Calibrating gravity and magnetic field sensor - Android Q&A, Help & Troubleshooting

Hello everybody
I'm using an Android 7.0 phone (Huawei P9 Plus) and I'm recording sensor data.
The recordings of the gravity sensor (`Sensor.TYPE_GRAVITY`) show always unreliable accuracy (using `SensorEvent.accuracy`). Is there a possibility to calibrate the gravity sensor?
For the magnetic field sensor (Sensor.TYPE_MAGNETIC_FIELD) I'm getting in the beginning also unreliable accuracy but after moving the phone a bit around it first changes to low accuracy, then to medium accuracy and then to high accuracy. Why is that? Can it not be all the time at high accuracy, also right from the start?

Related

sensor sampling rate

I have noticed, first with Speedx, but now in detail with the use of Z-DeviceTest that (at least) the accelerometer and compass are full of jitter and noise. They don't vary widely, but the sample rate is so high that even with the tablet sitting on a stable surface the X, Y, and Z are constantly changing at a rate of ~30Hz. Now, I also ran this same app on my Atrix and the axis values don't change as long as the device is untouched on a stable surface. When I move the Atrix then the axis values do update at a rate of ~30Hz, much like the Jetstream. The difference with the Jetstream is that even when physically stable the axis sensors are constantly changing and full of jitter.
Is this something you guys observe with your Jetstreams? As I said, it's typically not a big deal during normal use, but is a big annoyance when playing accelerometer-dependent games like Speedx. Using a diagnostic app really shows how noisy the sensors are even when the tablet is unmoved. Is there a way to smooth this out using some kind of filtering or sampling rate tweaks? I appreciate the accuracy, just not the noise.
Thanks,
Mike

[Q] Heart rate sensor accuracy

Hi. Just wondering if anyone has tested the accuracy of the heart rate monitor during vigorous exercise? I am cautious about this unit, given how poorly the HRM works on the Gear 2 Neo. I have a great fitness watch with an accurate strapless HRM, but I would love to combine the features of a smart watch with a fitness device. The Gear S is the best option at the moment, but I have not seen any valid comparison against a device that is known to be accurate.
Thanks!
I have worn the gear with my polar hrm at the same time. If you sit there and stare at it, sometimes it can look off from the polar. However, when the workout is over and you compare the two, the readings are comparable. Accurate enough for me.
I haven't done enough tests to be 100% sure, but using a blood pressure monitor(it shows heart rate as well) and then the watch right after. The difference was about 2 bpm. So while I can't be sure it is accurate in all scenarios and I only tested once. The results seem to be accurate.
Thanks for the replies. I don't need it to be perfect, so it should be OK. The Gear Neo is useless as a training tool, as it cannot hold the heart rate consistently. If the Gear S only has the occasional spike or drop, it should be OK.
Can the Gear S sync with a bluetooth heartrate monitor?
In my experience Gear 2 works fine, if you don't sweat.
If you want better accuracy, look into two sensor devices like Apple Watch, Fitbit HR, or Mio Alpha (especially MIO).
I would not expect any accuracy improvement, from any single sensor Hear Rate monitor.
I tested it a few times while at doctor appointments. The nurse did a manual reading the old fashioned way and it was always within 5 or less.
Sent from my SM-N915V using XDA Free mobile app
-_-

Amazfit fakes heart rates?

