Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Longjog @ GS-Sport-Servicecenter]

in that case: i think the first pace1k should be exactly the same as the first 1k autolap and it is not, it is a few seconds faster, i can't think of an explanation for that.

In general i dont understand why globalsat is making it so difficult, I would go for a simple time based algortihm, just measure the distance for a minute and calculate the average speed over the last minute. this can't be that different from the way how the current speed is calculated only the time interval is longer. perhaps the time interval can be user defined, so the user can choose between reliability and response time

Hi remcoh,
I agree with you, but only one minute as time-base will cause an error much more than 10 %.
Remember the first idea of pace1k. The reason why we would have it was, because the actual pace has an error which is caused by the natural inaccuracy of GPS. To reduce  this error it is nessesarry to take a longer distance for calculation base. I calculated the error will be less than 5% when the distance is around 700 m, so I suggested to take 1000 m as calculation base. This is the reason why I named it pace1k.
It should be the pace of the actual last 1000m. (with "actualy" I mean the pace 1k is updated not only every 1000m but every some seconds like lap-pace). And I really thought like you, dear remcoh, that this is possible in an easy way like the calculation of the AVG-pace for example. I do not really understand why it should be nessesary to build 100 m parts.
Maybe the easiest way to do it, is to modify the actual pace algorythm by setting a longer time-base (around 4 min) or distance-base (> 700m).

Regards

Longjog

[Hubert]

Hi Longjog,

I agree with you: "real distance could be much less than 100 m the watch shows, when GSP-signal is weak (4-5 satelites)."

Hi Hubert,

real distance could be much less than 100 m the watch shows, when GSP-signal is weak (4-5 satelites).
Take a look
...
I runned Angelstraße - Angelgartenstraße - with 5 min/km but the watch showed a full 1k-lap after around 650 m with a lappace around 3:20 min/km
This example is really very extreme but you see the GPS-failure can cause shorter real distances as the watch showes.

Sincerely, Longjog

Of course, you are right ! 

But we are talking about different problems:
* What you are describing, is a GPS error, caused by weak GPS signals.  This can be positive or negative. It can be
   a big error or a small one. You never know...
* When I was talking about GPS errors, I was talking about the "normal" inaccuracy, which is caused by the maximum
   resolution provided by the GPS system. And this is (worst case) +/- 15m.

So - everything I wrote about optimization of algorithms was based on "good" GPS signals (with "normal" errors).
It does not apply to situations with GPS errors caused by weak signals.

Regards, Hubert

[Hubert]

Hi Albert !

What are you going to tell me ??

Hello Hubert,

When you run for 1 hour with 4 Min/kmand then one more 1 hour with 6 Min/kmthe total (average) pace is not 5 Min/km !

You made an example here, but you don't follow the basic mathematic rules.I agree with what you want to achieve, your way is much better.The right approach is the following in order to do:  The rule is X², *, \ etc.result:2*60\(4+6)=1212 km/hour = 5 min/km by the way: Geoffrey has had the flu.Best regards,Albert[/]

I don't follow the basic mathematic rules ?  There is no rule in the red colored sentence ?! I just said, that it's wrong !
I didn't tell any rule or any method for the correct calculation. 

And what are you going to tell with this strange calculation: result:2*60\(4+6)=1212 km/hour = 5 min/km ?
Do you want to say that the average pace in my example is really 5 Min/km ?

I'm sorry, but I'm a little bit tired of explaining simple mathematics. Please just believe me: It's not 5 Min/km !

Please take some more time and recalculate. I'm sure, you will also find that I'm right.

Regards, Hubert

[calvin]

1 hour at 4min/km pace = 15km in 60min
1 hour at 6min/km pace = 10km in 60min

The overall: 25km in 120min
The average pace: 120min / 25km = 4:48/km

[palisander]

Hello from Spain.
Excuse my poor english.
I have downloaded the firmware version F-GGH-2A-0906111 from the your    New Athletic Product Website.
Is it different of the version FW: F-GGH-2A-0905191.GHX available on the global website and US website?
Also, I wolud like to try firmware X
Thanks

[Hubert]

Hi Calvin !

Welcome to the forum and thank you for your help !

1 hour at 4min/km pace = 15km in 60min
1 hour at 6min/km pace = 10km in 60min

The overall: 25km in 120min
The average pace: 120min / 25km = 4:48/km

Finally somebody who understands what I'm talking about...

Regards, Hubert

[calvin]

Hello from Spain.
Excuse my poor english.
I have downloaded the firmware version F-GGH-2A-0906111 from the your    New Athletic Product Website.
Is it different of the version FW: F-GGH-2A-0905191.GHX available on the global website and US website?
Also, I wolud like to try firmware X
Thanks

Hello from Portugal!
I would like to know the answer to the same question Palisander posted.

