garmin though. The whole server going down thing is a personal bugbear of mine, so I'm very motivated to make sure that doesn't happen with my hardware - even if that means not introducing features (like lots of server-side stuff) if I can't guarantee it'll hang around. I'm happy to say I'm still producing updates for and improving the original Espruino board 13 years later, and I hope to be doing the same with Bangle.js. |
|
Indeed, being so dependent on the Garmin servers was something I didn't quite expect when I bought it, and it always nagged me that I had little control over my data, both in the sense that downloading them was tricky (there are some reverse-engineering scripts because the fie format is not documented) and, increasingly, that health data is something I don't want to share with companies. But when the connection failed I found that the watch didn't even keep the time very well and lost a couple of minutes every day (!) without me noticing it. So, it's a great selling point for the Bangle that it works completely with my own hardware and I have full control. |
|
I didn't know about Garmin servers failing. I have a Garmin watch which I used a lot, and then decided to get a bangle.js 2 because I thought it was stupid that I couldn't see the data in the app unless my phone was connected to the internet. From a hardware point of view, however, this watch is the best gps device I have, and I use it to collect gps data, but everything goes over usb. It uses the mtp protocol, but somehow it is a bit tricky to use, but it is possible. To get files, aft-mtp-cli works pretty well, but it doesn't send files (don't know what the error means). To send files, mpt-sendfile has worked for me, but it is somehow very slow and fickle. |
|
I don't want to talk too much about the Garmin here, but just for completeness, I used garmindb (is somewhere on github) to download and convert the data, that's quite good. Copying the data per USB cable without server is possible, but inconvenient, especially when travelling. And yes, I can set the time with GPS, but again it's inconvenient having to do this at least once a day; that's like an old mechanical watch. For me, this was more hassle than it's worth it and I'm glad I had the bangle and it just works after very little setup. |
|
As I showed in this post (https://github.com/orgs/espruino/discussions/7292), I have modified my Bangle.js 2 for bigger battery (350mAh) and 4 additional buttons and piezo buzzer. I had to sacrifice the HRM to get more digital pins for the switches, but I use this watch for golfing with my golf app (https://banglejs.com/apps/?id=golf-gps, it's a touch panel input version). Even with the larger battery, I could only use it for about 5 hours with the GPS on continuously. Very recently, I added one minute GPS off right after a shot added. Also using one of the buttons, I added manual GPS on/off for long waiting between shots. It dramatically increase the battery life. Now after a 18 hole play I still have 80% of battery with active turn the GPS on/off. Yesterday, I tested only for the 1 minute pause after a shot and the battery was about 58% left. |
|
I'm curious how the accuracy and the battery consumption would be different between the built-in GPS and using the phone's GPS instead. I did check the "Overwrite GPS" option in the Android Integration and had the GadgetBridge installed on my phone and paired with my Bangle.js 2. I opened my golf-gps app on the watch and see if it immediately pass the GPS fix process (it usually take a few minutes if I don't use it everyday) because my phone's location service is always on. But it took the same amount of time to fix the GPS and looked like it's still using the built-in GPS instead of the phone's. |
|
yes, bangle gps gives better results for me than the integration whe overwrite is checked, however when I run with both and at the same time record also on phone, the phone is even better than bangle, just the gadgetbridge integration is somehow erratic for me |
LukasEdl made the integration originally. Should we ping him? |
|
well if someone can see it too. I'm not sure how to verify that my phone is not randomly giving wrong coordinates to gadgetbridge |
Yes, I experienced exactly this symptom and had to switch back to the bangle2 gps for the elapsed distance to be accurate (I use the regular Run app). I also found that the gpx track from Recorder showed random coordinate jumps of several km when using the Gadgetbridge Overwrite GPS, but those could be more easily smoothed out later using eg OsmAnd. I was newer to the watch then and didn't debug further. Maybe we need a new thread for this |
|
I know I mentioned the microphone already, but as I interact more with AI agents, I have built a “tower control” system and am now bridging it with custom software on the Bangle.js. That has made me realize how useful a microphone would be, even beyond recording voice notes. For example, I could speak a thought, task, or instruction into the watch and have it sent through the system to an AI agent. The agent could then capture it, organize it, or act on it without requiring me to pull out my phone or sit down at a computer. As AI becomes more capable and integrated into daily workflows, voice input could make the Bangle significantly more practical. A microphone would enable quick, hands-free interaction with AI agents and make the device more useful and capable than ever before. Sorry for repeating the request/suggestion, but it feels especially relevant given current developments in AI/tech and what is likely to become increasingly useful going forward. |