Hi,
I have an e-bike. Just to test the GPS accuracy of my new watch i used the "bike" sport program. The top speed of my bike is around 20-24 km/h. The GPS tracking went really well, I am impressed there. However I saw rather strange and high heart rate readings. It was 113 bpm in average and 146 bpm as highest. I actually do not have much physical activity using the e-bike in full assist mode. I really have to push hard myself to get my heart rate to about 150 pbm.
I did repeat the test and got the same result.
So I have to conclude that, the watch checks my speed and thinks I am pushing hard on a "real" bike. It does not take the actual sensor reading, but something "pre-programmed" = fake one.
ventura1977 said:
Hi,
I have an e-bike. Just to test the GPS accuracy of my new watch i used the "bike" sport program. The top speed of my bike is around 20-24 km/h. The GPS tracking went really well, I am impressed there. However I saw rather strange and high heart rate readings. It was 113 bpm in average and 146 bpm as highest. I actually do not have much physical activity using the e-bike in full assist mode. I really have to push hard myself to get my heart rate to about 150 pbm.
I did repeat the test and got the same result.
So I have to conclude that, the watch checks my speed and thinks I am pushing hard on a "real" bike. It does not take the actual sensor reading, but something "pre-programmed" = fake one.
Click to expand...
Click to collapse
Hmm, it's probably that. Although it gives me approx. the same readings on my regular bike, but I'm just a fat bastard.
If you have a chance to wear some other heart rate bracelet simultaneously, do that, and compare results.
Sent from my LG-D858HK using Tapatalk
Their faking algorithm is well known to be very advanced... it also takes temperature, atmospheric pressure, activity duration, your weight, height, and age into account. It also manages to pass comparative tests to other similar devices and external heart rate monitors, and detects when you give your watch to another person and measures his HR correctly. Ah, yes, and yesterday I saw some bright lights in the sky - be prepared, the aliens are coming.
trueruer said:
Their faking algorithm is well known to be very advanced... it also takes temperature, atmospheric pressure, activity duration, your weight, height, and age into account. It also manages to pass comparative tests to other similar devices and external heart rate monitors, and detects when you give your watch to another person and measures his HR correctly. Ah, yes, and yesterday I saw some bright lights in the sky - be prepared, the aliens are coming.
Click to expand...
Click to collapse
Hi,
Can't decide what you are saying... So you mean the measurements should be correct and I am "making up" the complaints on the bad readings? Or else?
ventura1977 said:
Hi,
and I am "making up" the complaints on the bad readings?
Click to expand...
Click to collapse
it's different to say that "i get bad readings" (or not very accurate) and different "i get fake readings"
btw gps has nothing to do with the hr measures (common sense , gps is only for outdoor activities)
ventura1977 said:
Hi,
Can't decide what you are saying... So you mean the measurements should be correct and I am "making up" the complaints on the bad readings? Or else?
Click to expand...
Click to collapse
I've just showed you what happens when way of thinking is flawed, based on what I see when using only my amazfit device and taking your hypothesis that there is a faking algorithm involved. Nothing more. I suggest you read this article: https://en.wikipedia.org/wiki/Scientific_method , then continue your investigation into why in your case the watch is showing faulty(if it is faulty) data and in what conditions. Otherwise your proposition about faking is no different than the one about an alien invasion coming every time there is a bright light in the sky.
I kinda ran into a similar issue. I was wearing the Pace past the wristbone and had my Schosche Rythmm+ on the forearm connected to my Under Amour App on my smartphone and went for a walk/Run - the heartrate readings and caloric burn were way off. The highest my Pace Watch would show is about 110 or 120, but i was running 400M intevals with 2:30 minutes of rest in between. My Rythm was showing 140-160 on the sprints. Which is about accurate.
So I am question the ability and accuracy of the device. If we could connect to an external HRM via Bluetooth this watch would be killa!
I really like the watch and not trying to make up stories guys. It was a simple observation about an odd behaviour.
Somewhere I have saw posted that single measurements seems to be accurate, but if you ran a sport program it is way off.
I suppose the easiest way is to verify this is to run a program while sitting in the car or bus with gps good signal.
Is there a possibility that while on your bike felt adrenaline when you were going fast causing your heart rate to go up?
Honestly, if you want accurate heart rates get an Amazfit Health Band w/it's ECG sensor. http://amazfitcentral.com/amazfit-health-band-specifications/
I use mine for jogging. The first firmware was very accurate but after the OTA update it always said that heart rate too high, over 170. Then I have to keep the strap pretty tight to get better reading.
Then I changed the firmware to PACIfied version. Reading was good again & without too much tightening of strap.
I think the watch fakes it too.
I think it fakes it too.
Did some tests. Ran an intervall where it got all the way up to 195. Which was 10 beats off what the treadmill shows. Which is fine, not expecting it to be that accurate so high.
Then I walked flat for 6kmh and it detected 110 which is also OK. Then I walked 6kmh with 10 degree uphill.
Now watch still show 110, mill showing 150.
It doesn't detect the uphill.
I'm so extremely disappointed in this thing. Wasted 110$
