Do you really need your iPhone all the time?

[Julien]

...With a programmed stride length, the UP band, with just it's accelerometer, is darn accurate to the distance my TomTom reports. So I'm calling BS on the claim that accelerometer distance measurements are poor and always off. When done right, they can be quite good.

They can be "darn accurate" or bad or in-between. However they can NOT offer location or tracking (map) info of your run as was stated by Tanegashima. This was the gist of the debate.

If you calibrate, have a constant cadence (170 to 190 is ideal [running, not sprinting] and should be maintained at all running speeds) and a consistent stride length they will be accurate. One problem is your running speed is determined by your stride length [cadence X stride length = speed] (unless you are unable to keep constant cadence). So if you slow down or speed up your accuracy will drift.

 

[flur]

They can be "darn accurate" or bad or in-between. However they can NOT offer location or tracking (map) info of your run as was stated by Tanegashima. This was the gist of the debate.

I said I wasn't an engineer, not that I had trouble reading. I got the debate. I'm responding the whole "it's inaccurate for distance" because no, when configured correctly, it's not.

TIf you calibrate, have a constant cadence (170 to 190 is ideal [running, not sprinting] and should be maintained at all running speeds) and a consistent stride length they will be accurate. One problem is your running speed is determined by your stride length [cadence X stride length = speed] (unless you are unable to keep constant cadence). So if you slow down or speed up your accuracy will drift.

So now you're going to tell me how fast I'm supposed to run? Lol!

If you calculate your average stride length properly, you won't have so much trouble. This isn't rocket science here.

 

[Julien]

.....So now you're going to tell me how fast I'm supposed to run? Lol!

If you calculate your average stride length properly, you won't have so much trouble. This isn't rocket science here.

Of course not, it just simple mathematics. Cadence X stride length = speed. So as you speed up you stride length increases and steps per mile decrees. As you slow down it shortens and steps per mile increase. For instance 180 cadence X 1.3 yards stride length = 8MPH running speed. 8MPH is 1 mile in 7:30, so 7.5 x 180 = 1350 steps per mile. Now slow down to 6MPH (10 minute mile) and you have 10 x 180 = 1800 steps per mile. That would be about a 30% difference in distance measured between 8MPH/1350 steps per mile and 6MPH/1800 steps per mile on a calibrated accelerometer.

 

[samiznaetekto]

Apple confirms it uses sensors in conjugation with GPS from the phone. That's a fact.

Link to your "fact"?

Simple, use the iPhone inside a faraday cage. Enough proof.

Link to your "proof"?

And there's proof that a low pass filter has been applied to it, because in previous iOS versions, it wasn't.

Again, where is your "proof"? I just downloaded and ran AccelerometerSample from Apple on my iPhone, set the update frequency to 100Hz, and bumped, kicked, and drummed like a rock star - it shows on the raw data graph as noisy/responsive data as it can be. I see no signs of LPF. So, show me your proof or stop trolling.

Civilian's gps is only good for about 1.1 meter precision, the iPhone can track better than this.

Where's your "proof"?

You have zero knowledge of physics and real world limitations.

You have zero knowledge of my knowledge of physics, iOS programming, sensor fusion, and real world limitations.

 

[samiznaetekto]

Do you know that it's NIKE+ app and the design has nothing to do with Apple's at all, right?

Ok, my mistake. But my point is still valid. Look at the Health app.

-1047 steps? 6404?

Look at the Stocks app. Vertical scales.

Scales are supposed to be in nice, easy to interpolate, intervals. 1, 2, 5, 10, 20, 50, 100, etc. When the interval is from 7.34 to 13.26, how can you estimate - without PIA - a point half way between?

 

[matrix07]

Look at the Stocks app. Vertical scales.

Image

Scales are supposed to be in nice, easy to interpolate, intervals. 1, 2, 5, 10, 20, 50, 100, etc. When the interval is from 7.34 to 13.26, how can you estimate - without PIA - a point half way between?

Scale? I don't have old iOS at the moment but I believe that's the price of the stock. 19.18 was the highest and 1.42 was the lowest so you got two important informations together at a glance.
Personally I prefer the old version. It's much clearer.

