18.11.2009 14:03:37 UTCgeändert am 18.11.2009 14:03:55 UTC
I'll answer in English because this will interest many users, I guess. If a German translation is required, please tell me and I'll add one.
Removing a complete column from the hr-data.dat file:
1) Open MS Excel (OpenOffice should do too)
2) Open the "hr-data.dat" file in Excel and select " " (space) as separator character
3) Mark the entire column by clicking on the letter on top of the column ("A", "B", ...)
4) Press delete
5) Save the file in CSV format as hr-data-mod-dat
6) Now you may have to remove lines that contain zero with a text editor
7) Open the file in Kubios
I'll probably add an easier solution for this later.
17.11.2009 23:05:12 UTCgeändert am 18.11.2009 12:06:04 UTC
In der Bedienungsanleitunmg steht: "Tipp: Falls es Probleme beim Import de
Datei gibt, kann es helfen, Zeilen mit
Nullwerten oder auch die komplette linke
Spalte zu entfernen."
Mit welchem Editor kann man senkrecht Zahlen markieren um die linke Spalte zu entfernen? Von Hand bei mehreren Tausend Zeilen sitzt man ein paar Monate dran.
Muss man jetzt Spass am Zeilencode entwickeln um das jetzt auslesen zu können?
Fakt ist, bei mir läßt sich keine einzigste Dat in Kubios öffnen und Tausende Zeilen von Hand editieren dient nicht einer Steigerung der Herzratenvariabilität.;-)
Gruß Dorphin
@Dima: Actually the zero values are not bugs, they are needed. That's because we use the file not only for exporting to other programs but also for when you load the training session in Run.GPS! We need this to be able to display all the HRV related values.
If you stop the recording and then start it again, the line would read like this:
323424 0
....
The first value is the timestamp of the heart beat, and there is no interval duration yet, so it's zero. But we need the timestamp to be able to reproduce all data.
@dorphin: you do need to insert the dots! It's okay to use milliseconds (this is also in the Kubios manual). Es ist nicht notwendig, Punkte einzufügen, Kubios akzeptiert auch Millisekunden.
@ALL: Please take a look at the new section in the Run.GPS 2.3.4 manual. It should be enough to just remove the lines where the second column (NN interval duration) is zero.
15.11.2009 14:19:27 UTCgeändert am 15.11.2009 14:39:01 UTC
Ich weiß jetzt warum sich die hr-data.dat bei mir nicht importieren läßt. in der Bedienungsanleitung von Kubios steht das das ASCII-Format so aussehen muss:
747 779
1543 796
2334 791
3118 784
3877 759
.
.
.
Nachdem ich per Hand die Nullen und die Pünktchen eingefügt hatte, ließ es sich problemlos importieren.
Yes, the reality exceeded all expectations. NN-is exactly what I needed and it is usable.
All other indicators need to be studied additionally (I am not a cardiologist :)
There is the same problem with HRV.
During the ortho-test (6-8 min) all data is sufficient (but also not readable by HRV), but during training (30 min) hr-data.dat has too much zero values, which sometimes as long as 10-20 seconds.
Actually current version is what I needed, and these bugs are just to report.
Hallo, finde ich ja super das es von der Idee bis zur Umsetzung nur 4 Tage brauchte! Ich wollte das jetzt gleich mal ausprobieren und habe die zuvor erstellte hr-data.dat in Kubios HRV geladen aber leider bleibt das Programm dann beim importprozess stehen. s.Screenshot:
13.11.2009 14:46:01 UTCgeändert am 13.11.2009 14:47:23 UTC
Thanks for your interesting suggestions!
We have decided to include the following in the 2.3.4 release:
- NN analysis
- real time values:
+ SDNN
+ pNN50
+ pNN100
+ avg NN interval
+ avg HRV
- a data file will also be produced which can easily be imported to other programs, below is a screenshot of Kubios HRV, filled with data from Run.GPS
Probably the excellent decision would be to include some information which allows to calculate R-R intervals.
t1: HR1= R-R1
t2: 60/(HR2 + heart rate variability1) = R-R2
...
I think it's enough, but just if time periods are close to R-R interval!
Actually I could calculate R-R right know if time periods in csv are not too long (now 1-4 s) and there is no averaging HR over time.
The best time interval for that calculation is 1/HR (=R-R :)).
This will follow the greater data files, which is not suitable for average user from one hand, but from another it significantly broadens options for data analysis.
Most software allows to import R-R data files and most sensor enabled systems provide that (Polar, Suunto etc.). The same option for RunGPS+HxM would improve it's attractiveness as a training system for potential users.
Probably it would be enough to just record the heart rate variability and not the exact timestamps of each heart beat. The interesting values are:
- the average heart rate variability in a training session
- maybe the current heart rate variability v = variability(time)