Reading SEG-Y files
This section documents how SEG-Y files are read using SegyMAT.
ReadSegy
ReadSegy.m can be used to read SEG-Y formatted files :
[Data,SegyTraceHeaders,SegyHeader]=ReadSegy('data.segy');
wiggle(Data,[],SegyHeader.time,[SegyTraceHeaders.cdp],'VA')
imagesc([SegyTraceHeaders.cdp],[SegyHeader.time],Data)
This will read data.segy using the revision and data sample format specified in the binary header (SegyHeader), and plot the data using the wiggle plotting function.
Data is a 2D variable containing the seismic data of size [Nsamples x Ntraces].
SegyTraceHeadersis a structure of size [1,Ntraces]' structure containing all the header values from the traces. TypeSegyTraceHeadersto see a list of header information.SegyTraceHeaders(9)`, list all header names and values of trace number 9.
>> SegyTraceHeaders(9)
ans =
SegyMAT_TraceStart: 91952
TraceSequenceLine: 0
TraceSequenceFile: 9
FieldRecord: 0
TraceNumber: 9
EnergySourcePoint: 0
cdp: 0
cdpTrace: 0
TraceIdenitifactionCode: 0
NSummedTraces: 0
NStackedTraces: 0
DataUse: 0
offset: 400
...
SourceEnergyDirectionMantissa: 0
SourceEnergyDirectionExponent: 0
SourceMeasurementMantissa: 0
SourceMeasurementExponent: 0
SourceMeasurementUnit: 0
UnassignedInt1: 0
UnassignedInt2: 0
SegyMAT_TraceDataStart: 92192
To access an array of trace header values simply use square brackets as :
cdp=[SegyTraceHeaders.cdp];
offset=[SegyTraceHeaders.offset];
...
SegyHeader is a structure containing all the Segyheader values. Typing
SegyHeader will list the names and values of all header
values.
>> SegyHeader
SegyHeader =
Rev: [1x2 struct]
TextualFileHeader: [3200x1 double]
Job: 0
Line: 0
Reel: 0
DataTracePerEnsemble: 0
AuxiliaryTracePerEnsemble: 0
dt: 1000
dtOrig: 0
ns: 2701
nsOrig: 0
DataSampleFormat: 5
EnsembleFold: 0
TraceSorting: 0
VerticalSumCode: 0
SweepFrequencyStart: 0
SweepFrequencyEnd: 0
SweepLength: 0
SweepType: 0
SweepChannel: 0
SweepTaperlengthStart: 0
SweepTaperLengthEnd: 0
TaperType: 0
CorrelatedDataTraces: 0
BinaryGain: 0
AmplitudeRecoveryMethod: 0
MeasurementSystem: 1
ImpulseSignalPolarity: 0
VibratoryPolarityCode: 0
Unassigned1: [120x1 double]
SegyFormatRevisionNumber: 100
FixedLengthTraceFlag: 1
NumberOfExtTextualHeaders: 0
Unassigned2: [47x1 double]
time: [1x2701 double]
A number of arguments can be given to ReadSegy, controlling what type of and which part of the data to read.
Read specific trace numbers
To read traces 100, 201 and 320 use e.g.
>> [Data,SegyTraceHeaders,SegyHeader]=ReadSegy(filename,'traces',[100 201 320]);
Use for example ReadSegyTraceHeadervalue.m and 'find' to find a list of trace ids (this is equivalent to using the 'minmax' option)
>> cdp=ReadSegyTraceHeaderValue(filename,'key','cdp');
>> traces = find(cdp>100 & cdp<200)
>> [Data,SegyTraceHeaders,SegyHeader]=ReadSegy(filename,'traces',[100 201 320]);
Read only every 5th trace
>> [Data,SegyTraceHeaders,SegyHeader]=ReadSegy(filename,'jump',5);
To read time slice 0.5 < t < 5
>> [Data,SegyTraceHeaders,SegyHeader]=ReadSegy(filename,'trange',.5,3);
Read data in a CDP header range : 5000<cdp<5800 (the 'minmax' option)
cdp can be changed to any other valid header name
>> [Data,SegyTraceHeaders,SegyHeader]=
ReadSegy(filename,'minmax','cdp',5000,5800);
Read only header values
In some cases it can be desirable only to read the header information (the SegyHeader and SegyTraceHeaders). This will return an empty Data variable.
