Forum: General Topics

Forums / General Topics / Calculating calories!

 

Subject:Calculating calories! 

diego

9:35
Thursday
19-Jun-2008

Yes, yes. I know that's a silly addon. But there's at least a couple of free gps java software out there that show calories consumption.

One of them, 'WalkMe', is even specifically designed for it. The other, 'GPSWatch', allows setting target calories, and it then returns calories burnt, time elapsed, and remaining time and distance. Wonderful, isn't it?

What does it take?

You can take a look at this:

Total Calories Burned = Duration in Minutes x ((Metabolic Equivalent MET x 3.5 x weight in kg)/200).

But as MET depends on the speed, it needs to be added up during the walk. Slope should also be taken into account as a factor.

Want to give it a try?
 

Stephen

16:19
Thursday
19-Jun-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

Hi Diego,

I've thought about adding something like this, though I'm not quite sure how the maths works out. What would need to be predefined in the phone - just the persons weight? Any how is the MET then calculated from the speed?

I could easily add this if theres a formula that can calculate the energy burned using the values from the TEXT/TRACK page, eg Avg Speed, Distance, Ascent, Duration, etc...

Any ideas?

Cheers,
Stephen
 

ChrisM

11:13
Friday
20-Jun-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

Hi Diego,

Great minds think alike! I was going to post almost exactly the same thing. My wife is wanting to lose a few lbs, and I thought being able to see how many calories she was burning while she was walking the dog etc. would be a great incentive...

Stephen, Ihad a quick shufty round the internet, there doesn't seem to be that many places that give you a formula, but I did come across this:
http://workout.ifit.com/v2/FAQ.html#22

Think it's something to do with some electronic treadmill, but I guess the formula could apply to general walking/running...

(as for cycling, now there IS a challange!!)

Cheers,

Chris.
 

ChrisM

11:41
Friday
20-Jun-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

Hmm, this is quite an interesting subject.
Typing 'walking calories burned formula' into Google comes up with all sorts of stuff, including this:
http://www.runnersworld.com/article/0,7120,s6-242-304-311-8402-0,00.html

The one thing I AM getting from this though is that is is almost impossible to get anything more that a very very rough estimate of calorie burn based simply on bodyweight, distance, speed and incline.

Even so, a basic calorie burn calculator WOULD be quite fun, if nothing else...

Chris.
 

ChrisM

11:48
Friday
20-Jun-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

One more link, then I'll stop. At least for now...
(Really should be working anyway!!)

http://www.wv-hsta.org/cdc_chc/walking_mets_table.htm

Seems like MET is not something that can be easily calculated...

Chris.
 

diego

12:48
Friday
20-Jun-2008

I found an excellent work on Energetics of Bipedal Running at http://jeb.biologists.org/cgi/reprint/201/19/2745.pdf.

To put it short,

Metabolic rate in running = cost coefficient * body weight * speed / lenght step

in units:

Watts = Jules/Newton * Newtons * (meters/second) / meters

Human sample used in the study shows a cost coefficient of 0.262 J/N, which looks quite stable with speed, so you can consider it to be a fixed coefficient. But your program should ask for personal mass, and compute weight of it, and also lenght step. A routine to assess lenght step could be provided, as with WalkMe. In that program, user is required to start walking while pressing a key for each step for a minimum of 100 steps.

So, lets assume a guy with a mass of 78.88 kg, that is weighting 773.1 N, and whose lenght step is 0.75 meters. At 2 m/s, his energy requirement will be 270 Watt each second. After one hour, he would have consumed 1 944 501 joules or 465 kcal.

Alternate simplified method
You can find around java calculators embedded into websites that perform the following calculation:

energy consumed = weight * met / speed * distance

units

kcal = kgs * met / (km/s) * km

where met depends on the speed according to the met column in the following lookup table:

speed met Simplfied Experimental
=========================
3.2   2.0   158     
3.2   2.5   197      207
4.0   3.0   237      258
4.8   3.5   276      310
6.4   5.0   394      413
7.2   6.3   497      465
8.0   8.0   631      516

If you cross check both simplified and experimental methods for the same speed range, simplified results look a bit asymptotic above 6 km/h, while experimental derivated method is linear.

I do not know the impact of slope in this calculations, but one can assume that metabolic rate increases with slope, something like its tangent. I can investigate a little further for you on that, but bare with me, I am only an economist

I hope that has been of any help.