Oh, and you don't have to estimate anything. You just turn your iPhone horizontally and drag a finger across the chart and it will show you the price at any given time.

It's a powerful app that is so simple and easy to use, like most of Apple apps.

 

[samiznaetekto]

19.18 was the highest and 1.42 was the lowest

You just turn your iPhone horizontally and drag a finger across the chart and it will show you the price at any given time.

And what do you see in between? I'm looking at 2Y stock in landscape right now. High 116.47, low 55.79. What do you see between them? Any special meaning of these random-looking numbers? Where are the grid lines located?

 

[matrix07]

And what do you see in between? I'm looking at 2Y stock in landscape right now. High 116.47, low 55.79. What do you see between them? Any special meaning of these random-looking numbers? Where are the grid lines located?

The meaning of it is obvious, isn't it? Even the kid will know. You're just annoyed that it's not .00, that's all.

 

[samiznaetekto]

The meaning of it is obvious, isn't it? Even the kid will know. You're just annoyed that it's not .00, that's all.

Yes, the meaning of it is obvious to me: the programmer who did it was too lazy or incapable and simply divided the range in 4, instead of implementing a more laborious, complex logic that make scale values rounded and looking nice. It's the same indication of low quality of programmers as the now famous pie chart with parts adding up to 101%.

 

[matrix07]

Yes, the meaning of it is obvious to me: the programmer who did it was too lazy or incapable and simply divided the range in 4, instead of implementing a more laborious, complex logic that make scale values rounded and looking nice. It's the same indication of low quality of programmers as the now famous pie chart with parts adding up to 101%.

And that would be stupid, because complexity isn't necessarily better than simplicity. Most of the time it's the opposite, but that would take an intelligence to know it.

 

[Mr. Buzzcut]

I have yet to see anyone prove accuracy of movement tracking through actual results. Sounds like the holy grail for fitness enthusiasts. Why would you need a gps device for a run or bike ride if any of this is true? Can anyone cite a single source for this, especially Tenagashima who claims to possess the knowledge. Where are the empirical results?

 

[Julien]

....Can anyone cite a single source for this, especially Tenagashima who claims to possess the knowledge. Where are the empirical results?

I can answer this for everyone, including Tenagashima. NO

Now about to head out the door for a 10 miler.

 

[samiznaetekto]

And that would be stupid, because complexity isn't necessarily better than simplicity. Most of the time it's the opposite, but that would take an intelligence to know it.

So, explain the unapologetic -1047 steps to me. I don't have the intelligence to understand how putting -1047 on a scale that shows number of steps is not stupid, lazy, and incompetent.

My theory: during rapid expansion following the success of iPhone, hired a lot of $5/hr Indian programmers. That would perfectly explain the quality of software.

 

[matrix07]

So, explain the unapologetic -1047 steps to me. I don't have the intelligence to understand how putting -1047 on a scale that shows number of steps is not stupid, lazy, and incompetent.

I don't have it. The least I can get is 2 steps. (yeah, some day I'm just too lazy.) That image was from August at least though so I'm not surprised it would have bugs since it's still in a beta state. Did it ever occur to you to check things, or put things into context before posting? Since you know it might make you look..uh.. dumb?
(I don't say Apple OS doesn't have bugs just that complaining beta software has bugs is stupid)

 

[Tanegashima]

Link to your "fact"?

You know, open any location app in a place that doesn't have GPS.

Link to your "proof"?

Do it yourself.

Again, where is your "proof"? I just downloaded and ran AccelerometerSample from Apple on my iPhone, set the update frequency to 100Hz, and bumped, kicked, and drummed like a rock star - it shows on the raw data graph as noisy/responsive data as it can be. I see no signs of LPF. So, show me your proof or stop trolling.

You don't understand what you are doing right, do you have a notion of what 100Hz is?

Where's your "proof"?

Where's yours?

You have zero knowledge of my knowledge of physics, iOS programming, sensor fusion, and real world limitations.