|
I'm looking at doing this using BlueWatch so iOS users can at least have a dictation option, but i'm not sure how to send audio packets reliably without wifi (to save bangle power) Can it be chunked and transmitted using standard ble packets or will it need a specialized sending function? |
I think the easiest for understanding it that the watch can record it as .wav file and then such file can be synced to phone like any other file (image, text). That would be working "using standard ble packets" whatever that is. However the more standard way = directly working as bluetooth headset won't be there. |
and of course it could be done also at realtime and data could be sent e.g as BLE notifications via some custom BLE characteristics but that is more complicated, more details about recording audio is here https://www.espruino.com/Waveform |
|
I'll probably attempt to send it the same way we send everything at the moment - over the BLE UART. So yes, if BlueWatch can receive the wave data and handle it, that'd be great. I'll have to do a bit more research on the format, but the audio quality isn't going to be great from a tiny mic inside the watch anyway so if I compress it with a voice codec like G.729 (or more likely Opus) it might only be 1kByte/second, which would be trivial to transfer. |
|
you can code up a python server if you use pc |
|
Interesting read. Long answer. I follow bangle js since its first model, never bought one, got close multiple times thought. I am very unhappy with closed source options, I would pay flagship money to get the hardware of current watches with a pure foss OS and even better a js or python api like bangle. What would make me buy a bangle 3
I've had watches where I'd just change the display, for that, a pebble is just superior. What would make me pay flagship money for a bangle3 (pretty sure that's not possible with nordic chips but a man can dream)
Good luck with what ever you can manage. I can imagine how hard it is to study your market and your customer type. Every tinkerer has different wants and perhaps not the budget to match. |
|
What I liked most in the banglejs 2 was its reflective display. It may not be what you want since you asked for more colours (that's 3 bits, à la pebble), but you don't need an always on display at all, it's just always visible under light with minimal power consumption. |
|
The Bangle.js 3 display is a round, 240x240, 6 bpp (64 color) one, and it does do a much better job of images than the Bangle.js 2 3bpp one (especially with dithering). It's a long way from 'popping' like an OLED would though - unless you compare them in the sun :) Thanks for the honest post. Many smartwatches now are effectively smartphones, and I think having something that powerful with a fully open OS would be amazing. Asteroid OS looks interesting, or I assume there must be some open Android builds for some of the Android watches out there that'd be really good. I'm not sure a Bangle.js 3 will really be what you want though. With Bangle.js 3 I'm trying to make something powerful enough to be useful while simple enough that the hardware and software is approachable - and using that simplicity to try and get a decent battery life. Right now I think that means avoiding going down the embedded Linux route. It definitely puts Espruino/Bangle.js in a niche, but it appears to be a big enough niche to support me working on it anyway :) |
and others
Uh oh!
There was an error while loading. Please reload this page.
As I mentioned in https://github.com/orgs/espruino/discussions/7341#discussioncomment-13959214 I'm looking a little more seriously at a Bangle.js 3 now (but we're still 18 months away at least!).
One big question is: is GPS worth it?
If I add GPS I want it to work well, so I'd probably use a Ublox chip. Battery life with GPS should be around 5x better by the look of it, and the fix should be more accurate.
It will add to the price (I'm still trying to get a better idea how much), but the new watch will be more expensive anyway as I'm having to design it from scratch, and will be going for higher quality materials and more memory.
The GPS will also likely push the physical size of the watch up too, but again because I'm hoping to make this one so that it's not glued shut (and the screen is bigger) it will be a larger watch than Bangle.js 2 anyway (although I hope it can be a bit thinner).
Any thoughts? How many of you use GPS? Personally I do use the GPS for walks/runs, and recently even used it for working out lap times on a track day, so I think I'd miss it (especially in 'recorder'). But if I'm being honest, running is the only time I couldn't have just used by phone's GPS as I don't take it with me.
All reactions