Decoder metadata interface #2672

197g · 2025-11-26T22:26:52Z

See #2245, the intended ImageDecoder changes.

This changes the ImageDecoder trait to fix some underlying issues. The main change is a clarification to the responsibilities; the trait is an interface from an implementor towards the image library. That is, the protocol established from its interface should allow us to drive the decoder into our buffers and our metadata. It is not optimized to be used by an external caller which should prefer the use of ImageReader and other inherent methods instead.

This is a work-in-progress, below motivates the changes and discusses open points.

ImageDecoder::peek_layout encourages decoders to read headers after the constructor. This fixes the inherent problem we had with communicating limits. The sequences for internal use is roughly:

let mut decoder = …;
decoder.set_limits(…); // Other global configuration we have?

{ // Potentially multiple times:
  let layout_info = decoder.peek_layout()?;
  let mut buffer = allocate_for(&layout_info);
  decoder.read_image(&mut buffer)?;
  decoder.xmp_metadata()?; // and other meta
}

// … for sequences, start again from `peek_layout()`

ImageDecoder::read_image(&mut self) no longer consumes self. We no longer need the additional boxed method and its trait work around, the trait is now dyn-compatible.

Discussion

mstoeckl · 2025-11-27T14:39:51Z

The main change is a clarification to the responsibilities; the trait is an interface from an implementor towards the image library. That is, the protocol established from its interface should allow us to drive the decoder into our buffers and our metadata. It is not optimized to be used by an external caller which should prefer the use of ImageReader and other inherent methods instead.

With this framing, I think Limits::max_image_width and Limits::max_image_height no longer need to be communicated to or handled by the ImageDecoder trait, because the external code can check ImageDecoder::dimensions() before invoking ImageDecoder::read_image(); only the memory limit (Limits::max_alloc) is essential. That being said, the current way Limits are handled by ImageDecoder isn't that awkward to implement, so to reduce migration costs keeping the current ImageDecoder::set_limits() API may be OK.

fintelia · 2025-11-27T18:04:35Z

A couple thoughts...

I do like the idea of handling animation decoding with this same trait. To understand, are you thinking of "sequences" as being animations or also stuff like the multiple images stored in a TIFF file? Even just handling animation has some tricky cases though. For instance in PNG, the default image that you get if you treat the image as non-animated may be different from the first frame of the animation. We might need both a read_image and a read_frame method.

The addition of an init method doesn't seem like it gains us much. The tricky part of our current new+set_limits API is that you get to look at the image dimensions and total output size in bytes when deciding what decoding limits to set. Requiring init (and by extension set_limits) to be called before reading the dimensions makes it basically the same as just having a with_limits constructor.

197g · 2025-11-27T19:06:55Z

Requiring init (and by extension set_limits) to be called before reading the dimensions makes it basically the same as just having a with_limits constructor.

It's a dyn-compatible way that achieves the goal of the constructor so it is actually an abstraction.

The tricky part of our current new+set_limits API is that you get to look at the image dimensions and total output size in bytes when deciding what decoding limits to set.

What do you by this? The main problem in png that I'm aware of is the lack of configured limits for reading the header in the ImageReader path, that motivated the extra constructor in the first place. With png we can not modify the limits after the fact but also we don't really perform any large size-dependent allocation within the decoder.

I'm also not suggesting that calling set_limits after the layout inspection would be disallowed but obviously is decoder dependent on whether that 'frees' additional capacity. I guess if that is sufficient remains to be seen? When we allocate a buffer (with applied allocator limits)´ that allows forwarding the remaining buffer size to the decoder. Or, set aside a different buffer allowance for metadata vs. image data. Whatever change is necessary in png just comes on top anyways, the init flow just allows us to abstract this and thus apply it with an existing Box<dyn ImageDecoder> so we don't have to do it all before. Indeed, as the comment on size eludes to we may want to different limit structs: one user facing that we use in ImageReader and one binding-facing that is passed to ImageDecoder::set_limits. Then settings just need to be

197g · 2025-12-04T12:17:45Z

@fintelia This now includes the other changes including to ImageReader as a draft of what I meant in #2679 (comment). In short:

The file guessing routines and the construction of the boxed decoder are split into a separate type, ImageFile, which provides the previous methods of ImageReader but also an into_reader for the mutable, stateful interface.
ImageReader:
- features all accessors for metadata considering that some formats fill said metadata (or a pointer to it) after an image is decoded.
- has viewbox as a re-imagining of the previous ImageDecoderRect trait but split into two responsibilites: the trait does the efficiency decision on an image-by-image basis with an interface that allows a partial application of the viewbox (in jpeg and tiff we would decode whole tiles); then the reader takes care of translating that into an exact layout. Note that another type of image buffer with offset+rowpitch information could do that adjustment zerocopy—I still want to get those benefits of the type erased buffer/image-canvas someday and this fits in.
The code also retrieves the CICP from the color profile and annotates the DynamicImage with it where available. For sanity's sake the moxcms integration was rewritten to allow a smaller dependency to be used here, I'll split these off the PR if we decide to go that route.
Conceivably there's a gain_map (or similar) that may be queried similar to the metadata methods. For that to be more ergonomic I'd like to seriously consider read_plane for, in tiff lingo, planar images as well as associated and non-associated mask data; and more speculatively other extra samples that are bump maps? uv? true cmyk?. While that does not necessarily all go into 1.* for any output that is not quite neatly statically sorted and sized as an Rgba 4-channel-homogeneous-host-order, I imagine it will be much simpler for a decoder to provides its data successively in multiple calls instead of a contiguous large byte slice. Similar to viewbox we'd allow this where ImageReader provides the compatibility to re-layout the image for the actual user—except where explicitly instructed. Adjusting ImageReader::decode to that effect should be no problem in principle.

RunDevelopment · 2025-12-06T13:41:50Z

I can't speak about image metadata, but I really don't like the new ImageDecoder interface as both an implementor of the interface and a potential user of it. Right now, it's just not clear to me at all how decoders should behave. My problems are:

init. This is just two-phase initialization and opens so many questions.
- Do users have to call it? The docs say "should" not "must" be called before read_image.
- Are users allowed to call it multiple times? If so, the decoder has to keep track of whether the header has already been read.
- Since init returns a layout, what's the point of dimensions() and color_type()? And what if they disagree?
- What should dimensions and co do before init is called?
- If init fails, what should happen to methods like dimensions and read_image? When called, should they panic, return an error, return default values?
- After calling read_image, do you have to re-init before calling read_image again?
viewbox makes it more difficult to implement decoders.
- Now they always have to internally keep track of the viewbox rect if rect decoding is supported.
- After calling viewbox, what should dimensions be? If they should be the viewbox size, should they reflect the new viewbox even before calling init?
- It's not clear what should happen if viewbox returns ok, but init errors.
- What should happen if users supply a viewbox outside the bounds of the image?
When calling viewbox, is the offset of the rect relative to the (0,0) of the full image or the last set viewbox?
What should happen if read_image is called twice? Should it read the same image again, error, read the next image in the sequence? The docs don't say.
- If the intended behavior is "read the same image again", then those semantics would force all decoders to require Seek for the reader (or keep an in memory copy of the image for subsequent reads). Not an unreasonable requirement, but it should be explicitly documented.

Regarding rectangle decoding, I think it would be better if we force decoders to support arbitrary rects. That's because the current interface is actually less efficient by allowing decoder to support only certain rects. To read a specific rect that is not supported as is, ImageReader has to read a too-large rect and then crop the read image, allocating the memory for the too-large image only to throw it away. It is forced to do this, because of the API.

However, most image formats are based on lines of block (macro pixels). So we can do a trick. Decode a line according to the too-large rect, and then only copy the pixels in the real rect to the output buffer. This reduces the memory overhead for unsupported rects from O(width*height) to O(width*block_height). Supported rects don't need this dance and can decode into the output buffer directly. I.e. that's kinda what DDS does.

And if a format can't do the line-based trick for unsupported rects, then decoders should just allocate a temp buffer for the too-large rect and then crop (=copy what is needed). This is still just as efficient as the best ImageReader can do.

For use cases where users can use rowpitch to ignore the exccess parts of the too-large rect, we could just have a method that gives back a preferred rect, which can be decoded very efficiently.