Yes, I have: complete ignorance, think you are more intelligent than what you do. And you think the world revolves around you.

----------

A: You leave you iPhone at home to go for a run. You drive to latitude/longitude 38.749999 -9.125644 (perhaps familiar) to start your run. How does the aWatch acquire the staring latitude and longitude to track you?

Apple watch has a path, it only needs 1 GPS point to position the rest.

 

[Julien]

...Apple watch has a path, it only needs 1 GPS point to position the rest.

I actually make turns when I run. At a minimum you would need a least one more GPS point (because of errors you actually need MANY more) to make even one hard 45º turn (which is not the way anyone runs) and another at the next 45º or stopping point.

Also how do you even get the starting GPS point without your iPhone?

Again you are WRONG. You can NOT get a mapped track (path) without GPS on a consumer device. Please provide even ONE example of your 'secret' tracking technology. Also contact DARPA and cash in.

 

[tagy]

So Tanegashima do you still think the iPhone has an array of accelerometers?

Presumably you disagree with the video I linked to?

Also if you can track position with just an accelerometer you should contact Oculus so they can get rid of the positional tracking camera with the rift, it already has an accelerometer so the camera should not be needed?

It could also be used in quadcopters to enable a position hold mode, which currently I have only seen quadcopters with GPS able to do.

 

[Tanegashima]

So Tanegashima do you still think the iPhone has an array of accelerometers?

array |əˈreɪ|
noun
1 an impressive display or range of a particular type of thing: there is a vast array of literature on the topic | a bewildering array of choices.
2 an ordered arrangement, in particular:
• an arrangement of troops.
• Mathematics an arrangement of quantities or symbols in rows and columns; a matrix.
• Computing an indexed set of related elements.
• Law a list of jurors empaneled.
3 literary elaborate or beautiful clothing: he was clothed in fine array.

The iPhone has 3 accelerometers: x, y and z.

Presumably you disagree with the video I linked to?

Which.

Also if you can track position with just an accelerometer you should contact Oculus so they can get rid of the positional tracking camera with the rift, it already has an accelerometer so the camera should not be needed?

OMG, this is so stupid. Can't you track position with 3 accelerometers on x, y and z? If you don't understand this, you better go back to elementary school.

It could also be used in quadcopters to enable a position hold mode, which currently I have only seen quadcopters with GPS able to do.

So... before GPS, quadcopters couldn't hold a position in air?

----------

I actually make turns when I run. At a minimum you would need a least one more GPS point (because of errors you actually need MANY more) to make even one hard 45º turn (which is not the way anyone runs) and another at the next 45º or stopping point.

You only need 3 axis for geopositioning: x, y and z. Rotation is not important for geopositioning. doesn't matter if whether or not you make turns, if the device turns with you. If the device doesn't turn solidarity with you, it can compensate with gyro data.

The fact that you write this

(because of errors you actually need MANY more)

Says a lot of your understanding of analytic geometry.

Also how do you even get the starting GPS point without your iPhone?

As said by Apple, the starting GPS point is uploaded by the iPhone.

Again you are WRONG. You can NOT get a mapped track (path) without GPS on a consumer device. Please provide even ONE example of your 'secret' tracking technology. Also contact DARPA and cash in.

So... I have a path, and one point of that path has a GPS position, and all the points on the path can obviously calculated in relative to that point with GPS position. Can't I calculate all points positions?

It's not my problem you can't figure this out.

 

[tagy]

An array of accelerometers to me would be several at different points on the device, not one 3-axis accelerometer, but ok.

Which.

This talk matches my experience: YouTube

OMG, this is so stupid. Can't you track position with 3 accelerometers on x, y and z? If you don't understand this, you better go back to elementary school.

You can.. I have done it.. Accuracy was terrible. So you saying Oculus don't need the camera? You should let them know.

So... before GPS, quadcopters couldn't hold a position in air?

That is correct, unless you can show me one that can?

 

[Julien]

As said by Apple, the starting GPS point is uploaded by the iPhone.




So... I have a path, and one point of that path has a GPS position, and all the points on the path can obviously calculated in relative to that point with GPS position. Can't I calculate all points positions?