Also I want to give a sugestion: a Changelog listing all the firmware releases till present, with respective version number, date, changes from previous version. Is this possible?

[albert]

Hi Calvin !

Welcome to the forum and thank you for your help !

1 hour at 4min/km pace = 15km in 60min
1 hour at 6min/km pace = 10km in 60min

The overall: 25km in 120min
The average pace: 120min / 25km = 4:48/km

Finally somebody who understands what I'm talking about...

Regards, Hubert

I agree with what you want to achieve, your way is much better.The right approach is the following in order to do:  The rule is X², *, \ etc.result:2*60\(4+6)=1212 km/hour = 5 min/km

Hello Hubert,
Of course you are right, sorry for my mistake! Like I said there are different way’s they look OK but they not. Sorry I did confuse you.
25/2= 12,5 km/hour. To show what I meant is here a little story, it looks right but isn’t.

3 men are deciding to rent a apartment  with 3 bedrooms. The pay the counter boy 100 Euro, then they go to their bedrooms. The boss from the apartment discovers the boy forgot to give 10% discount, so he sends the boy with 10 Euro to the 3 man. The boy thinks I take 4 Euro, because that is better to divide for 3 people. He gives each person 2 Euro back.
So instead of 100 they paid 100-6=94 Euro, can you agree with that?
But with the 4 Euro the boy took it is 94 +4= 98 Euro.
Where did the 2 Euro go? 

Have a nice day,

Best regards,

Albert

[TWG_TECH8]

Hi Calvin !

Welcome to the forum and thank you for your help !

1 hour at 4min/km pace = 15km in 60min
1 hour at 6min/km pace = 10km in 60min

The overall: 25km in 120min
The average pace: 120min / 25km = 4:48/km

Finally somebody who understands what I'm talking about...

Regards, Hubert

I agree with what you want to achieve, your way is much better.The right approach is the following in order to do:  The rule is X², *, \ etc.result:2*60\(4+6)=1212 km/hour = 5 min/km

Hello Hubert,
Of course you are right, sorry for my mistake! Like I said there are different way’s they look OK but they not. Sorry I did confuse you.
25/2= 12,5 km/hour. To show what I meant is here a little story, it looks right but isn’t.

3 men are deciding to rent a apartment  with 3 bedrooms. The pay the counter boy 100 Euro, then they go to their bedrooms. The boss from the apartment discovers the boy forgot to give 10% discount, so he sends the boy with 10 Euro to the 3 man. The boy thinks I take 4 Euro, because that is better to divide for 3 people. He gives each person 2 Euro back.
So instead of 100 they paid 100-6=94 Euro, can you agree with that?
But with the 4 Euro the boy took it is 94 +4= 98 Euro.
Where did the 2 Euro go? 

Have a nice day,

Best regards,

Albert

It seems a interesting logic question.
94 Euro consist of 90 (Boss) and 4 (boy).
The balance should be 2*3 (customers) not 4 Euro!
So the right way is 90+4+6=100
Am I correct?

Sincerely,

Geoffrey

[TWG_TECH8]

Hello from Spain.
Excuse my poor english.
I have downloaded the firmware version F-GGH-2A-0906111 from the your    New Athletic Product Website.
Is it different of the version FW: F-GGH-2A-0905191.GHX available on the global website and US website?
Also, I wolud like to try firmware X
Thanks

Hello from Portugal!
I would like to know the answer to the same question Palisander posted.

Also I want to give a sugestion: a Changelog listing all the firmware releases till present, with respective version number, date, changes from previous version. Is this possible?

Hello Portugal & Spain!
Thanks for interested on our firmware.
The firmware (F-GGH-2A-0905191.GHX ) is the same in US and Taiwain website.
You only can download firmwareX (0906111) from Taiwan website or GS-Sport offical webiste which our total new website for our sport product line. http://www.gs-sport.com/pro_dl.php?en_product_id=1&en_product_cnt_id=23
To put change log is fine, just need time to translate because our internal system is using Chinese.

Sincerely,

Geoffrey
 

[TWG_TECH8]

1 hour at 4min/km pace = 15km in 60min
1 hour at 6min/km pace = 10km in 60min

The overall: 25km in 120min
The average pace: 120min / 25km = 4:48/km

Hi Calvin,

The calculation seems resonable and correct. 

Have a nice day,

Geoffrey

[TWG_TECH8]

Hi Longjog and remcoh !

So - there are really some people out there, having doubts
about the famous Hubert theories ?  Unbelievable ! 