 

diego

12:55
Friday
20-Jun-2008

I am checking the pages you submited. May thanks.

And sorry, I forgot to mention before that last table expresses kcal/hour, and also speed it's related to km/h, not second!
 

Stephen

21:38
Friday
20-Jun-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

Thanks guys for the info/links. It must have been hilarious for the people doing the biped calorie experiments trying to get their avian subjects running on the treadmill. Imagine if they started with a quail and found it a bit uncooperative, they must have been dreading getting the emu on the ramp! I suppose at least it couldn't cheat by flying!! ;-)

So here's where I'm at...

There seem to be a number of different way of calculating the energy used:

1. Metabolic rate in running (power) = cost coefficient * body weight * speed / length step
2. energy consumed = weight * met / speed * distance
3. kcal = MET value of activity x body weight in kg x time in hours

Method 1 is specific to running, ideally I'd like to be able to use a more general formula. 2 and 3 look to be pretty much the same thing (since distance/speed = time), though the units may differ.

I've been looking at this document (found through one of your links):
http://prevention.sph.sc.edu/tools/docs/documents_compendium.pdf

This gives quite a comprehensive list of MET values. Almost too comprehensive actually - I can honestly say I've never ever sat back and wondered how much energy one might use in Digging For Worms (With Shovel), though its fascinating to know that Butchering Animals is more than twice as energetic as just Feeding them. Much more friendly as well!

So, in practice, how about if for every segment of a Track I get TMJ to calculate:
kcal = MET x body weight (kg) x time (hrs)

It would calculate and display this for three activites - Walking, Running and Cycling (leaving it up to the user to decide which activity is closest to what they were doing). The MET value would be looked up from the table in the above PDF document, extrapolated from the average speed in that segment, eg Cycling 16-19mph = 12.0 METs.

It could then add up the total kcal used in each track section.

I'm not sure what to do about including altitude in this equation. GPS doesn't give a particularly accurate altitude reading, but perhaps I could use Newtons good old 'potential energy = m x g x h' formula, eg if you 'rise' by 10m you'd need to convert (at least) mass x 9.81 x 10 Joules of stored energy to achieve this. But then what happens if you're going downhill? It has a big effect on cycling (you effectively reclaim energy), but walking still requires energy input... Perhaps I'll just ignore altitude altogether - its only a rough estimate after all...?

Its also got me thinking about other forms of transport. Slightly off-topic, but how about a 'grams of CO2 used whilst driving' readout, or something similar!? Theres probably some equation to estimate car carbon usage vs speed... Any thoughts?

Cheers!
Stephen
 

ChrisM

0:03
Saturday
21-Jun-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

I think I like the way you are going with the calorie calculator. A way of selecting if you were walking/running or cycling in each section, and showing the calculation using the most appropriate MET value, or even just (as you said) showing a value for each of the above, and leaving it to the user to read the value the want.
As far as alititude/incline goes, if I were you, I'd be tempted to ignore it. Certainly with my GPS unit, the alititude is (relatively speaking) all over the place, and I think trying to take it into account would probably make the calculations less accurate than if you just leave it out.

CO2 usage is a tricky one, as the CO2 emissions vary hugely from one car to another, and there are so many other factors (eg a car driven at 40 mph in 5th gear is , I'd imagine, generating less Co2 than the same car being thrashed along at 40 in 3rd). Having said that, a very rough estimate might be fun, and even a bit thought provoking...

Interesting stuff. Will give it all some more thought when I'm less tired/drunk...
:-)

Chris.
 

diego

6:47
Saturday
21-Jun-2008

Oh, honestly it would be a pity not to use speed, since walking from 2 mph to 4 mph doubles the calories consumption, and it's a data you directly get from the MNEA.

Regarding slope, it is also very important, crucial I may say, to trekkers and mountainers. But we can think of it a little more. Maybe adding it later.
 

diego

6:58
Saturday
21-Jun-2008

Ops, also on biclycing... I would not consider it. How much energy you waste on a cycle depends on many factors. Cycle weight, gears you use, etc.

All GPS java programs focus on walking/running (and hopping!)
 

diego

7:19
Saturday
21-Jun-2008

How do others make it?

http://fitness2me.tk/ uses type of activity, and gives vocal clues
http://www.i10n.com/gpswatchplusfeatures Scroll down to 'Are you a calorie conscious exerciser ?'
http://www.sportsdo.net/ commercial
http://www.getjar.com/products/15419/WalkMe, the simplest

Now, the power of TMJ is its superior design.