So the API could look like this:

trait ImageDecoder {
    // ...
    /// Returns a viewbox that contains all pixels of the given rect but can potentially be decoded more efficiently.
    /// If rect decoding is not supported or no more-efficient rect exists, the given rect is returned as is.
    fn preferred_viewbox(&self, viewbox: Rect) -> Rect {
        viewbox // default impl
    }
    fn read_image_rect(&mut self, buf, viewbox) -> ImageResult {
        Err(ImageError::Decoding(Decoding::RectDecodingNotSupported)) // or similar
    }

This API should make rect decoding easier to use, easier to implement, and allow for more efficient implementations.

197g · 2025-12-07T20:06:18Z

init. This is just two-phase initialization and opens so many questions.

That was one of the open questions, the argument you're presenting makes it clear it should return the layout and that's it. Renamed to next_layout accordingly. I'd like to remove the existing dimensions()/color_type methods from the trait as well. There's no point using separate method calls for communicating them.

For use cases where users can use rowpitch, […]

That is ultimately the crux of the problem. I'd say it's pretty much the only problem even though that does not appreciate the complexity. A lot of what you put forth is overly specific to solving one instance of it, obviously focusing on DDS. That's not bad but take a step back to the larger picture. There's no good way to communicate all kinds of layouts that the caller could handle: tiled, planar, depths, sample types …. With the information being exchanged right now, no-one can find a best-match between the requirements of image's data types (and Limits) and what the decoder can provide. This won't be solved by moving complexity into the decoders, we need to get structured information out of them primarily, then make that decision / handling the resulting byte data in image's code.
1. viewbox makes it more difficult to implement decoders.
The point of the default implementation in this PR is that it is purely opt-in. Don't implement the method for decoders that can not provide viewbox decoding and everything works correctly. The documentation seems to be confusing, point taken. We're always going to have inefficiencies, I'm for working through the distinct alternative layouts that allow an optimization one-by-one. More importantly for this PR immediately is what outcome a caller may want and what interface would give it to them—in this case I've worked on the use-case of extracting part of an atlas.
However, most image formats are based on lines of block (macro pixels). So we can do a trick.

I'm not designing anything in this interface around a singular "trick", that's the wrong way around. That is how we got here. That's precisely what created ImageDecoderRect, almost to the dot. Falsehoods programmer's assume about image decoding will lead that to this breaking down and to be horrible to maintain. The trick you mention should live in the decoder's trait impl and nowhere else and we can bring it back where appropriate and possible. (Note that if you do it for a specific format, some formats will be even more efficient and not require you decode anything line-by-line but skip ahead, do tiles, … That's just to drive home the point that you do not want to do this above the decoder abstraction but below it in the ImageDecoder impl).
It is forced to do this, because of the API.

The decoder impls is only forced to do anything if we force it via an interface—this PR does not; read_image_rect(&mut self, buf, viewbox) does force a decoder to be able to handle all possible viewboxes—this PR does not. I'm definitely taking worse short-term efficiency over code maintenance problems—the latter won't get us efficiency in the long run either.

When calling viewbox, is the offset of the rect relative to the (0,0) of the full image or the last set viewbox?

It's suppose to be to the full image. Yeah, that needs more documentation and pointers to the proper implementation.

fintelia · 2025-12-06T21:59:50Z

src/io/image_reader_type.rs

+        self.limits.check_dimensions(layout.width, layout.height)?;
+        // Check that we do not allocate a bigger buffer than we are allowed to
+        // FIXME: should this rather go in `DynamicImage::from_decoder` somehow?
+        self.limits.reserve(layout.total_bytes())?;


This has a subtle behavior change. Previously we reserved the space for the output buffer and then passed the reduced memory limit onto the decoder to limit its metadata/scratch buffer allocations. But at this point in the code we've already told the decoder that it is free to use the entire memory limit amount for those purposes. So we effectively only limiting memory to 2x the requested amount.

That's true. The quicker fix here is to pass limits only once we have gotten the layout which would be closest to what we're doing right now on the main branch but obviously not really address if decoders are able to handle it. In gif we .. duplicate the allocation arbitrarily, right now, too:

image/src/codecs/gif.rs

Lines 172 to 174 in 7f44147

self.limits.reserve_usize(buffer_size)?;

let mut frame_buffer = vec![0; buffer_size];

self.limits.free_usize(buffer_size);

We must rework the limits system regardless if it is to work with animations, too. That is to say at some point the struct will have been moved into the decoder but some higher level part of IO is still going to do allocations; but also de-allocations (!). Should those count? How do we access the limits if so?

