Misleading text in SEGY header?

[jcrivenaes]

We have just refactored xtgeo, added segyio export to xtgeo, cf. equinor/xtgeo#1529. When I scan the text header, I see now:

C 1 DATE 2026-02-12                                                             
C 2 AN INCREASE IN AMPLITUDE EQUALS AN INCREASE IN ACOUSTIC IMPEDANCE           
C 3 Written by libsegyio (python)                                               
C 4                      

I wondered why the line "AN INCREASE IN AMPLITUDE EQUALS AN INCREASE IN ACOUSTIC IMPEDANCE" is relevant. And is it correct?

Line C2 seems to be a hardcoded default in segyio. Asking chat:

Is this always true?: AN INCREASE IN AMPLITUDE EQUALS AN INCREASE IN ACOUSTIC IMPEDANCE



Short answer: No.

Why:

In linear acoustics, acoustic impedance is a property of the medium and geometry, primarily depending on frequency, not on signal amplitude.
Characteristic (specific) acoustic impedance of a medium: Z0 = ρ c (density × speed of sound). This does not change when you turn the volume up.
Local/specific impedance: Z = p/u (pressure over particle velocity). If you increase the source amplitude in a linear system, p and u scale proportionally, so Z stays the same.
What amplitude does affect:

Intensity and power. For a given medium, I = p_rms^2 / Z0, so intensity grows with the square of pressure amplitude while Z0 remains constant.
When amplitude could affect “impedance”:

Nonlinear regimes (very high levels) where the medium’s properties change (heating, shock formation).
Materials with amplitude- or flow-dependent behavior (some porous/duct systems), or transducer nonlinearity/saturation.
Bottom line: In ordinary, small-signal acoustics, increasing amplitude does not increase acoustic impedance.

[ajaust]

I think that is a good question. I had a brief digging in the segyio history and it turns out that this comment has been around since the beginning of segyio. It is already part of the Initial commit (5c05b1b) and it was moved around once (033f3f7).

I guess we would need some more digging to figure out how this comment got into segyio.

[ajaust]

Is this an issue for your application? You should be able to set/overwrite the corresponding text header from xtgeo when writing the file.

[jcrivenaes]

Is this an issue for your application? You should be able to set/overwrite the corresponding text header from xtgeo when writing the file.

It is not a big issue, just a "nuance". When we e.g. take the cubes into RMS, it is easy for any users to view the header. I think, in principle, that comments in the headers should be relevant and scientificly correct. I propose just to remove it as default.

A cube can also be other properties (often related to seismics, but also other non-seismic properties), not only amplitudes.

[ajaust]

For future reference: We have gotten further input by another specialist who agrees with you. This comment should not be the default. The comment might not be true/relevant depending on the data stored in the cube. Additionally, there is a dedicated field in the binary header (bytes 3257–3258) to indicate polarity.

[mferrera]

segyio/python/segyio/create.py

Line 25 in ce7bf86

rows[-1] = 128 # \x80 -- Unsure if this is really required...

I think this should also be removed, or at least converted to a NUL-byte. It's a fixed-length field and x80 seems an odd terminator for a string.

We see this pop up when using OpenVDS' SEGYImport to convert segyio-exported files to vds, which prints the text header. If this is run as a Python subprocess with text=True it raises a Unicode decode error.