I can't wait to put TMJ on top of this list! ;)
 

Stephen

15:40
Sunday
22-Jun-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

" it would be a pity not to use speed, since walking from 2 mph to 4 mph doubles the calories consumption, and it's a data you directly get from the MNEA. "

Dont forget that the speed is still very much taken into account since it is used to derive the MET figure. In your equation:
energy consumed = weight * met / speed * distance

(which for clarity should probably be written as: (weight * met * distance) / speed  )

Since distance/speed=time, this is equivalent to:
energy consumed = weight * met * time

I did a little experiment yesterday. I've added 3 items to the TRACK readout screen: Energy Walking, Running and Cycling (had to split it into two separate pages to fit it all in). It gets each figure by adding up the energy calculated for each track segment (ie the 'gap' between subsequent trackpoints). The MET is interpolated from a lookup table, which is defined separately for each activity. For example, for walking I've used the following table:

Spd m/s : MET
0.00000 : 1.0
1.11760 : 3.0
1.34112 : 3.3
1.56464 : 3.8
1.78816 : 5.0
2.01168 : 6.3

(Speed is stored internally in metres/sec hence its use in the table, in mph the figures are much rounder numbers - 0.0, 2.5, 3.0, 3.5, 4.0 & 4.5mph respectively).

So, lets say that two trackpoints are 40m apart with a 20sec gap, the average speed is therefore 2m/s making the interpolated MET about 6.2321. Energy consumed for a 75kg person would be:
75 x 6.2321 x 20/3600 = 2.6 kcal.
(the 3600 figure is needed to convert seconds into hours).

Does this sound about right?

I've also attempted to include gradient in this as well, though theres a lot of guesswork here, so any ideas would be welcome. Essentially I've got it to firstly work out the gradient of the segment and then to modify the MET depending on the activity, the amount of slope, and whether the slope is ascending or descending. The formula I've come up with is:
NewMET = (1 + c * gradient) * OriginalMet

The coefficient c is a constant chosen from the type of activity and the type of slope (ascent or descent). The coefficients that I've chosen are:

Walking Uphill = 0.5
Walking Downhill = 0.2
Running Uphill = 1.0
Running Downhill = 0.75
Cycling Uphill = 1.0
Cycling Downhill = 4.0

So lets say you're walking up a 1 in 10 gradient. The modified MET would be 1+0.5*0.1=1.05 times the original MET (ie 5% more calorie burn). Walking down the same hill would be 1+0.2*-0.1=0.98 (ie 2% less calorie burn). I figure that walking downhill isn't actually *that* much easier than walking on the flat, whereas for cycling I've assumed a much bigger effect. eg: Cycling down the same hill would be 1+4.0*-0.1=0.60 (40% less calories). If it was a really steep hill, say 1 in 5 then the calorie burn would work out at 80% less calories.

I'll make all these gradient calculations an option, but do you think they're on the right track? Diego, I agree that Cycling is perhaps a difficult one to accurately calculate, but the fact that there are published METs for cycling as well as walking and running suggests that I might as well include it, if only to give a rough comparison between separate bike rides - ie, it might not be a true calculation of exact calories, but could still help someone training to try and achieve a similar or increasing workout. I couldn't find any MET figures for hopping unfortunately ;-)

I was up in London yesterday, took a round-about walk to the pub. TMJ recorded 2.5 miles, average speed 3.5mph, duration 44 minutes, Calories used 247kcal. I left it a bit late leaving the pub and ended up running part of the way back to the station (figured it was a good experiment too!!). Recorded 2.5miles, avg spd 6.2mph, duration 24minutes, Calories used 321kcal. Do these figures sound about right?

Chris, as you say, the CO2 calculations would be very much ball-park figures, I've been trying to find some examples, with little success. The graph on this page is about the best I could find. It seems that CO2 emmisions are pretty much directly proportional to the amount of fuel burnt, so I was hoping to just be able to estimate the fuel consumption at different speeds (probably with an option to set the cars engine size, etc), but perhaps theres a bit too much guesswork in this!!