Hi Hubert,

a nice idea, but I think this is not the solution.
How do you explain that the first kilometer the pace1k is smaller than lap-pace?

Regards, Longjog

Okay - I agree, that the first part of my  explanation
was just to explain the basic problems for pace1k.
I didn't want to waste my (and yours) time with the details.

But if you want to...

Okay, let me explain my theorie of the first kilometers:

First, we assume that the algorithm for pace1k basicly works
like follows:

The watch tries to save time intervals for each 100m distance,
stores 10 of those values and adds them to get the pace1k.
After a new pace1k is shown on the display, the 10 values
are shifted by one, so that the next value can be gathered.

This algorithm doesn't work directly after starting the watch.
So it has to be a little different at the beginning. After
starting the watch, only the times needed for the 100m distances
are being saved. Nothing more. This is done 9 times.

Continuing my example of my last posting ( 5:00 Min/km pace
and an average GPS error of let us say: 7m) we get the following:

    1    107m   0:32 sec.
    2    214m   0:32 sec.
    3    321m   0:32 sec.
    4    428m   0:32 sec.
    5    535m   0:32 sec.
    6    642m   0:32 sec.
    7    749m   0:32 sec.
    8    856m   0:32 sec.
    9    963m   0:32 sec.

(Please note: only times are being saved, meters are just to explain!)


Now to start the pace1k algorithm completely, the watch does one
more check: It checks, if 1000m are already reached.
This is done independantly of the 100m distances and will be
reached (again with 7m error) at 1007m.
Time for this is also saved.

The values are now:

    1    107m   0:32 sec.
    2    214m   0:32 sec.
    3    321m   0:32 sec.
    4    428m   0:32 sec.
    5    535m   0:32 sec.
    6    642m   0:32 sec.
    7    749m   0:32 sec.
    8    856m   0:32 sec.
    9    963m   0:32 sec.
   10   1007m   0:13 sec.

Now pace1k values are complete, can be calculated and shown.
The first value for pace1k is:

   pace1k (1.0 km) = 5:01 Min/km

This explains, why there is no 20sec. delay at the beginning !

And - if pace1k is calculated at the beginning of the 2 or 3 sec. update cycle
of the watch and LapPace is calulated at the end (or at the beginning of the next cycle),
LapPace will have 2 or 3 sec. more !
Or - maybe, LapPace has another problem.
Maybe it's calculated by the famous pace = (p1+...+pn) /n formular...
I don't want to think about that. 

Now - let's see how the story continues:

The stored values are shifted and after another 100m we have
the following situation:

    1    214m   0:32 sec.
    2    321m   0:32 sec.
    3    428m   0:32 sec.
    4    535m   0:32 sec.
    5    642m   0:32 sec.
    6    749m   0:32 sec.
    7    856m   0:32 sec.
    8    963m   0:32 sec.
    9   1007m   0:13 sec.
   10   1114m   0:32 sec.

As we can see the pace1k will not change, because one 32 sec.
entry was removed, one new 32 sec. was added. So:

   pace1k (1.1 km) = 5:01 Min/km

This will continue up to the end to 2nd kilometer, when we have:

    1   1114m   0:32 sec.
    2   1221m   0:32 sec.
    3   1328m   0:32 sec.
    4   1435m   0:32 sec.
    5   1542m   0:32 sec.
    6   1649m   0:32 sec.
    7   1756m   0:32 sec.
    8   1863m   0:32 sec.
    9   1970m   0:32 sec.
   10   2077m   0:32 sec.

Here the "shorter" entry disappears, and pace1K will jump for about 20sec. !
The result will be:

   pace1k (2.0 km) = 5:20 Min/km


This is why you see "good" values for pace1k starting at km 1
and continuing up to km 2 and then jump for about 20 seconds.

Okay - now it should be enough of explanations.
And enough postings about this item. I'm waiting for a solution...

Regards, Hubert

Hi Hubert,

Thans so much for providing so detail explanation!
Now, we are thinking about there are 3 suspect reasons:
1. GPS inaccuracy to make every 100m shift likes Hubert said; I think not only the ending point will affect calculation, starting point to get 100m is also need to be considered. Because it is not the absolute 100m starting point.
2. CPU power limition; if the speed is quite fast likes Roger's driving test; it means 00:06 for 100m interval at 60km/hr. Maybe it will be overloading for CPU.
3. The condidtion of GPS signal ;

Today we have done a simple running and cycling test comparing with GH-505 which used more powerful CPU with the same alogrithm and checked with Lap-pace.

