This should be noted in the Metadata Modul, but, the bigtiff branch is only for saving BigTIFF.
Opening a BigTIFF and work on the Metadata was already possible (or not impossible) before the branch, so the capability to save as a BigTIFF has nothing to do with opening a BigTIFF and working on metadata as it in was implemented before (without saving).
Oh it is very ok that you remember that. But one thing is the need to save as file and the other one to manipulate metadata. The one that needs both had to look how it is possible.
The one is BigTIFF and if i want to save 10GB as TIFF it must be BigTIFF.
The other one is manipulating metadata around, which we had to look how it works.
BigTiff is not supported by Exiv2. However it is discussed in my book and handled by the book’s code: tvisitor.cpp. Joris Van Damme of AWare Systems maintains the BigTiff website and was very helpful when I was working on the book. Image Metadata and Exiv2 Architecture
Adding support for BigTiff to Exiv2 will involve updating src/tiffimage.cpp to handle the uint16/uint32 and uint32/uint64 differences between 32-bit Tiff and 64-bit BigTiff.
My interest in BigTiff derives from discussions with Imperial College, London who have 100Gbyte multipage Tiff files generated by medical scanners.
Adding multi-page support to Exiv2 may be desirable, however, it will involve architectural changes which could ripple to other multi-page formats such as PNG and GIF.
Oh, that’s nothing special. These are simple panoramas, many photos stitched together into one large image. Just as the focus is often stacked in macro shots, the individual shots are stitched in panoramas. But I don’t do 360 degrees or aerial shots of places, I just take photos. So imagine you go into the forest and take a picture. And then you look at thick branches and perceive them as sharp. For me, the sharpness not only extends to the thickest branch, but also to the finest leaves and green moss.
These files are industrially usually handled by Photoshop PSB. As you are mostly taking RAW images for up to 14bit image quality the intermediate format while grading is mostly TIFF. At the end i export to PSB (where i will save my IPTC information).
With EXR and lossless compression, what is your suggestion for using RT?
I guess that’s where my curiosity for the need for significantly higher than uint16 bitdepths comes from. EXRs have at least half-float precision, which should surpass uint16 for most applications in my very humble opinion. It’s an open format, it supports multipages, it offers float and double precision for data-type pages, it has a variety of lossless compression options, it has lossy compression options…
It is far from a suggestion. I wanted to understand why EXRs are not used for the application at hand. Not sure if RT currently reads/writes EXRs! In compositing/vfx EXRs could be called a standard.
With intermediate files, storage size is probably not an issue, but as a storage format I get a bit uneasy when I hear 100GB per file.
Wow.
No worries. Maybe BigTIFF is more convenient for many peoples applications because of that. I was just baffled by the need for BigTIFF in the first place.
I’ll stop derailing this now. Thank you all for your insights!
TIFF can also contain half float no problem, as well as 24-bit float, but the compression options are limited (at least the standard ones).
Both EXR and TIFF are good for HDR work. EXR kinda got a foothold in the film and VFX industry (I guess they have a specialized workflow to handle metadata and timing data as separate streams/formats), TIFF everywhere else I think…