>> [Data,SegyTraceHeaders,SegyHeader]=
ReadSegy(filename,'SkipData',1);
SEG-Y format revision
SEG-Y format revision number can be '0' (1975) or '1' (2002). By default the SEG-Y format revision number is read in the binary header, but this can be overruled using :
>> [Data,SegyTraceHeaders,SegyHeader]=
ReadSegy(filename,'revision',0);
A specific Data Sample Format
One can overrule the Data Sample Format listed in the binary header,
using the dsf argument. See Data Sample Formats for a list
of valid and supported values.
>> % Rev 0, IBM FLOATING POINT
>> [Data,SegyTraceHeaders,SegyHeader]=
ReadSegy(filename,'revision',0,'dsf',1);
>> % Rev 1, IEEE FLOATING POINT
>> [Data,SegyTraceHeaders,SegyHeader]=
ReadSegy(filename,'revision',1,'dsf',5);
If dsf is set to 5 and revision to 0, a warning message will occur,
since data sample format 5 is only defined in revision 1. The revision
is then automatically set to 1.
Force the use of a specific SegyHeader
>> [Data,SegyTraceHeaders,SegyHeader]=
ReadSegy(filename,'SegyHeader',SegyHeader);
ReadSegyFast
ReadSegyFast.m is a faster implementation of ReadSegy.m
since no trace header values are read. Thus this function will just
return the seismic data and the SegyHeader. e.g. :
>> [Data,SegyHeader]=ReadSegyFast('data.segy');
>> imagesc(Data)
If ReadSegy is called with only one output argument, ReadSegyFast will be used instead of ReadSegy.
ReadSegyFast options
Most of the same options that works for ReadSegy.m will also work for ReadSegyFast.m. The data sample format can be chosen using the 'revision' and 'dsf' tags. Also a 'SegyHeader' can be specified.
ReadSegyFast.m is currently optimized only for reading the whole SEGY-Y file, but the options 'jump' and 'trange' can be used (but will currently not result in faster read times).
Since the trace header values are not read, the 'minmax' option is not supported.
ReadSegyHeader
ReadSegyHeader.m reads the Binary Segy Header only. It can be called
with the same options as ReadSegy.m
Force using little endian :
>> SegyHeader=ReadSegyHeader(filename,'endian','l');
ReadSegyTraceHeaderValue
ReadSegyTraceHeaderValue.m reads one trace header value into an array. This approach is much faster than to read the whole file
using a keyword
To read a trace header value by its trace header key. See the definition of all the Trace Header names to use the correct key:
cdp=ReadSegyTraceHeaderValue(filename,'key','cdp');
SourceX=ReadSegyTraceHeaderValue(filename,'key','SourceX');
SourceY=ReadSegyTraceHeaderValue(filename,'key','SourceY');
plot(SourceX,SourceY)
using location+type
To read a trace header by its position in the trace header using a specific data sample format, use:
SourceX=ReadSegyTraceHeaderValue(filename,'pos',70,'precision','int32');
ReadSegyConstantTraceLength
Assuming a constant trace length (which is much more common than not) allows much faster reading of parts of large file. For example to read trace number 2030, the whole SEG-Y file must be sequentially read, assuming variable trace length. Assuming constant trace length the trace can be directly (and fast) located in the data cube.
To read trace 2030 use
[Data,SegyTraceHeader,SegyHeader]=ReadSegyConstantTraceLength(filename,'trace',2030);
To read traces 1-2000 and 2020-2040 use
[Data,SegyTraceHeader,SegyHeader]=ReadSegyConstantTraceLength(filename,'trace',[1:2000,2020:2040]);
using keywords
Several keywords can be used to efficiently read parts of larger files.
To read only the part of a file with SourceX between 1000-2000 and SourceY between 4000-5000 use :
[Data,SegyTraceHeader,SegyHeader]=ReadSegyConstantTraceLength(filename,'minmax','SourceX',1000,2000,'minmax','SourceY',4000,5000]);
ReadSu
ReadSu.m works similar to ReadSegy.m and the same input parameters can be used. A SuHeader can optionally be returned, but as there is no (SEG-Y)-Header information in a SU file it is mostly empty.
>> [Data,SuTraceHeaders,SuHeader]=ReadSu(filename);