1. We found the result is the same, so CPU power won't impact a lot at lower speed. Need to check high speed.

2. first 1K is ok; But after that Lap-pace is different with it. And we found 3090m for Pace1k at 3rd km calculation. So 100m distance shift is very suspect reason.


3. Pace 1k value will be little worse after 2km not very obvious, maybe or GPS inaccuracy (7m) contrast with 100m is too large. considering to 500m will be better.

Tomorrow it still needs a lot of job to do. 

Have a nice dream,

Geoffrey 

[calvin]

I have downloaded the firmware version F-GGH-2A-0906111 from the your New Athletic Product Website.
Is it different of the version FW: F-GGH-2A-0905191 available on the global website and US website?
Also, I wolud like to try firmware X

The firmware (F-GGH-2A-0905191) is the same in US and Taiwain website.
You only can download firmwareX (0906111) from Taiwan website or GS-Sport offical webiste which our total new website for our sport product line.

Ok Geoffrey, now I'm really confused!! 

Palisander asked if F-GGH-2A-0905191 is different from F-GGH-2A-0906111 (my guess is: they are different  ).
Palisander also requests to try the new firmwareX.

In your reply you don't explain the difference in F-GGH-2A-0905191 and F-GGH-2A-0906111.
And to make it worse you call firmwareX to F-GGH-2A-0906111 !!!    (Hey Palisander, you allready have it! )
Where's firmware F-GGH-2A-0910072 then? 

Can you see why I'm confused? Indeed a Changelog is really needed! 
Related to this, how do you name your firmware?

Sorry for so many questions! Regards!

[Hubert]

Hi Calvin,

Yes - you are right ! Geoffrey sometimes likes to confuse us... 

Ok Geoffrey, now I'm really confused!! 

Palisander asked if F-GGH-2A-0905191 is different from F-GGH-2A-0906111 (my guess is: they are different  ).
Palisander also requests to try the new firmwareX.

In your reply you don't explain the difference in F-GGH-2A-0905191 and F-GGH-2A-0906111.
And to make it worse you call firmwareX to F-GGH-2A-0906111 !!!    (Hey Palisander, you allready have it! )
Where's firmware F-GGH-2A-0910072 then? 

Can you see why I'm confused? Indeed a Changelog is really needed! 
Related to this, how do you name your firmware?

Sorry for so many questions! Regards!

And I understand your questions. I think, I had a similar question, when I started to join this forum:
http://www.usglobalsat.com/forum/index.php?topic=906.msg1165#msg1165

My current knowledge is:
* Forget the US web site. Taiwan website provides current versions.
* There are "official" firmware versions, released and downloadable from the website.
   These are numbered like "F-GGH-2A-0905191" where the ending "0905191" tells
   about the date, here: 2009 May 19.
* Additional there are test versions, sometimes called Beta versions. This time it is
   called "X-version".
* Those test versions are not downloadable. You have to ask Geoffrey to send it to you.

As far as I know, this is the current changelog of firmware versions:

04.11.2008   F-GGH-2A-0811041                          Formal release

11.02.2009   F-GGH-2A-0902111 (beta version)      1. Add set lap page
                                                                      2. Smooth Heartrate curve
                                                                      3. Improve cumulate

08.04.2009   F-GGH-2A-0904082 (beta 2 version)    1. Add filter into HR calculation and smooth alogrithm
                                                                      2. Modify threshold of cumulate calculation

13.05.2009   F-GGH-2A-0905132 (beta 3 version)    1. Add filter of HR zone limition at start up
                                                                      2. Add smooth filter to optimize HR curve
                                                                      3. Fix UTM not correct issue
                                                                      4. Fix Calories huge of last Lap problem

19.05.2009   F-GGH-2A-0905191 (beta 3.1 version) 1. Fix HR stick
                                                                      2. Correct Itilan wording"
                                                                      Formal release.

07.10.2009   F-GGH-2A-0910072 (Version X)          1. Add HR% in custom page
                                                                      2. Add HR% Alert
                                                                      3. Add a arrorw at 1/3 display in Track back mode
                                                                      4. Fix cumulative lap distance error
                                                                      5. In mile system, the height unit keep using "ft", not change to "mile"
                                                                      6. Add pace 1K in custom page


Latest version is called "Version X" and you can get it from Geoffrey.

I'm not quite sure... but I remember, there was one of the forum members, who put this
version on his web server... but I don't remember who...

I hope, this could clarify some of your questions.

Good luck, Hubert

[Hubert]

Hi albert,

Hello Hubert,
Of course you are right, sorry for my mistake! Sorry I did confuse you.

Best regards, Albert

Okay - No problem.

And thank you for the nice story. I like these kinds of stories !
(this is why I bought the GH-625...) 

Regards, Hubert