Documentation is still work in progress; we'd like to apologise for the momentary inconvenience.
["SeqN", "Digital0", "Digital1", "Digital2", "Digital3", "EDA"]
, which means you can load it as a dictionary-like structure in pretty much any modern scientific computing or programming language.
This portion of the header actually contains the string with the labels associated with the content of each column. The first 5 columns are always there and correspond to a sample Sequence Number ("SeqN")
, to the state of digital channels labelled on the board as I1-I4 ("Digital0"-"Digital4")
, and in your particular case to one of the analog channels
to which the user assigned the channel label "EDA"
- SeqN: it starts at 0 and goes to 15 and starts again at 0? Why? Is there a connection to the magic 16? Wouldn't it be bette to count the sample rate? If I record with a sample rate of 100, SeqN from 0 to 99 would be easier to understand
This is related with the maximum packet size
sent by the device. Given the number of bits required to send all the analog and digital inputs, the performance we wanted to achieve, and other facilities we needed to include (e.g. CRC), only 4-bits were left. Since we recognise the usefulness of having a sequence number to detect gaps in the samples, even with only 4-bits we decided to include it.
Given that the sampling rate is user-defined, a sample counter or the logic to create a time vector is fairly easy to implement on the base station side. You know that each line corresponds to a sample
, the number of samples per second (sampling rate), and that's all you'll need. Detecting sample gaps due for example to operating the device with low battery or near the Bluetooth coverage range is much more difficult to accomplish at the base station side.
- Digital 1, 2, 3,: What does it mean? Why are the digitals equal 1 while recording only eda?
To make OpenSignals as simple to use as possible UI-wise, we chose to always sample and store the digital inputs
I1-I4. The value 1 means that it is pulled up; if you connect for example a manual switch or a third-party device, you'll be able to toggle this number as shown here: https://www.how.do/guide/bitalino-a-tilt-switch
- EDA: when I start recording EDA with µS, is there the same unity in the txt?
Yes, given that the digital codes of the ADC
(e.g. 0-1024) are much more time- and space-efficient to store in real time and load afterwards than floats for example. Still, you can easily post-process the data to convert it into the correct units via the transfer function of the sensor:http://forum.bitalino.com/viewtopic.php?f=12&t=128
- And what does "digital input" and "digital output" means while recording in opensignals?
is a 1-bit input port that can either be at 0V (GND) or 3.3V (VCC) and is recorded synchronously with the analog inputs; these are typically used for example for annotation using a manual trigger or a third-party device.Digital output
is a 1-bit output port that can either be at 0V (GND) or 3.3V (VCC) and is triggered synchronously with the sampling process; these are typically used for example to trigger third-party devices or peripheral (e.g. the LED).
It would be great to have another page in the manual which describes the output files.
The image in attachment doesn't add much to what we've described above, but maybe it helps.
Note as well that OpenSignals has the source code fully available upon install
, which means that you can tweak it to have a few of the details you've questioned worked out to your preference. For the time being it is lacking documentation, but we'll be completely available to support.
The BITalino Team