Volume from Flow signal

Dieses Thema hat 3 Antworten sowie 2 Teilnehmer und wurde zuletzt vor vor 20 Jahren, 6 Monaten von paulwarshawsky@yahoo.com aktualisiert.

Anzeigen von 4-Stellen - 1 bis 4 (von insgesamt 4)

Autor

Beiträge
23.11.2004 um 01:53 Uhr #34166

paulwarshawsky@yahoo.com
Teilnehmer

I need to create a signal of volume (vs. time) from a flow signal – a record of a person breathing. This should be relatively easy except that the volume drifts, so I would like to reset the volume to zero at every -ve to +ve zero crossing of flow (i.e. at the start of a new breath.) I am going to need to do this every time I analyze a recording that I make so the process needs to be automated.

Suggestions?

Thanks,

Paul

23.11.2004 um 02:11 Uhr #34164

Bernhard Kantz
Teilnehmer

There are two options:

1.) Write a FPScript formula which gets the dataset as an argument and does the de-trending. You should use the LevelCrossings function to find all the positions and then loop accross each breath to substract the offset. You can use Drag & Drop for example to produce new formulas which call this function for a couple of datasets in one step.

2.) Use Automation to access and modify the dataset by reading the data in chunks, processing it and writing it back to the same or to a different dataset.

23.11.2004 um 21:44 Uhr #34163

paulwarshawsky@yahoo.com
Teilnehmer

I need to create a signal of volume (vs. time) from a flow signal – a record of a person breathing. This should be relatively easy except that the volume drifts, so I would like to reset the volume to zero at every -ve to +ve zero crossing of flow (i.e. at the start of a new breath.) I am going to need to do this every time I analyze a recording that I make so the process needs to be automated.

Suggestions?

Thanks,

Paul

23.11.2004 um 21:44 Uhr #34165

Bernhard Kantz
Teilnehmer

Maybe, the following approach makes sense for this application:

Use the FPScript functions UpperEnvelope and LowerEnvelope, to compute the enevelopes of the signal, re-sample them with the original X-data and then form the arithmetic mean (L + H) / 2. Substract this from your data to perform the de-trending.

support@weisang.com
Autor

Beiträge

Anzeigen von 4-Stellen - 1 bis 4 (von insgesamt 4)

Du musst angemeldet sein, um auf dieses Thema antworten zu können.