Cheers!
Stephen
 

ChrisM

11:26
Tuesday
24-Jun-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

If you google for something like 'Calories burned walking 1 mile' you get all sorts of numbers, but they seem to put your 247 calories for 2.5 miles in the right sort of ballpark. One site reckoned 100 cal/mile at 3mph, so that is pretty close. Another said 58-64/mile, but that was for a woman weighing 117lbs. You did say you are taking weight into account? If so, as long as you weigh more than 8 stone, it puts your figures pretty much on track...
Not sure about the gradient calculations, I know it should be taken into account for a more accurate estimate, but I think you should only work out an average gradient every few minutes or something, otherwise, GPS errors are going to screw the figures quite a lot. Certainly with my GPS, even if I'm walking along a fairly flat field, the altitute can jump up and down by 30-40 feet, and I can sometimes get some quite steep gradient figures in the display.

Cheers,

Chris.
 

Stephen

9:38
Wednesday
25-Jun-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

Hi Chris,

I forgot to mention that the TMJ figures were calculated for an 'average' weight of 75kg.

Regarding the gradient, all of the energy calculations are performed on the stored track, rather than the live stream (ie, the Calories reading only increases each time a new trackpoint is added to the track). This is a lot more accurate than the gradient gauges, since these show the difference between subsequent points coming from the GPS, which are only a second apart and can be fairly 'noisy', whereas the stored track smooths this out somewhat.

How does your altitude graph look (screen 9) - is this also very noisy or does it give a better representation of the altitude profile (you'll need to set the x-axis to Distance and then zoom in a lot to get a 1:1 view of the track profile). It is this data in effect which is used for the Energy calculations. Do you think this is sufficient?

Cheers,
Stephen
 

diego

13:15
Wednesday
25-Jun-2008

Interpolating figures from activity based MET tables may probably be as accurate (or inaccurate) as using a linear formula based on a coefficient, so I have no objections here.

I just thought using a formula would make your code shorter.

In any case, it would probably be wise to restrict energy calculations around the tabulated speed range, to avoid nonsense results when you drive or fly your gps enabled mobile!

Regarding the gradient, I also favour using stored tracks to drop incoherent results.

All in all, I am confident TMJ will meet a good solution to bring users a sophisticated and meaningful energy consumption estimate tool, probably better than most other software.


 

diego

13:31
Wednesday
25-Jun-2008

Where to put all this energy calculations in the interface?

You may get weight, activity and energy comsuption target from menu point '7. Settings'.

A new screen under key 1 may reflect relevant energy statistics. I am not sure whether it could be good to allow users to change the 3 parameters, or just some of them, here directly instead of accessing the menu.

The profile plots under key 9 may add a new graph to reflect the energy evolution with time or distance. This graph would only be displayed whenever an energy calculations option has been selected in '7. Settings'.

Just ideas.
 

Stephen

12:17
Sunday
29-Jun-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

Hi Diego,

For the moment I'm just adding the energy values to the text page - I guess I could add an option to set which type of exercise should be displayed, but I figure it might as well just show all types together on one page. If a particular type is 'invalid' due to an impossible track speed (eg walking at 50mph!) then it just gives a 'n/a' (not applicable) readout.

I haven't yet looked into adding a calorie graph, this will have to wait for the next version - I'll need to make a number of changes to enable this...

I'm hoping to have a new version ready in a week or two, the calorie calculations can then be refined further if necessary...

Cheers,
Stephen
 

ChrisM

8:16
Monday
30-Jun-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

Stephen,

Regarding your question about my altitude graph earlier in this thread, I did a little test yesterday. Jogged and walked round a very short circular walk (only about 3/4 mile). The general trend of the graph is ok, but it seems to have a lot of (for example) ups and downs, where the ground was only sloping down. Also seems to have more varience than it should. Interestingly, there was a lot more variance when jogging than walking...
You said that you get a value from the GPS unit once/second or so, but only write a point to the track every 20s (or whatever has been set in the config). When you write the altitude, do you just write whatever the last value from the GPS was when it's time to write a trackpoint? If so, how would it work if you calculated an average of all the values between each trackpoint, then wrote that average to the track??
 

Stephen

21:02
Monday
30-Jun-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

Hi Chris,

Funny you should say this - I was thinking the same thing yesterday and have already added a Settings option to 'Apply Extra Altitude Smoothing'. This effectively does exactly what you've suggested - the altitude graph now looks a lot smoother, with quite a significant reduction in Total Ascent and Total Descent values!