For consuming animations you'd typically have a swapchain-like fixed capacity queue setup. The decoder would consume some of the limits while allocating for each new frame, then those limits should get refreshed over time when the frames have been fully consumed and dropped.. We can't fully handle this either right now (as seen in gif).

I've ported this to taking half the allocation for ImageReader instead. That works reasonably well and we can consistently do it (maybe even more explainable now). One problem with the interface that remains from the previous iteration is that there is no guidance for dealing with multiple calls to it, when some allocations are already consumed.

src/io/image_reader_type.rs

fintelia · 2025-12-07T20:21:16Z

src/io/image_reader_type.rs

    }
 }
+
+impl ImageReader<'_> {


nit: it is easier for scrolling through the file if the impl blocks for each struct immediately follow the definitions

fintelia · 2025-12-07T22:26:03Z

src/io/decoder.rs

+    /// The decoder should attempt to restrict the amount of decoder image data to the indicated
+    /// box. It should return `Ok` if it supports the viewbox directly, or `Err` if it does not
+    /// support that exact viewbox. In the either case the decoder must indicate the new image
+    /// dimensions in subsequent calls to [`Self::init`].


Which formats do we expect to be able to implement viewboxes?

TIFF should be able to do it by only reading the necessarily strips/tiles. And DDS can because GPU compressed formats are designed for random access. But for other compressed formats, I think it might be rather annoying?

Even jpeg could profit. We'll need to be more precise in the description but I had intended that, in either return case, the values outside the initially indicated viewbox will be allowed to be treated as 'dont-care' by the decoder in their original (or only partially cropped) layout. For these pixels the decoder not need present the actual underlying image data.

With some integration that would allow wins beyond just a binary decoding decision. Some form of integration would be possible for:

all the PAM varieties, bmp, tga save on IO with full support; for small viewboxes if we Seek/read_at

in jpeg we need not process the data from any of the MCU's that do not intersect the data even if the bitstream must be decoded to align with the file

avif and webp seem similar to that where they might avoid processing values (obviously needing integration of it).

For anything non-RGB we get to save on conversion, too. I expect we uncover different forms of gains once we have a better understanding of a declarative form.

fintelia · 2025-12-07T22:29:08Z

src/io/image_reader_type.rs

+        self.viewbox = Some(viewbox);
+    }
+
+    /// Get the previously decoded EXIF metadata if any.


The "previously decoded" part here makes me a bit nervous. I think we'll want to be clearer about what the user has to do to make sure they don't get None for images that actually do have EXIF data

This is a bit odd and will depend on the format and metadatum. For instance, XMP is encoded per-image in tiff but only once in gif (despite us representing this as an image sequence) and also only once in png (no word about APNG). Problematically, in gif and png the standard requires absolutely no ordering with any of the other chunks. So it might be encountered before all of the header information is done; or after all the images have been consumed.

The problem with back references is of course the unclear association. And when multiple are included we always have a problem with consuming them 'after the end' since it should need to be buffered or the decoder able to store seek-back points (like png). Encoding all that in the interface is a challenge, i will incur some unavoidable complexity.

src/io/decoder.rs

197g · 2025-12-16T19:17:10Z

Resolving the naming question as peek_layout, hopefully satisfactory for now.

This was one of the silently broken features when adding the new methods. While reviewing how to address the underlying cause with a clippy fix it became obvious a few other methods were missing, too.

Change ImageDecoder::read_image to `fn(&mut self)`. There are some codecs where metadata and other information is only available after decoding. Add ImageDecoder::init for codecs where reading the header needs to take into account limits and other configuration we are to introduce. This sketch compiles as: `--no-default-features --features=png`

Renames it `peek_layout` to clarify the relation to `read_image`, that it is to be called potentially multiple times for each such call. Document the trait more with this relationship, moving some methods around to clarify.

The latter interface will be something that interfaces and interacts with the underlying decoder. Introduces ImageDecoder::set_viewport as a stand-in for what is intended here: color, subsampling, gain map application will take more interaction, clarification of how the library will request their data, and adjustments to the layout definition by the decoder.