jhenrikb said:
I think it fakes it too.
Did some tests. Ran an intervall where it got all the way up to 195. Which was 10 beats off what the treadmill shows. Which is fine, not expecting it to be that accurate so high.
Then I walked flat for 6kmh and it detected 110 which is also OK. Then I walked 6kmh with 10 degree uphill.
Now watch still show 110, mill showing 150.
It doesn't detect the uphill.
I'm so extremely disappointed in this thing. Wasted 110$
Click to expand...
Click to collapse
I'm going to run on the treadmill for thirty minutes, then put the watch on as soon as i stop the machine and see what the heart rate is.
FYI, had a Cambridge University educated doctor take my pulse and he said it was a lot less than what the amazefit watch was saying.
[Amazfit][Bip] heart rate sensor is a joke (firmware 0.9.40)
Hi, I have purchased the Amazfit Bip recently, it was cheap.
Several threads on the heart rate sensor of the Pace have try to warn against a possible fake feature.
I would like to extend this notice on the Bip after few tests.
I have posted a video on youtube to illustrate. unfortunately I am not able to post the link, here the title :
"Amazfit Bip heart rate sensors test."
Hope this helps.
here the description of the video :
A video of a heart rate test of the Huami Amazfit Bip watch. It can be seen that the watch does not offer a measurement but an estimate of the heart rate. This estimation rely on the time elapsed in the activity as seen in the first part of the video; but also on the information that can give the optical sensor, in the second part of the video. After turning off the camera, I started to move and put away my stuff and thus stress the accelerometer: the estimate of the heart rate has recovered.
The problem is that my heart rate could not exceed 60 bpm during the whole test (nothing doing it is around 54).
After several tests, this estimate seems made from several parameters:
- time: over time the heart rate increases in the effort
- what the optical sensor gives.
- the accelerometer : to detect an activity. (although in the first part of the video the watch does not move)
- parameters on age, weight, height, reported in mi fit, to establish a range of variation in the estimation of the heart rate.
Why that ? Maybe the sensor is unable to provide a measurement: the flash intensity is very low on this watch, the sensor covers a very small area.
What I expected: a measure like my polar f6
(without signal nothing is displayed: "--")
With a phone that has GPS and pedometer, it would have been better to invest in a Bluetooth module with electrodes.
Toni Maltes said:
Hi, I have purchased the Amazfit Bip recently, it was cheap.
Several threads on the heart rate sensor of the Pace have try to warn against a possible fake feature.
I would like to extend this notice on the Bip after few tests.
I have posted a video on youtube to illustrate. unfortunately I am not able to post the link, here the title :
"Amazfit Bip heart rate sensors test."
Hope this helps.
Click to expand...
Click to collapse
You're right, when missing data, the BIPseems to interpole the resullts with the data available.
If you put the BIP on your wrist during that time, it sloooooowwly go down to the HR. It's still rely to the old incomplete data that globally at this moment gives a false result.
That could be improved by software I hope.
EMPTY
RisenVe said:
Same here,
shaved my hairy arm at the bip position. Measured while riding my bike. Drived with 23km/h over all time and some sprinting intervalls to vo2/max. The Bip says at normal riding 110bpm +/-5 beats. But then, holy moly, at sprinting.... 130bpm +/-5 beats. On right arm i have used my mi band 2, on same height compared to bip. It says heart rate is between 80bpm at start up to 180bpm in sprint intervalls.
Click to expand...
Click to collapse
Are you sure it's tighten well ?
The sensor, is fully on your skin and doesn't move while riding?
Toni Maltes said:
Hi, I have purchased the Amazfit Bip recently, it was cheap.
Several threads on the heart rate sensor of the Pace have try to warn against a possible fake feature.
I would like to extend this notice on the Bip after few tests.
I have posted a video on youtube to illustrate. unfortunately I am not able to post the link, here the title :
"Amazfit Bip heart rate sensors test."
Hope this helps.
here the description of the video :
A video of a heart rate test of the Huami Amazfit Bip watch. It can be seen that the watch does not offer a measurement but an estimate of the heart rate. This estimation rely on the time elapsed in the activity as seen in the first part of the video; but also on the information that can give the optical sensor, in the second part of the video. After turning off the camera, I started to move and put away my stuff and thus stress the accelerometer: the estimate of the heart rate has recovered.
The problem is that my heart rate could not exceed 60 bpm during the whole test (nothing doing it is around 54).
After several tests, this estimate seems made from several parameters:
- time: over time the heart rate increases in the effort
- what the optical sensor gives.
- the accelerometer : to detect an activity. (although in the first part of the video the watch does not move)
- parameters on age, weight, height, reported in mi fit, to establish a range of variation in the estimation of the heart rate.
Why that ? Maybe the sensor is unable to provide a measurement: the flash intensity is very low on this watch, the sensor covers a very small area.
What I expected: a measure like my polar f6
(without signal nothing is displayed: "--")
With a phone that has GPS and pedometer, it would have been better to invest in a Bluetooth module with electrodes.
Click to expand...
Click to collapse
very interesting test, I already realized that the cardiac sensor of this watch gives me strange results sometimes.