I've set it as an option, just in case people want a more accurate (but also more spikey) graph...

Cheers,
Stephen
 

ChrisM

21:47
Monday
30-Jun-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

Altitude smoothing sounds like a good option!
To be honest, I think that'll be better. Not sure if the more spikey graph is truly more accurate anyway, as it picks up a lot of 'interference'. Unless you are travelling through countryside that is extremely steep, or if the track-points are a very long distance apart, the 'smoothed' version would probably be closer to the truth.

Probably is still best to have it as an option like you say though. I suppose if you were hill walking in the desert or something then the non-smoothed graph would definitely be more accurate...
:-))

Chris.
 

diego

8:46
Tuesday
1-Jul-2008

On altitude errors, be aware of antenna sensitivity.

Pointing your devide's antenna downwards to the ground will maximize multipath errors that will be specially significant in calculating altitude for an energy comsuption measurement.
 

ChrisM

11:41
Tuesday
1-Jul-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

Ahh, that's an interesting point. I didn't really think about that. I suppose if I considered it at all, I was kind of thinking the aerial was omni-directional, if it isn't then maybe that's a part of the problem!
I usually have my GPS 'mouse' stuck in the back pocket of my jeans when I'm out and about. Which probably means the aerial is at 90% to the ground (and possibly facing towards me).
Maybe I'll see if it's possible to keep it with the aerial pointing upwards, and see if it makes a difference...
 

diego

13:36
Wednesday
2-Jul-2008

Definitely.

Just keep looking at the Satellite screen along various antenna attitudes, and you'll see a dramatic drop in SNR whenever you let your gps drop pointing to the ground. Also, look the effect on your position on the map.
 

ChrisM

14:19
Wednesday
2-Jul-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

Hmm,

Just drawing two threads together a bit, it occurs to me that there are two values - HDOP and VDOP, which I understand to be some sort of measurement of the 'accuracy' of the GPS fix.
With respect to the problem discussed both here and elsewhere of getting rid of inaccurate trackpoints that are messing up distance travelled/average speed/calories consumed etc. would a filter to discard any points with HDOP/VDOP that were above a configurable threshold help to solve the problem??

Chris.
 

Stephen

16:10
Wednesday
2-Jul-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

A friend of mine sometimes puts his GPS on his head (under a hat) in order to maximise the reception quality - not too sure what effect the bluetooth will be having on his brain though!!

Chris, TMJ does in fact already look at the HDOP value to try to estimate how 'trustworthy' the current fix is. The higher the HDOP (it goes up to 50), the worse the signal accuracy is, so TMJ applies more 'smoothing' than at lower values. Although this helps slightly its not actually hugely useful - most 'spikes' seem to be caused *before* the GPS realises that the signal quality has dropped (with my GPS device at least). For example, you're walking along outside with a clear view of the sky and an HDOP of, say 0.9. If you then walk into a tunnel the location might quite suddenly jump quite a distance, but still be reported as HDOP 0.9 for a couple of seconds, before eventually increasing - what can TMJ do in this case?!

Overall, all of these 'smoothing' algorithms are in a way a bit like 'closing the stable door after the horse has bolted'! After all, if the GPS device can't work out a reliable location (especially considering all the money/research/technology etc that has led to its development, and the fact that it has direct access to the raw satellite signals), then I don't think theres going to be an easy way for TMJ to substantially improve it!!

Incidentally, my first attempt at using the speed 'moving average' to filter the spikes wasn't too successful - it would either filter too many points, or completely miss the important ones. I'm still looking into it, but probably wont have it ready for the forthcoming release. The altitude smoothing works surprisingly well though, and the 'Pause Track Recording' function does make it easier to manually pause the track when going into an area of known poor reception, so its not all bad news!

Cheers,
Stephen

 

ChrisM

8:48
Thursday
3-Jul-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

Hi Stephen,

I think I see the problem you have. You're right, there probably isn't a simple solution (not sure if there's even a difficult solution!).

Like any system, TMJ can only really be as good as the data it's given. If the GPS is feeding it rubbish, then there is only so much you can do to filter that out. It's the old acronymn GIGO (Garbage In, Garbage Out) to some extent - Don't get me wrong, I'm not critisising you or TMJ, it sounds like you have already given a lot of thought about ways of improving things, and there is only so much that can be done...

