Тема: Обробка астрофото із Siril (Прочитано 8373 раз)

lock042

Цитувати

However conversion doesn't work for Fujifilm X-A3 (bayer sensor).

OK. I think it is normal. It is a relatively new DSLR. We use the library libraw to decode images. Just wait they take into account the camera. (in fact, It is the case now, but the lib version embeded in your Siril does not). Is your platform Windows ?

Take a look here: https://www.libraw.org/news/libraw-0-19-snapshot-20170812
Version available on Windows seems to be:
mingw/mingw-w64-libraw 0.18.9-1
Library for reading RAW files obtained from digital photo cameras (mingw-w64)

Please share me an image of your X-A3

Цитувати

I also thinking that "bayer pattern" is not really correct naming for all types of raw files in output, probably more common like "CFA pattern" would be more close?

I think you are right since XTRANS is not Bayer

AYN

Цитата: lock042 від 07 Вересня 2018, 10:27:56

OK. I think it is normal. It is a relatively new DSLR. We use the library libraw to decode images. Just wait they take into account the camera. (in fact, It is the case now, but the lib version embeded in your Siril does not). Is your platform Windows ?

I'm using Siril 0.9.9 on Linux Ubuntu. Latest libraw (0.19) has support for X-A3 definitely as I'm also using Sequator ( https://sites.google.com/site/sequatorglobal/ ) which also based on libraw (0.19-release) and it opens X-A3 raw files nicely.

P.S. Sample raw file from X-A3 placed here:
https://jmp.sh/zX2yCFc

lock042

On my debian testing I have no problem to open the file.
On Ubuntu the libraw version is not the 0.19: https://packages.ubuntu.com/bionic/libraw16

Код: [Select]

log: Decoding Fujifilm X-A3 file (ISO=3200, Exposure=117.0 sec)
log: CFA pattern: RGGB
log: Reading RAW: file FXA38386.raf, 1 layer(s), 6016x4014 pixels
Loading image OK, now displaying

AYN

Цитата: lock042 від 07 Вересня 2018, 14:18:50

On my debian testing I have no problem to open the file.
On Ubuntu the libraw version is not the 0.19: https://packages.ubuntu.com/bionic/libraw16

Thanks, you are right. For some reason I thought siril package installs automatically related latest libraw version, didn't look at it closely.

lock042

you can compile libraw0.19 in your Ubuntu and it should work

.

mpyat2

Цитата: lock042 від 05 Вересня 2018, 14:58:45

Hello. Did you improve the translation ?

Hi Cyril,

I've attached improved translation (mo lines translated).

Could you please calculate a percentage of translated lines? My free poedit lacks this function.

Best regards,

Maksym

lock042

Thank you

Цитувати

Could you please calculate a percentage of translated lines? My free poedit lacks this function.

Are you sure

I also have the free edition and at the bottom you have the percent. Here: 54%

mpyat2

Ha-ha! Indeed, you are right. I have not used poedit for years and tried to use "Statistics" button which warn me that it requires poedit pro.

AYN

Цитата: lock042 від 08 Вересня 2018, 17:07:25

you can compile libraw0.19 in your Ubuntu and it should work .

Will do... when I have some time

Thanks.
BTW - tried to use Windows siril package 0.9.9 and it also unable to process X-A3 raf files.

lock042

Цитувати

BTW - tried to use Windows siril package 0.9.9 and it also unable to process X-A3 raf files.

Yes, when I did the compilation libraw was not at the right version on the system

.

You can use Debian testing

.

lock042

Hello mpyat2. I hope you're going well.
How is the translation ?

.

You may be happy to know that I've implemented a platesolving tool in Siril (available for now, in beta version, in linux with the PPA repositories). The aim of this tool is to (slowly) introduce a Johnson BV photometric color calibration tool. It will take some time but it is kind of exciting !!

mpyat2

Hi Cyril,

There was no actual progress in translation yet. Rainy November is coming so I (hope) will move forward with the translation.