Amazfit GTR sensors

while we are waiting situation with terrible GPS to be resolved in some way....
does anyone know how rest of GTR sensors are being used. specification says:
BioTrackerTM PPG biological tracking optical sensor, 6-axis acceleration sensor, 3-axis geomagnetic sensor, Air-pressure sensor, Capacity sensor, Ambient light sensor
Click to expand...
Click to collapse
what are geomagnetic sensor, air-pressure and capacity sensors being used for?
There is an altimeter, and pressure sensor inside that you could check within some menus to check how high and the air pressure.
Geomagnetic is the magnetic North.. compass app will help you here

Question Screen dims a lot in low light (ligh sensor poor sensitivity?) - POCO M4 Pro 4G

Hello everyone, has anyone noticed that the light sensor of the Poco M4 Pro 4G seems to have poor sensitivity in very low light environments (compared to other models)?
In low-light environments, it dims the screen a lot (automatic adaptive brightness). Using the "Sensor Test" app or entering the MIUI Test Tols ( * # * # MIUI # * # *) -> light sensor, I compared the light sensor sensitivity of this model with two other models: Redmi Note 6 PRO and Poco X4 PRO 5G. I noticed that the sensor of the Poco M4 Pro 4G seems to be less sensitive in low light environments. For example, in the room with only the brightness of the TV, the Redmi Note 6 PRO detected 5lx and the Poco X4 PRO 5G about 10lx. The Poco M4 Pro 4G, zero lx.
In environments with a more light and outdoors, it works perfectly. Only when it is a place with very low light is it less sensitive. And that makes the screen darker more. I don't believe it's a hardware defect, because the sensor works normally and detects brightness variation. In environments with normal or external light, its sensor reading is close to that of other models.
Did you notice that too?
In my case it's either too bright or too dark. Tbh I'm yet to come accross a headset that has a flawless autobrightness feature. Maybe this just doesn't work well in the budget smartphone area. That's why I don't rely on the autobrightness at all, I keep changing the brightness as I feel.
At times it just dials down the brightness all of a sudden without any change in the light conditions, weird.
I have a feeling it's more of a software issue than a hardware one.

Categories

Resources