VipsJpeg: Bogus JPEG colorspace -tiff2vips: decompress error tile 0 x 0

[petoor]

Hello.

I have a problem regarding loading a tiff image into memory.

img = pyvips.Image.new_from_file(vips_filename = "field.tif", access='sequential')
img.numpy() # Throws   VipsJpeg: Bogus JPEG colorspace -tiff2vips: decompress error tile 0 x 0
# img.colourspace("srgb").numpy() - # also doesnt work

The image is readable by for instance Dask using

from dask.array.image import imread

Could you point me to what im doing wrong?
Versions:
vips: vips-8.16.0 (build from source)
pyvips : 2.2.3
Running Ubuntu 2024.04

The image in question :
Created by open drone map
field_image.zip

[petoor]

For good measure, I've made a dockerfile where i can reproduce the error.
(Using the packaged libvips version (8.15.2))

Create file pyvips_test.py and copy the field.tif image into same folder.
**pyvips_test.py **

import pyvips
img = pyvips.Image.new_from_file(vips_filename = "field.tif", access='sequential')
img.numpy() # Throws   VipsJpeg: Bogus JPEG colorspace -tiff2vips: decompress error tile 0 x 0
# img.colourspace("srgb").numpy() - # also doesnt work
print(img)

Dockerfile

RUN apt-get update && apt-get upgrade -y && apt-get -y install \
    build-essential ninja-build wget libglib2.0-dev libjpeg-turbo8-dev libtiff5-dev libarchive-dev libvips python3-pip

RUN python3 -m pip install numpy pyvips --break-system-packages

COPY pyvips_test.py pyvips_test.py
COPY field.tif field.tif

ENTRYPOINT ["python3", "pyvips_test.py"]

Building
docker build -t pyvipstest .

Finally running:
docker run pyvipstest

Yields

pyvips.error.Error: unable to write to memory
  VipsJpeg: Bogus JPEG colorspace
tiff2vips: decompress error tile 0 x 0
VipsJpeg: Bogus JPEG colorspace
tiff2vips: decompress error tile 512 x 0
VipsJpeg: Bogus JPEG colorspace
tiff2vips: decompress error tile 1024 x 0

[jcupitt]

Hello @petoor,

I tried with tiffcp:

$ tiffcp field.tif x.tif
TIFFReadDirectory: Warning, Unknown field with tag 33550 (0x830e) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 33922 (0x8482) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 34735 (0x87af) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 34737 (0x87b1) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 42112 (0xa480) encountered.
JPEGLib: Bogus input colorspace.
x.tif: Error, can't write tile at 0 0.

So your image can't be read by libtiff. I suppose dask have their own tiff reader which allows a bigger range of colourspaces. Or maybe this is something specific to GDAL?

I tried tiffinfo as well:

$ tiffinfo field.tif 
TIFFReadDirectory: Warning, Unknown field with tag 33550 (0x830e) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 33922 (0x8482) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 34735 (0x87af) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 34737 (0x87b1) encountered.
TIFFReadDirectory: Warning, Unknown field with tag 42112 (0xa480) encountered.
=== TIFF directory 0 ===
TIFF Directory at offset 0x8 (8)
  Image Width: 5506 Image Length: 6289
  Tile Width: 512 Tile Length: 512
  Bits/Sample: 8
  Sample Format: unsigned integer
  Compression Scheme: JPEG
  Photometric Interpretation: RGB color
  Extra Samples: 1<unassoc-alpha>
  Samples/Pixel: 4
  Planar Configuration: single image plane
  Tag 33550: 0.015538,0.015540,0.000000
  Tag 33922: 0.000000,0.000000,0.000000,562759.412575,6125026.769753,0.000000
  Tag 34735: 1,1,0,7,1024,0,1,1,1025,0,1,1,1026,34737,8,0,2049,34737,8,8,2054,0,1,9102,3072,0,1,32632,3076,0,1,9001
  Tag 34737: unknown|unknown|
  GDAL Metadata: <GDALMetadata>
  <Item name="DESCRIPTION" sample="0" role="description">red</Item>
  <Item name="DESCRIPTION" sample="1" role="description">green</Item>
  <Item name="DESCRIPTION" sample="2" role="description">blue</Item>
</GDALMetadata>

  JPEG Tables: (73 bytes)

It's a four-band image, I suppose RGBA, but with JPEG (RGB and CMYK only) compression. I don't think many TIFF readers will be able to handle this.

[petoor]

Hi @jcupitt

Thank you very much for pointing me in the right direction.

gdal_translate -co COMPRESS=JPEG -co TILED=YES

will return the bogus error, however, stripping alpha channel will result in a tif readable by pyvips.

gdal_translate -b 1 -b 2 -b 3 -co COMPRESS=JPEG -co TILED=YES
or for even better compression.
gdal_translate -b 1 -b 2 -b 3 -co COMPRESS=JPEG -co PHOTOMETRIC=YCBCR -co TILED=YES

In case anyone is having the same question in the future.

[petoor]

Or straight from the documentation : https://gdal.org/en/stable/programs/gdal_translate.html
gdal_translate rgba.tif withmask.tif -b 1 -b 2 -b 3 -mask 4 -co COMPRESS=JPEG -co PHOTOMETRIC=YCBCR
In case anyone want to keep the alpha channel

[jcupitt]

Ah, that's helpful. Thanks for the update @petoor