It's not my problem you can't figure this out.

You can NOT track geo position using 2 consumer grade accelerometers even with an accurate stating point. The aWatch is attached to your wrist and in running it is moving dynamically through all 3 planes (sagittal, frontal and transverse). Also the accelerometers are effected by acceleration/deceleration (speeding up/slowing down), angular momentum (turning) and vertical oscillation (just to name a few). All of these movements would add errors in the calculations from the accelerometers. The errors would multiply exponentially and without GPS there would be no way to correct the track.

Below is an actual run I did Saturday. Even using the highest quality military grade (multi million dollar) accelerometers (fixed mounted and not swinging dynamically on a wrist) connected to a super computer the track would be far different form my actual recored GPS run. It would likely be off by a wide margin in the 1st mile. There is NO WAY it would end anywhere near the starting point. It would probably be more than a mile (radius) off.

And you are saying a $350 wrist mounted device can do this???

It can't be done with todays technologies.


Please provide PROOF of this nonsense with just one example of a consumer device that offers map tracking data without using GPS.

 

[MartinAppleGuy]

How good will The Apple Watch be without an iPhone

How useful would the Apple Watch be without an iPhone? Would it still be useful if you have an iPad? I'm guessing basic aspects such as the Clock and Reminders could work with an iPhone? Could the Apple Watch work without an iPhone but connect to WiFi? With full-fledged watch apps coming after release, would a second gen Apple Watch be more stand alone friendly?

 

[Julien]

How useful would the Apple Watch be without an iPhone? Would it still be useful if you have an iPad? I'm guessing basic aspects such as the Clock and Reminders could work with an iPhone? Could the Apple Watch work without an iPhone but connect to WiFi? With full-fledged watch apps coming after release, would a second gen Apple Watch be more stand alone friendly?

The aWatch can't connect to a WiFi. It must do all of it's internet communication through your iPhone. Not sure if you can run the aWatch master app on an iPad.

You can do many things without having the iPhone on you. Here is an example.

You go for a run without your iPhone (I would never do it this way but it's your choice). You can listen to music while motoring your HR. You can check your approximate distance, calories burned, steps taken, time of activity, time of day. You can set a timer or alarm. You can then stop by a grocery store use Apple Pay and a Loyalty Card. You can show pictures of you cat/dog/significant other. You can have all of this and more without having your iPhone with you.

 

[MartinAppleGuy]

The aWatch can't connect to a WiFi. It must do all of it's internet communication through your iPhone. Not sure if you can run the aWatch master app on an iPad.

You can do many things without having the iPhone on you. Here is an example.

You go for a run without your iPhone (I would never do it this way but it's your choice). You can listen to music while motoring your HR. You can check your approximate distance, calories burned, steps taken, time of activity, time of day. You can set a timer or alarm. You can then stop by a grocery store use Apple Pay and a Loyalty Card. You can show pictures of you cat/dog/significant other. You can have all of this and more without having your iPhone with you.

What about not having the iPhone at all though?

 

[Julien]

What about not having the iPhone at all though?

May not work at all since initial setup may be through the aWatch control App. At best you would have just a watch. You might could see your current HR, steps and some other health metrics, but you couldn't record them.

 

[Tanegashima]

You can.. I have done it.. Accuracy was terrible. So you saying Oculus don't need the camera? You should let them know.

If the device is fixed in rotation, or has a way to know it's rotation (gyro, wheel speed on cars), yes, you can know the path, with the greatest accuracy possible, even more accurate than GPS, specially civilian's GPS.




[quteo]That is correct, unless you can show me one that can?[/QUOTE]

I'm not into quadcopters, but obviously there are flying machines that can stay in place without GPS, and were invented before GPS even...

----------

You can NOT track geo position using 2 consumer grade accelerometers even with an accurate stating point.

Proof?

No, you don't have it, so shut up.

Please provide PROOF of this nonsense with just one example of a consumer device that offers map tracking data without using GPS. [/url]

Can you provide a proof that it doesn't, first?