[OpendTect_Users] Users Digest, Vol 24, Issue 3

[Bert Bril]

Hi all,



Maurizio wrote:

> 73-76 4 byte float format X source coordinate
> 77-80 4 byte float format Y source coordinate

It happens from time to time. Just for the record: this practice is 
against any of the SEG-Y standards. The SEG-Y headers are not allowed to 
have any (explicit) floating point format data.


> The 4 byte values should not be used because of the too small precision,
> and in fact, after better checking the SEGY, it stores also 8 byte float
> values 8 (bytes 181-196) but (I think...I'm not so expert on this
> topic..) without using a standard 8 byte double format...

Coordinates can also not be put in 8-byte double precision. The standard 
simply does not allow that.


> I'd like to read some comments about or if some of you had similar
> problems...

We get all sorts of files here that are not according to the standard. 
And then, the only thing we are all trying to do is exchange traces of 
data samples. How hard can it be? Well, real hard, as long as the 
standard is much too complex, much too big, much too vague. No wonder 
the producers of the files often don't comply with the standard.


> I can upload a raw file if you are curious..

Ehhh ... not really. Seen many variants on this theme ...


Luckily, there is OpendTect's 'Simple format'. Coordinates are read into 
double-precision floating point numbers, samples are read into 4-byte 
floats. Easy to make, easy to read. And if you need a good SEG-Y file, 
simply export the data from OpendTect. Because OpendTect actually 
complies with all SEG-Y standards, including revision 1.



- Bert