[Gwyddion-users] Height data precision and significant digits

[Erika Callagon]

I've been working on VSI data, and I'm wondering if there is a way to
change the units displayed (which is by default in m, at least for my data)
to nm. Changing the color scale and doing a 'fix mean zero' changes the
units to microns, but with only one decimal place. I'm trying to measure
nm-level height changes, but the data is not displayed up to this precision
(the original data contains this information).

The 'copy' option in 'Statistical Quantities' sometimes would give me 1
more decimal place, but if I can set the number of significant digits
displayed.

Thank you very much!

-Erika

[David Nečas (Yeti)]

Post by Erika Callagon
I've been working on VSI data, and I'm wondering if there is a way to
change the units displayed (which is by default in m, at least for my data)
to nm. Changing the color scale and doing a 'fix mean zero' changes the
units to microns, but with only one decimal place. I'm trying to measure
nm-level height changes, but the data is not displayed up to this precision
(the original data contains this information).

The information definitely is not lost in Gwyddion as it represents
everything in double precision internally (about 16 significant digits).

The colour scale generally covers the entire data range so the precision
of values displayed there is just sufficient to distinguish the dozen
tick marks displayed. It serves for orientation, not for value reading
to many significant digits.

As for reading, in value reading, distance, ... tools, values are
typically displayed with 4 digits (and exported with an extra one).
So you have 10-micron range data and need to read sub-nanometer
differences?

Yeti

[Erika Callagon]

Thanks for the reply. The values that I read (for instance, in Statistical
Quantities) sometimes will only display two significant digits. I'm trying
to measure average height differences from VSI data (in the nm-range), but
the displayed value for Average Height in Statistical Quantities would be
for example "1.2 µm". I also find that the precision displayed varies
between data from same measurement or data set (sometimes, it would be
X.XXX µm, but I can't seem to be able to control it). For example, if I
crop the image, the precision in the displayed readings change. It would be
useful if there is a way to set the precision in the reported measures. Is
there a current work-around to this?

The various functions in Gwyddion look really good, and it'd really be
great if I can maximize them. Thank you very much!

-Erika

Post by Erika Callagon

Post by Erika Callagon
I've been working on VSI data, and I'm wondering if there is a way to
change the units displayed (which is by default in m, at least for my

data)

Post by Erika Callagon
to nm. Changing the color scale and doing a 'fix mean zero' changes the
units to microns, but with only one decimal place. I'm trying to measure
nm-level height changes, but the data is not displayed up to this

precision

Post by Erika Callagon
(the original data contains this information).

The information definitely is not lost in Gwyddion as it represents
everything in double precision internally (about 16 significant digits).
The colour scale generally covers the entire data range so the precision
of values displayed there is just sufficient to distinguish the dozen
tick marks displayed. It serves for orientation, not for value reading
to many significant digits.
As for reading, in value reading, distance, ... tools, values are
typically displayed with 4 digits (and exported with an extra one).
So you have 10-micron range data and need to read sub-nanometer
differences?
Yeti
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
Gwyddion-users mailing list
https://lists.sourceforge.net/lists/listinfo/gwyddion-users

[David Nečas (Yeti)]

Post by Erika Callagon
The values that I read (for instance, in Statistical
Quantities) sometimes will only display two significant digits. I'm trying
to measure average height differences from VSI data (in the nm-range), but
the displayed value for Average Height in Statistical Quantities would be
for example "1.2 µm".

This really should not happen. Could you maybe send me some example
data where you encounter this (possibly with a hint what to try reading
if it is not obvious)?

Post by Erika Callagon
It would be useful if there is a way to set the precision in the
reported measures. Is there a current work-around to this?

It is not possible to set the precision manually. And I am quite
resistant against such feature request, always trying to make things
work automatically better instead – because that actually improves the
software and makes it more pleasant to use, as opposed to enabling
workarounds for things that work poorly. So I am trying to understand
what goes wrong for your data.

Regards,

Yeti

[Erika Callagon]

Thank you, Yeti. I have sent some sample data to your e-mail.

-Erika

Post by Erika Callagon
The values that I read (for instance, in Statistical
Quantities) sometimes will only display two significant digits. I'm trying
to measure average height differences from VSI data (in the nm-range), but
the displayed value for Average Height in Statistical Quantities would be
for example "1.2 µm".

This really should not happen. Could you maybe send me some example
data where you encounter this (possibly with a hint what to try reading
if it is not obvious)?

Post by Erika Callagon
It would be useful if there is a way to set the precision in the
reported measures. Is there a current work-around to this?

It is not possible to set the precision manually. And I am quite
resistant against such feature request, always trying to make things
work automatically better instead – because that actually improves the
software and makes it more pleasant to use, as opposed to enabling
workarounds for things that work poorly. So I am trying to understand
what goes wrong for your data.