On the brighter side, good to hear that the altitute smoothing and the pause is improving things!

Keep up the good work!!

Cheers,

Chris.
 

diego

14:06
Monday
14-Jul-2008

Just to bring the thread back to the calories issue, I am not happy with the current solution, but I have tried to test its wellness to support my opinion, and I have found my TMJ resets the track to zero after a while once and again, so I can not get value for a walk of, let's say, 50 minutes long. Anyone else has experienced that?

Also, where do we enter the weight?



"Calories are calculated using the formula: kcal = MET * weight(kg) * time(hrs), where MET is the 'average' metabolic rate (per kg) burned when performing the associated activity at the calculated speed. (So running faster has a higher MET value, and hence higher calorie burn per second)." TMJ user's guide.
 

Stephen

18:46
Monday
14-Jul-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

Hi Diego,

I'm not sure why you're finding that TMJ resets the track all by itself. Have you perhaps got a low value set in Menu/Settings/Limit Track Size to ... Points? When the track becomes bigger than this value it is automatically cropped to conserve memory. Just set this to something really big to effectively disable it, or use a larger value for Settings/Min Time Betweeen Points or Settings/Min Time Betweeen Points so that the track doesn't get big so quickly.

If this isn't the issue then maybe there's a bug that hasn't shown itself before, or only on your handset for some reason - has anyone else noticed this problem?

The weight is entered in Menu/Settings/Users Weight.

Other than the disappearing track issue, what is it that you're not happy about with the calorie calculations?

Cheers,
Stephen
 

sentinal

21:12
Monday
14-Jul-2008

Location:
UK

Phone Model:
-

Hi Diego,

I've got a Sony Ericson W810i. I was out on my bike for 5 hours Sunday and 2:30 hours today (you know, tour de France and all that :) I didn't get any calorie data resets.

Hi Stephen,

I may have missed it while reading the posts but does the calorie information take into consideration altitude (extra smoothing option has made a big difference)? I can understand some users may not want this feature if the gps can't be located in a good position.

Cheers,

David


 

diego

13:46
Tuesday
15-Jul-2008

Hello.

After your recommendation, I set 9999 points as maximum track size and took the phone for a walk.

After 6.42 km at 6.0 km/h average speed, TMJ showed an energy consumption of 200 kCal walking for 1:04.32 hours.

Using the formula built into TMJ:

   MET * weight(kg) * time(hrs), = kCal;
   4.65 kCal/hr * 40 kg * 1.076 hrs = 200 kCal

Using speed dependant MET calculations, as you can get in websites, you get 201 kCal for an speed of 6.4 km/h, which seems pretty good approach.

But same time walking at half the speed should yield half the energy, unlike TMJ's calculations, that will keep showing 200 kCal.

Don't you think this is too simplistic?


 

diego

13:50
Tuesday
15-Jul-2008

Well, Stephen, I am sorry I overlooked your test trips to the pub!

Then you are interpolating MET values from speed tables? Ok, then.

I must confess I didn't get to test the same track at different speeds as you did, and only believed what you declared in your manual on the formula used. That is, that you were not considering speed...

 

Stephen

17:54
Tuesday
15-Jul-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

Hi Diego,

No worries - I might not have explained it the best way in the manual but the " performing the associated activity at the calculated speed " bit was meant to imply that the MET *is* interpolated from a MET vs Speed table (as given in an earlier post). And don't forget that this interpolated MET is calculated separately for each tiny track segment, so I'd have thought that this would give a more accurate estimate than just using the tracks total time, average speed, etc (as the online calculators probably do).

Therefore " But same time walking at half the speed should yield half the energy, unlike TMJ's calculations, that will keep showing 200 kCal. " isn't the case since the MET used by TMJ will be less due to the lower speed (though not necessarily half, as by interpolating from 'actual' MET values I figure that it would actually be more accurate - ie not necessarily half the speed, half the calories, etc).

David, yes you're right - I had meant to include a Settings tickbox to enable/disable the use of altitude in the calorie calculation (it is currently always enabled) - I'll fix this for the next version.

