WatchUSeek Watch Forums banner

241 - 260 of 309 Posts

·
Registered
Joined
·
12 Posts
- The normal step-tracking might well correspond to some level of physical activity, just not steps. Since 10,000 steps or whatever is pretty arbitrary anyway, it might be enough to say "I want to target 15,000 Casio-steps, because that corresponds well, for me, with 10,000 steps elsewhere."
I've had the exact same thought. I think over a bit of time I'll be better able to determine what my "Casio-steps" should be. As we know, the watch resets the count to 0 at midnight, which here in Norway was about an hour ago. I purposfully took the watch off because I wanted a clean slate tomorrow morning, but woke up hungry about 30 minutes ago. According to the watch I've taken 458 steps from the bedroom into the kitchen to grab a piece of bread. I know for a fact that is way off, heh.

- Even if the pedometer and distance-estimation are completely worthless -- which I'm not yet convinced of, and which might be fixed in software -- I think this is a nice watch.
Agreed. Beyond this particular thing (and some others), my overall impression of the watch is quite positive.

Oh, and @stian133 -- if you have further correspondence with Casio, you might want to include a link to this thread. Social media has a way of incentivizing corporations, as well it should. And, I'd love to hear more of your (and others') in-depth thoughts. That's why I'm sharing mine. Geek is as geek does, if you'll forgive me for likening us.
Jon
Will do. Unfortunately I didn't get an automatic ticket reply to my inbox when filling out the support contact form on casio-europe.com, but I'll pursue this further. Probably takes a few days for a reply to land though, and I can follow up with this thread then.

Cheers!
 

·
Registered
Joined
·
4,848 Posts
I've had the exact same thought. I think over a bit of time I'll be better able to determine what my "Casio-steps" should be. As we know, the watch resets the count to 0 at midnight, which here in Norway was about an hour ago. I purposfully took the watch off because I wanted a clean slate tomorrow morning, but woke up hungry about 30 minutes ago. According to the watch I've taken 458 steps from the bedroom into the kitchen to grab a piece of bread. I know for a fact that is way off, heh.



Agreed. Beyond this particular thing (and some others), my overall impression of the watch is quite positive.



Will do. Unfortunately I didn't get an automatic ticket reply to my inbox when filling out the support contact form on casio-europe.com, but I'll pursue this further. Probably takes a few days for a reply to land though, and I can follow up with this thread then.

Cheers!
I noticed that mine seemed to count steps when sedentary, so I was really curious about it. Just for fun I then counted my steps to the subway stop which is a little more,than a 20 minute walk and it was surprisingly dead on. For me, the problem seems to be that it is too sensitive when stationary and not actually walking.
 

·
Registered
Joined
·
20 Posts
I respect that dedication, @stian133 -- I'm not sure I'd have the wherewithal to put my watch on before hunting a midnight snack. But. For science. I'll try taking it off tonight and checking the step count before and through the day tomorrow
 

·
Registered
Joined
·
20 Posts
If screen burn in is an actual issue then why does the display only turn off for 3 hours?
My best guess so far: this has to do with image persistence (wikipedia). I haven't heard of LCDs being permanently burned, nor of MIP being used in anything other than a LCD. So, a few hours can relax the pixels and let them change state more efficiently ... and also save battery (from updating the screen every second or so). With an LCD, you can fix the problem by removing power for a while (maybe an hour, maybe a week); this contrasts with OLED displays, which permanently degrade over time. LCDs just need a bit of a vacation from time to time, as do we all.

I don't fully understand the hardware used here or the issues with the hardware I think is being used, but this is my understanding from folks who do seem to know such things.

tl;dr: not something to worry about.
 

·
Registered
Joined
·
7,000 Posts
Has anyone tried selecting a Southern Hemisphere port to see if the illuminated portion of the moon phase graphic is correctly flipped left for right? Technically it's also upside down, but the lunar markings aren't visible on the watch, so this is of no concern.


Northern Hemisphere

phases.gif


Southern Hemisphere

phasessouth.gif
 

·
Registered
Joined
·
7,002 Posts
For those of you who have the watch: How accuraty are you finding the pedometer? On mine, the step count is way off. I've had the watch for a few days now, and I like to wear my watches 24/7. When I wake up in the morning I've already taken around 800 "steps".

Today, my phone has been in my pocket all day as a comparison, and it reports 898 steps (I haven't moved much, I know), but the watch is reporting 2839 steps. Thats two-thousand steps more.

I've asked around reddit a bit and it's not just me reporting this behaviour.

I've updated the firmware (version 03) and set my height, weight, sex, etc correctly.

So: How are you guys finding the pedometer? Also, if you have other g shocks with pedometers, how does it compare to those?
put your watch in your pocket and tape your phone to your wrist :-!
 