Regards,

Yeti

[Joe Smerdon]

Hi all,

The only reason I use anything other than Gwyddion these days is when I have reams of data and need to produce browse documents (I use ImageSXM for this, but WSxM still kind of works). I was wondering if we can have this function in Gwyddion? That is, the option to produce a document containing all the thumbnails for a folder, annotated with the file name (and maybe exp conditions and scale). In ImageSXM it's a paged TIFF file with a screenful of images per page.

thanks

Joe

[Andrés Muñiz Piniella]

Post by Joe Smerdon
Hi all,
The only reason I use anything other than Gwyddion these days is when I
have reams of data and need to produce browse documents (I use ImageSXM
for this, but WSxM still kind of works). I was wondering if we can
have this function in Gwyddion? That is, the option to produce a
document containing all the thumbnails for a folder, annotated with the
file name (and maybe exp conditions and scale). In ImageSXM it's a
paged TIFF file with a screenful of images per page.

Hi Joe,
https://sourceforge.net/p/gwyddion/wiki/Scripts/
Some time ago I wrote a script that would generate a pdf via LaTeX with all the images in a gay file. Last time I used it it worked. Hopefully this is something you are looking for?
Maybe add to the script a facility to merge the file in a folder to the document?

[Joe Smerdon]

Hi Andres,

thanks for the script - it is definitely useful. But really I have hundreds of images, only a few of which are ever going to be useful, and the script generates one page per image. I don’t intend to teach granny to suck eggs (hopefully you get that English phrase - if not, it means I don’t wish to presume to teach an expert, but to avoid confusion, I’ll say something very simply) but if you have a look at how ImageSXM manages it, you will see what I currently consider the best handling of biggish SPM datasets.

I do see how the script could be modified to put several images on one page, but I think that while this modification would be ok, a simple function integrated into Gwyddion to do the same (ideally with checkboxes or something to decide what information to include with the thumbnails) and with the possibility to subsequently click on a thumbnail to load it into Gwyddion, would be a functionality with wide appeal.

Thanks again

Joe

Post by Andrés Muñiz Piniella

Post by Joe Smerdon
Hi all,
The only reason I use anything other than Gwyddion these days is when I
have reams of data and need to produce browse documents (I use ImageSXM
for this, but WSxM still kind of works). I was wondering if we can
have this function in Gwyddion? That is, the option to produce a
document containing all the thumbnails for a folder, annotated with the
file name (and maybe exp conditions and scale). In ImageSXM it's a
paged TIFF file with a screenful of images per page.

Hi Joe,
https://sourceforge.net/p/gwyddion/wiki/Scripts/
Some time ago I wrote a script that would generate a pdf via LaTeX with all the images in a gay file. Last time I used it it worked. Hopefully this is something you are looking for?
Maybe add to the script a facility to merge the file in a folder to the document?
--
Andrés
------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most
engaging tech sites, SlashDot.org! http://sdm.link/slashdot
_______________________________________________
Gwyddion-users mailing list
https://lists.sourceforge.net/lists/listinfo/gwyddion-users

[David Nečas (Yeti)]

...a simple function integrated into Gwyddion...

Here we probably need a reality check. What is the most humongous
single Gwyddion function, more than 60% larger than any other module?
Yes, the image export. It is several times larger than some other
things I already consider monstrosities. And people still keep asking
for additional features and options.

So I doubt any ‘simple function’ would do. It would end up as a major
project for sure. In addition, the most natural implementation for me,
writing a TeX document and then having an actual well-designed
typesetting system typeset the entire thing, would unlikley make
non-Unix users happy.

While I do not say it is completely impossible or against my religion or
anything, I must say I find the idea of spending time with this pretty
unattractive – as opposed to experimenting with some interesting new
data processing methods. I wish good luck to anyone who endeavours to
implement it.

Regards,

Yeti

[David Nečas (Yeti)]

Post by Erika Callagon
Thank you, Yeti. I have sent some sample data to your e-mail.

I revised the statistical quantities module and the unit and precision
format choices should be improved, in particular they are chosen
according to the selected part instead of the entire image and
dispersion-like quantities were decoupled from total-range-like
quantities (and a few other things). The changes will appear in
tomorrow's development snapshot.

That said, I do not see the tool doing anything that wrong in your
screenshots. When the Z range is 20+ µm then 10 nm precision of the
minimum, maximum and mean values seems about right. The precision of
dispersion-like quantities was affected by the global value offset, and
that should be fixed now. Generally, if you data have micron range (and
you even want nm-like differences) but are globally offset by 3.6 cm
some things can be expected to be displayed in an odd and possibly
unhelpful manner. Similarly, extracting smaller regions of interest
from a 4.3k×2.6k image for detailed analysis is always going to be
beneficial for data processing speed, if nothing else.

Regards,

Yeti