Cheers,
Stephen
 

shchenka

16:10
Sunday
20-Jul-2008

Hi all,
I've played with the calories calculator today. I took a stroll but before that I carefully cleared all the track sections. How surprised I was at the end, when I looked at the screen where the calories are displayed to realize that I walked with a max speed of 83..6 km/h. This brought me to an idea, perhaps a silly one.
Stephen, how about a "walking mode" that could be manually turned on when walking only and in which every distance covered with a speed above a certain pre-set value would not be counted for calories calculator. And maybe ignored in the track, too. In other words, tn the walking mode, TMJ would not record any track when moving faster than the maximum walking speed. Just an idea as I said.
 

diego

12:08
Monday
21-Jul-2008

Yes, I second the idea.

At low speeds, position errors can exceed the distance between points and yield to absurd calculated speeds and headings.
 

shchenka

8:31
Tuesday
22-Jul-2008

" At low speeds, position errors can exceed the distance between points and yield to absurd calculated speeds and headings "
I think this is especially true for the receivers with MTK chipsets, which have the tendency to 'float' compared to the SiRF III based ones.
 

Stephen

10:07
Tuesday
22-Jul-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

I had originally decided not to add a 'speed-limiter' for track recording (to avoid a situation where a track is accidentally not recorded at all), but I guess if it is completely configurable then it can always be disabled if not wanted!

I'm thinking now of adding an 'Activity Profile' option, which would allow custom settings to be stored for different activities (probably just a small number, perhaps Walking, Running, Cycling, Driving, Sailing, Flying). Each profile would have independent settings for Max Track Speed, Min Time Between Points, Min Distance Between Points and Start New Section After ... Seconds. Would this be of use do you think?

Cheers,
Stephen
 

diego

20:27
Tuesday
22-Jul-2008

Activity Profile sounds great. It may improve vastly your software useability, and it's a truly original idea, since I can't think of a competing product.

By the way, as silly as it can sound, I can't make my TMJ running on a Nokia 6151 to save my weight on 'Settings'. It always returns to 40 kgs.

Any thint?
 

ChrisM

7:57
Wednesday
23-Jul-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

Just my 2p's worth...

Activity profile is a great idea. It will solve one of the problems I have with TMJ which is trying to find a balance between track settings that'll work both when walking and driving.

Chris.
 

shchenka

8:09
Wednesday
23-Jul-2008

" I'm thinking now of adding an 'Activity Profile' option, which would allow custom settings to be stored for different activities (probably just a small number, perhaps Walking, Running, Cycling, Driving, Sailing, Flying). "
... and geocaching ;)
which might require its own settings.
 

ChrisM

8:20
Wednesday
23-Jul-2008

Location:
Bedfordshire, UK

Phone Model:
Sony Ericsson W770i, Blackberry Playbook(?)

Maybe there should be 1 or 2 'Custom' settings for uses that you might not have anticipated...

Chris.
 

Stephen

14:29
Wednesday
23-Jul-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

Yep, there'll be at least a 'General' and 'Custom' option (and each one can be manually altered anyhow, so you could always re-use 'Sailing' for something else if necessary!). To be honest I can't really see much use for having more than a handful of profiles anyhow (as similar activities would generally boil down to the same thing).

Shchenka, perhaps the profile would be a good place to have the 'Maximum Autopan Zoom Level' setting!?

Diego, did you download v0.6.0 very soon after I originally uploaded it? There was a bug which saved the mass from the Screen Height setting - I fixed it fairly quickly so you could try re-downloading again. (If this doesn't work then I might have missed the Nokia version - let me know)...
 

shchenka

20:44
Wednesday
23-Jul-2008

" Shchenka, perhaps the profile would be a good place to have the 'Maximum Autopan Zoom Level' setting!? "
Yes, I believe so. This would also help saving the valuable context menu space.
Probably also the min time and distance between points (quite dense for geocaching mode) could be profile-dependent.

By the way, Garmin has introduced the concept of application modes in its recent receivers, too.
 

diego

17:22
Tuesday
9-Sep-2008

Okay, I must confess I am now very pleased with this implementation of the calculation of energy consumption.

I use it very often, in 'planned walks' of 300 kCal each... why do I do it?

For the free meal, you know!

So thanks so much, you should be living only on that, you have made a quality product, the java client, the web portal, etc... out of almost nothing.

Congratulations once more for all that.
 
 

Stephen

11:11
Wednesday
10-Sep-2008

Location:
Surrey, UK

Phone Model:
BB 9800 Torch
BlackBerry 8900
SE W910i
Nokia 5800

Cheers Diego - glad you find it useful!!
 

(You must be logged in to post a reply to this thread)