Turning AirPods into a fitness tracker
richarddas.com
richarddas.com
It’d work for pushups, squats, chinups, etc etc. So many exercises are the same idea of “move body repeatedly”.
I don’t want my phone in my pocket while doing it, I like having my phone in front of me while I do pushups/situps/chinups/etc so I can see the timer/stats.
I’d happily pay for an app that could do that and just let me tag what each one was to add to a log.
I'm mostly doing dips pushups pullups and abs, and it counts each rep.
The simplest form is starting a generic strength activity, where pressing the lap button moves to "rest" and displays a quick optional "Hey I counted this, fix the count" thing before resuming to the next exercise with the lap button. Once you're done with your workout and it synced to the phone both each exercise type and reps can be fixed after the fact.
You can also set up more structured, scripted workouts with exercises known beforehand, which slightly improves rep counting (otherwise it's very generic and approximate)
When I enter an activity such as weight lifting it seems to be able to infer what exercise I performed, it seems like it should be able to infer when I am going for a run or a walk.
Might have to do with the fact that during an activity it's actively tracking, full sensors blaring (notably HR) whereas otherwise it's lazier to save battery.
Still, I agree it'd be nice to have the option to either auto-create after the fact, or auto-start (and stop) live, or both.
This is on the Forerunner 945 and will auto tag things like walks, runs, and cycling.
I use Strong. It’s free to log and works perfectly. It costs to get all the graphs and I found it to be too expensive for what it is. But it has a free export function to csv, so I’m making my own PowerBI dashboard. Just have to export once a month to see the progress for example of max weight, calculated 1RM, (max) volume per exercise.
Looking for alternatives tho.
log weightxreps@rir in comma seperated long text field, calculates repRMs, and other stats like volume/inol. Visualizations in appsheets very weak though.
Most important feature for me, having gone off program/fuckarounditis is having UI that clearly indicates how long ago I've done an exercise, and associated stats / percentages. This way I can train intuitively using my own meso/microcycles, i.e. right now my microcycle is is every 10-12 days instead of 7, which helped me blow past some stalls. I try to visualize in: how many days ago I last did an exercise [how many rest days in between]. I can glance at exercise charts (can be fuzzy to match all exercise of certain tag, i.e. #bp shows data for all bench variations), and get a sense of periodization/progress. Trying to figure out a fatigue/recovery system (PRF = performance, RCV = fatigue). Big part is finding out how body responds to stress in various blocks, i.e. figuring out whether to push/add sets despite soreness/fatigue. Anyway apart from occasionally slow syncing, the new appsheet has made logging very convenient, since it precalculates dates/sesssion, just have to pick exercise and enter weightsxsets.
TLDR flexible training.
---
Before this, I've used for years where you log using syntax in giant textarea:
Their website is a little busy and they had some redesigns that I didn't like, but it's easily one of the best logging experiences out there since you're not wasting time filling individual setsxreps etc. They have an app now, which I haven't tried, but the web experience was very good for me. I logged in a note, and copy and paste into log file when done. Visualizations pretty good.
---
My ideal logging, if I knew how to code (only decent in excel formula), would be basically something like a population pyramid style chart where y axis is training days, x axis is inol/weight x or whatever stat. You tap on the horizontal/bar training day to view relevant details but otherwise have high level overview of essential stats.
https://epirhandbook.com/en/images/pop_pyramid_baseline.png
Last couple photos in album shows old spread / visiualization sheet format, the 2nd last photo last like rows that represent past training days (empty = rest days) with inol for each exercise tag. I've blacked out some of my repmaxes in case it identifies me online powerlifting competition DBs.
Disclaimer: Anorak Ventures is an investor.
A combination of watch + airpods should probably solve the overall body pose and positioning.
However, people would still want videos for reference and to get visual feedback.Perhaps not. Just an audio feedback supplanted by occasional video should help.
(1) Thanks for the writeup, this is classic HN content, a simple app that solves a person's problem with enough interesting technical detail to make it interesting.
(2) Definitely post this to the App Store! I'd happily pay a few $ for it even.
(3) How would it handle burpees? Would it just ignore the 'jump' phase? But still count the push-ups since it's zeroed in on the up/down motion of the push-up itself?
Since there’s interest, I’ll publish the source code to help others learn, and put the app on the App Store for those who just want to use it.
Regarding burpees and other types of motion, I’m looking into their motion characteristics, so that I can include them for detection.
Burpees would likely require recognizing different phases (the jump, the stand, the push, etc) and then when it detects them in sequence, incrementing the count. Something like that, but stay tuned! :)
https://www.dcrainmaker.com/2024/05/sennheiser-momentum-temp...
Respiration rate is harder. It requires precise heart rate variability data to detect the start of a breath.
Respiration can be measured with microphones, and again, movement, but I suspect this may be more challenging than filtering HR.