·
Registered
Joined
·
6 Posts
Just got it today. I like it a lot but find that the tide information on G Shock Move app which used to sync with the watch is not correct. For example, on my location in Vietnam last night, the next low tide is 1039pm (check from many of tides info web sites) but in Casio app shows the same time at 1039pm but as the next high tide which was opposite and also makes the tide information in the watch incorrect. I am not sure this is a bug or I miss something during set up. But I am sure I had already set up the correct home time, etc.
 

·
Registered
Joined
·
12 Posts
I got this on reddit today. Hopefully it is a calibration issue that'll sort itself out.

I dunno about the GBX, but my GBD-800 took about 2 weeks to record my steps correctly! In early days, it said I did 20k steps when in reality it was probably close to 14-15k steps. But now it more or less records the same number of steps as my Google Fit app and/or Samsung Fit band. Maybe it still needs to get use to your daily activities?
 

·
Registered
Joined
·
20 Posts
Has anyone tried selecting a Southern Hemisphere port to see if the illuminated portion of the moon phase graphic is correctly flipped left for right?
Using the preset Rio de Janiero port for today, the left side of the moon is lit; for San Francisco, the right side is lit. Same way with Melbourne vs Portland. I'd say this works as expected.
 

·
Registered
Joined
·
20 Posts
Another run this morning, after setting both phone and watch to use miles, the watch (and imported run log on the app) still report values off by a factor of 1.6.

Otherwise, accuracy continues to impress. As part of a ~2.6 mile run, ~1 mile was around a park where I avoided mapped trails. I didn't do manual laps on my phone and I don't want to dig into a GPX file, so a rough comparison between GPS mile 1 and mile 1 vs that same portion mapped on milermeter.com; and between the phone's manual lap of the off-map portion vs milermeter.com shows:

- GPS was 5.4% short of milermeter.com
- GBX-100 was 4.7% short of milermeter.com

I'm not going to say the watch was better, but I am going to say that the GPS isn't better than the watch when the GPS wasn't able to use established roads or trails to calculate distance. This was tested in a park in flat ground, under some tree cover but 90% with a clear view of the sky (e.g., next to a pond, through a baseball outfield). The watch can only look better in comparison where GPS signals are difficult to receive.

Tomorrow I will run with my watch connected to the phone, with GPS enabled, and with the same GPS logging app recording at the same time. This should make the watch even more accurate, and might correct for the unit confusion.

I've been running at a fairly relaxed pace (~10 min/mile). My legs are fairly tired from a busy month, otherwise I'd try some sprinting before calibration, but I think that'd just be begging for an injury right now. So, I'll report on more aggressive running when my legs are ready for that.

---

The pedometer reported that I took 50 steps while washing my hands and 200 steps while putting on my running outfit. I do hope this shakes out.
 

·
Registered
Joined
·
535 Posts
I appreciate your updates a lot, but as a general question: how is a watch that is essentially computing footsteps by wrist movements (this is an algorithmically intractable problem) supposed to filter out the motion of a wrist while washing ones hands or putting on clothes?

Are you just saying that you want your watch to measure the same number of steps as other watches that compute footsteps based on wrist movements?

I guess I don't really understand how CASIO is supposed to solve that problem. There's no software that could do that accurately without the input of additional hardware sensors. I suppose they could try to make the sensor less sensitive to movement, but that could have an undesirable effect? Perhaps with some kind of interventional calibration, but is there even a calibration that a user can perform?

There's other questions too...is the GBX-100 using an IMU for step measurement? orthogonally arranged ball bearings? Is it an analog measurement or a digital one?
 

·
Registered
Joined
·
12 Posts
I appreciate your updates a lot, but as a general question: how is a watch that is essentially computing footsteps by wrist movements (this is an algorithmically intractable problem) supposed to filter out the motion of a wrist while washing ones hands or putting on clothes?

Are you just saying that you want your watch to measure the same number of steps as other watches that compute footsteps based on wrist movements?

I guess I don't really understand how CASIO is supposed to solve that problem. There's no software that could do that accurately without the input of additional hardware sensors. I suppose they could try to make the sensor less sensitive to movement, but that could have an undesirable effect? Perhaps with some kind of interventional calibration, but is there even a calibration that a user can perform?

There's other questions too...is the GBX-100 using an IMU for step measurement? orthogonally arranged ball bearings? Is it an analog measurement or a digital one?
I can't really answer your questions, other than to say that I've worn other watches in the past that have done a much more accurate job.
 

·
Registered
Joined
·
3,473 Posts
I appreciate your updates a lot, but as a general question: how is a watch that is essentially computing footsteps by wrist movements (this is an algorithmically intractable problem) supposed to filter out the motion of a wrist while washing ones hands or putting on clothes?

Are you just saying that you want your watch to measure the same number of steps as other watches that compute footsteps based on wrist movements?

I guess I don't really understand how CASIO is supposed to solve that problem. There's no software that could do that accurately without the input of additional hardware sensors. I suppose they could try to make the sensor less sensitive to movement, but that could have an undesirable effect? Perhaps with some kind of interventional calibration, but is there even a calibration that a user can perform?

