Export data formats

List of file formats used by BV Workbench.

CSV

Comma separated 64-bit floating point values.

• Signals : time column and value column for each signal.
• Images : pixel values exported in the same arrangement of row,columns as the data.

Apache Parquet

Binary format that can be read by MATLAB, Python, etc.

It is generally more efficient than CSV. See more details here.

DAT

Custom binary format used by BV workbench. The first 4 bytes indicates the data type.

DATA_TYPE	Description
0x00001D01	Time series
0x00002D03	Spatio-temporal data
0x00002D04	Time-frequency data
0x00002D05	Scalar map data
0x00002D06	Velocity map data
0x00003D02	Phase map data

NOTE

Map values are stored in row-major order.

Time series

Current version : 1

Offset	Field name	Type	Size	Units
0	DATA_TYPE	Int32	4
4	VERSION	Int32	4
8	START_TIME	Float64	8	seconds
16	SAMPLING_TIME	Float64	8	seconds
24	INPUT_RANGE_MIN	Float64	8
32	INPUT_RANGE_MAX	Float64	8
40	LENGTH	Int32	4
...	RESERVED	-	-
512	SIGNAL VALUES	Float64	LENGTH * 8

NOTE

The units depends on the input. Data from analog inputs are stored in volts.

Scalar map data

Current version : 1

Offset	Field name	Type	Size	Units
0	DATA_TYPE	Int32	4
4	VERSION	Int32	4
8	WIDTH	Int32	4
12	HEIGHT	Int32	4
40	SCALE_X	Float64	8	mm/px
48	SCALE_Y	Float64	8	mm/px
56	SAMPLE_COUNT	Int32	4
60	SCALAR_TYPE	Int32	4
...	RESERVED	-	-
512	BACKGROUND	UInt16	WIDTH * HEIGHT * 2
512 + WIDTH * HEIGHT * 2	VALUES	Float32	WIDTH * HEIGHT * 4

NOTE

The units depends on the scalar type field.

enum SCALAR_TYPE {
    ActivationTime = 1, // s
    RiseTime = 2, // ms
    PeakTime = 3, // s
    PeakAmplitude = 4,
    PeakToDecayTime = 5, // ms
    DecayTime = 6, // ms
    DecayTau = 7, // ms
    APD = 8, // ms
    UpstrokeVelocity = 9, // units/ms (a.u.)
    PeakToPeakInterval = 10, // ms
    DiastolicInterval = 11, // ms
    Frequency = 12, // Hz
    Velocity = 13,  // m/s
    Alternans = 14, // % change relative to previous
    ApdAlternans = 15, // ms
};

Velocity map data

Current version : 1

Offset	Field name	Type	Size	Units
0	DATA_TYPE	Int32	4
4	VERSION	Int32	4
8	WIDTH	Int32	4
12	HEIGHT	Int32	4
40	SCALE_X	Float64	8	mm/px
48	SCALE_Y	Float64	8	mm/px
56	SAMPLE_COUNT	Int32	4
...	RESERVED	-	-
512	BACKGROUND	UInt16	WIDTH * HEIGHT * 2
512 + WIDTH * HEIGHT * 2	VECTORS	Vector	WIDTH * HEIGHT * 8	m/s

struct Vector {
    float x; // m/s
    float y; // m/s
};

Phase map data

Current version : 1

Offset	Field name	Type	Size	Units
0	DATA_TYPE	Int32	4
4	VERSION	Int32	4
8	WIDTH	Int32	4
12	HEIGHT	Int32	4
16	FRAME_COUNT	Int32	4
40	SCALE_X	Float64	8	mm/px
48	SCALE_Y	Float64	8	mm/px
56	START_TIME	Float64	8	seconds
64	SAMPLING_TIME	Float64	8	seconds
...	RESERVED	-	-
512	BACKGROUND	UInt16	WIDTH * HEIGHT * 2
512 + WIDTH * HEIGHT * 2	PHASE	Float32	WIDTH * HEIGHT * FRAME_COUNT * 4	radians
512 + WIDTH * HEIGHT * (2 + FRAME_COUNT * 4)	SINGULARITIES	PointArray	Length prefixed array

// for each frame
struct Singularities {
    int count;
    Point* y;
};
struct Point {
    double x;
    double y;
};

Time-frequency data

Current version : 1

Offset	Field name	Type	Size	Units
0	DATA_TYPE	Int32	4
4	VERSION	Int32	4
8	WIDTH	Int32	4
12	HEIGHT	Int32	4
...	RESERVED	-	-
512	MAGNITUDE	Float32	WIDTH * HEIGHT * 4
512 + WIDTH * HEIGHT * 4	TIME_VALUES	Float64	WIDTH * 8	seconds
512 + WIDTH * HEIGHT * 4 + WIDTH * 8	FREQ_VALUES	Float64	HEIGHT * 8	Hz

Spatio-temporal data

Current version : 1

Offset	Field name	Type	Size	Units
0	DATA_TYPE	Int32	4
4	VERSION	Int32	4
8	WIDTH	Int32	4
12	HEIGHT	Int32	4
16	START_TIME	Float64	8	seconds
24	SAMPLING_TIME	Float64	8	seconds
32	SCALE_X	Float64	8	mm/px
40	SCALE_Y	Float64	8	mm/px
48	POINT_COUNT	Int32	4
...	RESERVED	-	-
512	AMPLITUDE	Float32	WIDTH * HEIGHT * 4
512 + WIDTH * HEIGHT * 4	POINTS	Point	POINT_COUNT * 8

NOTE

In this context, WIDTH is the number of time points. HEIGHT is the total number of divisions along the (poly)line. POINTS are coordinates in the source image sequence. Use SCALE_X and SCALE_Y to get distances between points.

struct Point {
    int x;
    int y;
};