Sigma sd Quattro .x3f files only **1/4 detail**

[DrJulian]

Raw files saved in Sigma sd Quattro / H

I've raised this issue before though I'm not sure it's been understood, I hope the attached crops make it clear.

/

.x3f raws are developing with only 1/4 of the detail (the file is the correct resolution but is 1/4 detail 2x2 pixels)

.dng raws are good (this is how the .x3f raw should look).

Please see attached 200% crops.

/

Raws, .x3f and .dng are here :

https://we.tl/t-6OrfquyunL

[DrJulian]

I have been informed that the higher detail present in the .dng files is due to interpolation, and contains no more detail than the .x3f files (when both developed in RPP).

It is clear to me that the .x3f files would not suddenly gain the extra detail that is present in the .dng developed files. It might smooth some edges, but it would not magically produce the extra detail that's seen in the .dng developed file.

The .x3f files are not being developed correctly, there is detail (from the the top layer of the Foveon Quattro sensor) that is missing.

See attached, enlarged further (view at full-size - pixels will be large), with the x3f and dng developed files, much closer together.

The right image could not be an interpolated version of the left, there is more detail, more information to it.

Best regards,

Julian

[Iliah Borg]

You can run FFT and see it is interpolation.

On 2/7/23 05:17, DrJulian wrote:
> I have been informed that the higher detail present in the .dng files is due to interpolation, and contains no more
> detail than the .x3f files (when both developed in RPP).
>
> It is clear to me that the .x3f files would not suddenly gain the extra detail that is present in the .dng developed
> files. It might smooth some edges, but it would not magically produce the extra detail that's seen in the .dng developed
> file.
>
> The .x3f files are not being developed correctly, there is detail (from the the top layer of the Foveon Quattro sensor)
> that is missing.
>
> See attached, enlarged further (view at full-size - pixels will be large), with the x3f and dng developed files, much
> closer together.
>
> The right image could not be an interpolated version of the left, there is more detail, more information to it.
>
> Best regards,
> Julian
>
>
> On Monday, February 6, 2023 at 7:34:30 PM UTC DrJulian wrote:
>
> Raw files saved in Sigma sd Quattro / H
>
> I've raised this issue before though I'm not sure it's been understood, I hope the attached crops make it clear.
>
> /
>

> .*x3f raws are developing with only 1/4 of the detail* (the file is the correct resolution but is 1/4 detail 2x2 pixels)
>
> *.dng raws are good* (this is how the .x3f raw should look).

>
> Please see attached 200% crops.
>
> /
>
> Raws, .x3f and .dng are here :
>

> https://we.tl/t-6OrfquyunL <https://we.tl/t-6OrfquyunL>
>
> --
> You received this message because you are subscribed to the Google Groups "Raw Photo Processor" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to
> raw-photo-proce...@googlegroups.com <mailto:raw-photo-proce...@googlegroups.com>.
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/raw-photo-processor/db40679f-1ac1-44d6-a115-0a9234fc9ebbn%40googlegroups.com
> <https://groups.google.com/d/msgid/raw-photo-processor/db40679f-1ac1-44d6-a115-0a9234fc9ebbn%40googlegroups.com?utm_medium=email&utm_source=footer>.

--
Best regards,
Iliah Borg
LibRaw, LLC
www.libraw.org
www.rawdigger.com
www.fastrawviewer.com

[DrJulian]

If is it interpolation, is there a way in RPP we can get .x3f developed images to look like the .dng images?

i.e. can a routine be added that will interpolate the information in the .x3f Quattro files, to provide the extra 'perceived' detail that's seen in the dng files.

As at present, the x3f files don't look too good, yet if the potential is within them to look good (as good as the dngs) it would be great to see it.

Best regards,

Julian