Regarding photometry tool etc: IMHO more important task is to solve a problem with possible negative values (due to noise) in frames after removing offset (bias). Now all negative values are simply truncated...
Moreover, if I try to open a calibrated image (processed by third-party software) which has negative numbers, I get a message: "FITS error: Numerical overflow during type conversion while reading FITS data."
Also, Siril cannot open FITS with bitpix=-32 (floating point) even if the actual range of data is within 0..65535 (Siril ignores bitpix?)

So, honestly, the program currently has many limitations which are critical to me (well, it's only my opinion, not a requirement nor criticism

)

Best regards,
Maksym

lock042

Цитувати

Regarding photometry tool etc: IMHO more important task is to solve a problem with possible negative values (due to noise) in frames after removing offset (bias).

For the tool I want to introduce, it is not critical. What do you want is something light + pedestal - offset during preprocessing ? And at the end we remove pedestal value ? Negative would be truncated again.

Цитувати

Moreover, if I try to open a calibrated image (processed by third-party software) which has negative numbers, I get a message: "FITS error: Numerical overflow during type conversion while reading FITS data."

Share this kind of FITS, we can't test all files, so we need example. This message is not a message of Siril but an error of the lib: cfitsio. So I need to check it.

Цитувати

Also, Siril cannot open FITS with bitpix=-32 (floating point) even if the actual range of data is within 0..65535 (Siril ignores bitpix?)

Again we need example to try to open this kind of file. Of course Siril take care of bitpix, and a of BZERO, BSCALE that are sometime forgoten by the software which produce the FITS file.

mpyat2

Цитата: lock042 від 23 Жовтня 2018, 21:36:08

Again we need example to try to open this kind of file. Of course Siril take care of bitpix, and a of BZERO, BSCALE that are sometime forgoten by the software which produce the FITS file.

Hi Cyril,

Here are examples of FITS files https://drive.google.com/drive/folders/1w4U74SgrCi7M4nC7WmhaqLIma1zq7KiZ

c1.fit: calibrated (using IRIS). It contains several negative pixels and cannot be read by Siril
light1.fit: image before calibration converted from CR2. This file reads normally.
light1float.fit: the same as previous, resaved by FITSwork4 using floating-point values (BITPIX=-32). It cannot be opened by Siril normally: after opening I see white image, statistics shows that all pixels have UINT_MAX value (65535).

Regards,
Maksym

lock042

Цитувати

light1float.fit: the same as previous, resaved by FITSwork4 using floating-point values (BITPIX=-32). It cannot be opened by Siril normally: after opening I see white image, statistics shows that all pixels have UINT_MAX value (65535).

I just fixed this one. It is because to open these kind of file, the keyword DATAMAX was needed. Indeed, it is very difficult to know if this is range [0, 1] or other. So now, I compute stats before opening files.

lock042

Цитувати

c1.fit: calibrated (using IRIS). It contains several negative pixels and cannot be read by Siril

About this file. I can now open it as PixInsight do, i.e, with minval = 32726 (32768 - 42) instead of -42

mpyat2

Цитата: lock042 від 24 Жовтня 2018, 23:07:06

Цитувати
c1.fit: calibrated (using IRIS). It contains several negative pixels and cannot be read by Siril

About this file. I can now open it as PixInsight do, i.e, with minval = 32726 (32768 - 42) instead of -42

Yet this file conforms FITS standards (see, for example, https://fits.gsfc.nasa.gov/fits_primer.html), BITPIX=16 corresponds to 16-bit signed integers.
It can be opened smoothly using IRIS, FITSwork4, AstroImageJ, AstraImage, XNViewMP, GIMP 2.10, MuniWin, and even PixInsightLE! (see picture)
Probably by other programs too, I checked those installed in my PC.

P.S. By the way online FITS Viewer http://www.clearskyinstitute.com/fits/fits.html opens it too.

lock042

And siril too

mpyat2

Цитата: lock042 від 25 Жовтня 2018, 10:24:24

And siril too

I mean c1.fit example having negative values. I have Siril 0.9.9 installed, it cannot open it. Should I update?

lock042

If you do use the version of the svn server, yes.
If not you will have to wait the 0.9.10.

Statistics of the c1.fit in PixInsight and Siril. You can see the shift of 32768 in order to preserve negative value.

Новини:

Автор Тема: Обробка астрофото із Siril (Прочитано 8373 раз)