There's other questions too...is the GBX-100 using an IMU for step measurement? orthogonally arranged ball bearings? Is it an analog measurement or a digital one?
I don’t exactly know how the accelerometer works, but good ones use various factors to tell if you are walking or doing something else. I have put fitness trackers to the test from Fitbit, Garmin and COROS, and they generally don’t register steps in regular activity or in a car or going up an elevator. I find the iPhone steps to be less accurate than running watches I have tried as fitness trackers. Not sure if it’s because it’s in the pocket, or just less refined overall. I will be interested to pit the GDX-100 against my Coros Pace GPS watch which I usually wear to record running and walking.


Sent from my iPhone using Tapatalk
 

·
Registered
Joined
·
20 Posts
...how is a watch that is essentially computing footsteps by wrist movements (this is an algorithmically intractable problem) supposed to filter out the motion of a wrist while washing ones hands or putting on clothes?
You're right that this is a difficult problem. Here's one paper that discusses it. As I understand it, the basic problem is to find a pattern of accelerometer data that includes the regular arm movements in walking but excludes other movements. Putting on clothes doesn't have large motions that repeat over 10+ seconds, which the operations guide suggests is the threshold. Hand washing includes repeated motions but not of the same duration or magnitude of velocity changes as in walking. I'm not saying I could do better than Casio, just that these inaccuracies seem fixable.

Then again, I'm comparing it with a device that didn't have an instant digital readout of steps, so maybe I've always counted hand washing in my steps and just never knew about it before now. :)

What I want is a reasonably linear measure of physical activity. A report of the volume under the acceleration curves in each axis might be a better measure than "steps." I don't know. What I do know is that I'm uncertain as to what is being represented by the number of "steps" that my watch counts, and that uncertainty diminishes my perceived value of this function.

You might have more understanding of the problem than I do; if so, feel free to tell me I'm confusing the difficulty of flying to Chicago with that of going to the moon.
 

·
Registered
Joined
·
1,757 Posts
Super annoying fact: when changing the time manually, you cannot adjust the seconds.

So if your seconds are off, which they will be at some point (my experience with all G-Shocks is that they run fast), you must connect the watch to the Bluetooth phone app.

For most people buying this watch that probably won't be a big deal (since a majority of people will specifically buy this watch to use with their phone for the added features), but I have no interest in connecting it to my phone whatsoever.
Oh well, I guess I will have to hook up the app once a month to keep it accurate.

Other than that minor annoyance, this G is a home run!


"boys support boys"
 

·
Registered
Joined
·
20 Posts
Can anyone confirm that their watches provide remotely accurate distance estimates for runs?

When I try to calibrate with GPS, I see no meaningful change. When I try running with metric units, I get distance estimates that are off by a factor of 2.56 (1.6 x 1.6). As far as I can tell, the primary function of the headline feature of this watch is fundamentally broken. It might just be my watch, however, so if other folks are getting good estimates then I'll treat this unit as defective rather than this design.

I think this is an excellent platform. I hope my issues can be fixed with an update. But I am disappointed at the apparent lack of testing shown here.
 

·
Registered
Joined
·
3,473 Posts
Despite my previous posts about trying to get one of these, I eventually overpaid to get the black one from Japan. It’s here, and I think I like it. The display is phenomenal...still perfectly legible at an angle in direct sunlight.

The app is relatively easy to use, though a little frustrating to change the units. I changed it on the watch, but then the profile setup on the app defaults to metric, which resets the watch units. You can change it back, but only once you get past the initial setup, which is annoying.

I look forward to testing how well it measures distance. I do a lot of walking, and know the approximate distance of every intersection and landmark in range.




Sent from my iPhone using Tapatalk
 

·
Registered
Joined
·
12 Posts
Can anyone confirm that their watches provide remotely accurate distance estimates for runs?

When I try to calibrate with GPS, I see no meaningful change. When I try running with metric units, I get distance estimates that are off by a factor of 2.56 (1.6 x 1.6). As far as I can tell, the primary function of the headline feature of this watch is fundamentally broken. It might just be my watch, however, so if other folks are getting good estimates then I'll treat this unit as defective rather than this design.

I think this is an excellent platform. I hope my issues can be fixed with an update. But I am disappointed at the apparent lack of testing shown here.
Steps and distance are atrocious on mine as well.

I'm done with this watch for now, putting it on the shelf and will only check from time to time if new firmware is pushed.

Maybe nobody cares, but to those who are considering buying, I've compiled a list of things to consider (not exhaustive). Cells/scores marked with an upper-right tick contain a comment made visible upon click/hover. See list: Google Sheets

It's a great looking watch, but it's become a source of irritation for me at this point. Everytime I look at it all I can think about is how poorly they've tested it. It seems they didn't care at all before pushing this to market. Wearing the GW-B5600 again feels like a relief in comparison.
 
241 - 260 of 309 Posts
Top