FUJIFILM DBP for GX680 RAF support about rawspeed HOT 26 OPEN

@LebedevRI
I will. Please allow me 2 days as it is overcast today and tomorrow.

from rawspeed.

Minor correction: RawDigger (beta version you already have) indeed supports this format. The only thing is not supported is RGB render (this is not supported for all cameras where sensor resolution does not match RGB render resolution)

from rawspeed.

That looks like uncompressed raw with some tiling.
But it does not look like just tiling, since all the 'tiles' on the same line appear to have the same content, so only the rows are different.
I'm guessing there is some kind of sub-samplin going on.

from rawspeed.

I'm having trouble finding pdf usermanual for that back.

The only thing is not supported is RGB render (this is not supported for all cameras where sensor resolution does not match RGB render resolution)

So there is more than one raw format?

from rawspeed.

RawDigger does not shows RGB rendering for 'non-square pixels' (pixel 'aspect ratio' is not 1.0), e.g. Nikon D1X, most Fuju SuperCCD cameras, some very old Kodaks. Only RAW display (composite and per-channel) is possible for such cameras.

The only cause for that is interface problem: RAW/RGB rendering are switchable via single button, there is no way to do that if RGB/RAW render size or orientation do not match each other.

from rawspeed.

@LibRaw i think there may be a translation issue, but to me this really looks like another very good, explanatory, but non-answer... :/

(i'm only looking at the sample we have at raw.pixls.us)
After staring at it a bit more, i think:

1. the camera orientation is wrong, it needs to be rotated -90deg
2. there are 8 columns and 8 rows.
3. the image was split into 8 strips (tile with width = full sensor width, height = 1/8 sensor width)
4. each strip is split further somehow, current guess is planar-like, i.e. given image
   | A | B |
   | C | D |
   We split it into:
   plane 0: | A |
   plane 1: | B |
   plane 2: | C |
   plane 3: | D |

from rawspeed.

Sorry that I delete the snapshot since I do not have the permission from Alex to publish it before testing.

from rawspeed.

Sorry that I delete the snapshot since I do not have the permission from Alex to publish it before testing.

I did not download it anyway.

from rawspeed.

Here is decoder:

nTiles = 8;
  tile_width = raw_width / nTiles;

  ushort *tile;
  tile = (ushort *)calloc(raw_height, tile_width * 2);

  for (tile_n = 0; tile_n < nTiles; tile_n++)
  {
    read_shorts(tile, tile_width * raw_height);
    for (scan_line = 0; scan_line < raw_height; scan_line++)
    {
      memcpy (&raw_image[scan_line * raw_width + tile_n * tile_width], &tile[scan_line * tile_width], tile_width * 2);
    }
  }
  free (tile);

from rawspeed.

@wtanmdt before i proceed with this issue, do you own that back?
I need an answer to #125 (comment)

from rawspeed.

@LebedevRI
Yes, I own that back.
Your help is really very appreciated.

from rawspeed.

Note that was only the first part of the question.
Is there more than one raw format selectable on that back?

from rawspeed.

@LebedevRI
There is only one raw format available on that back.

from rawspeed.

OK, thank you.

from rawspeed.

@LebedevRI
Thank you for your time and efforts.

from rawspeed.

@wtanmdt if you could contribute some better (than we currently have) sample to https://raw.pixls.us, that would be great, i don't really like the current one https://raw.pixls.us/getfile.php?type=preview&id=2069
Something like outside nature landscape would be best.

from rawspeed.

Is this photo better? @LebedevRI

from rawspeed.

Looks underexposed by one-two stops, but yes, better.
(RPU seems down at the moment :/)

from rawspeed.

Maybe this one better?

from rawspeed.

If there are more than one, just upload them all to https://raw.pixls.us, i will deduplicate them.

from rawspeed.

and this one

from rawspeed.

Thanks, deduplicated down to https://raw.pixls.us/getfile.php?type=preview&id=2574

from rawspeed.

After looking at the exif (e.g. diff), i'm not seeing any obvious tag that would identify this raw variant from the others :/
Maybe other than RawImageFullSize vs RawImageSize mismatch..

from rawspeed.

This is beyond my knowledge. Many thanks for your time. @LebedevRI

from rawspeed.

@wtanmdt could i get a pic of that back? i'm really interested in it

from rawspeed.

@wheropl here you are:

from rawspeed.