The purpose of the trait is to be careful with the boundary to the `moxcms` dependency. We use it because it is a quality implementation but it is heavy weight for what we need. There's other possible ways to provide transfer functions and color space transforms. Now this also introduces ICC profile parsing but again that could be done with a *much* lighter dependency as we only need basic information from it. The trait should make every little additional cross-dependency a conscious decision. Also it should be the start of a customization point, by feature flag or actually at runtime.

No longer responsible for ensuring the size constraints are met under the new policy and with available of constructing a reader from an instance of a boxed decoder.

This allows us to write a generic iterator which uses the same decoder function to generate a whole sequence of frames. The attributes are designed to be extensible to describe changes in available metadata as well, very concretely some formats require that XMP/ICC/… are polled for each individual image whereas others have one for the whole file and put that at the end. So there's no universal sequence for querying the metadata, and we need to hold runtime information. This will be the focus of the next commit.

197g · 2025-12-22T22:17:18Z

@fintelia I understand this is too big for a code-depth review but I'd be interested in the directional input. Is the merging of 'animations' and simple images as well as the optimization hint methods convincing enough? Is the idea of returning data from read_image something that works for you? The struct is meant to be Default-able and fills in the information for into_frames() but I'll sketch out some way of putting the metadata indicators in there (i.e. should you poll xmp for this frame, or wait until the end, or is it constant of the file).

As an aside, in wondermagick we basically find that sequence encoding is a missing API to match imagemagick. We can currently only do this with gif, despite tiff and webp having absolutely no conceptual problems with it (avif too, but imagemagick behaves odd, does not match libavifs decoding and the rust libraries don't provide it). It would be nice to make those traits symmetric so the direction here influences the encoding, too.

RunDevelopment mentioned this pull request Nov 28, 2025

Suggestions: New ImageDecoder and ImageSequenceDecoder APIs #2679

Open

This was referenced Nov 30, 2025

Rip out ImageDecoderRect #2681

Merged

Prevent OOM in ICO due to PNG #2682

Merged

197g force-pushed the decoder-metadata-interface branch from 86c9194 to cdc0363 Compare December 7, 2025 18:22

fintelia reviewed Dec 7, 2025

View reviewed changes

RunDevelopment reviewed Dec 9, 2025

View reviewed changes

src/io/decoder.rs Outdated Show resolved Hide resolved

This comment was marked as outdated.

Sign in to view

This comment was marked as resolved.

Sign in to view

197g force-pushed the decoder-metadata-interface branch 2 times, most recently from 1a114c3 to 306c6d2 Compare December 16, 2025 18:40

This was referenced Dec 18, 2025

Add a -write option Shnatsel/wondermagick#77

Open

Add {ImageBuffer,DynamicImage}::crop_in_place #2701

Merged

197g force-pushed the decoder-metadata-interface branch 6 times, most recently from fa259bb to 46487ad Compare December 20, 2025 23:08

197g added 4 commits December 22, 2025 18:43

Ensure ICO forwards all methods

27d15b3

This was one of the silently broken features when adding the new methods. While reviewing how to address the underlying cause with a clippy fix it became obvious a few other methods were missing, too.

Migrate all codecs to init with layout

9ff8d1e

Renames it `peek_layout` to clarify the relation to `read_image`, that it is to be called potentially multiple times for each such call. Document the trait more with this relationship, moving some methods around to clarify.

197g added 7 commits December 22, 2025 18:50

Provide limited metadata decoding with ImageReader

f000de7

Provide ImageReader::skip

cd804c4

Rename ImageFile -> ImageReaderOptions

c71b611

Fixup gif testing

eb03633

No longer responsible for ensuring the size constraints are met under the new policy and with available of constructing a reader from an instance of a boxed decoder.

Adjust tests for ImageReader limits usage

4325060

197g force-pushed the decoder-metadata-interface branch from e8d2713 to 4325060 Compare December 22, 2025 17:54

197g marked this pull request as ready for review December 22, 2025 22:17

	self.limits.reserve_usize(buffer_size)?;
	let mut frame_buffer = vec![0; buffer_size];
	self.limits.free_usize(buffer_size);

Decoder metadata interface #2672

Are you sure you want to change the base?

Decoder metadata interface #2672

Uh oh!

Conversation

197g commented Nov 26, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Discussion

Uh oh!

mstoeckl commented Nov 27, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Uh oh!

fintelia commented Nov 27, 2025

Uh oh!

197g commented Nov 27, 2025

Uh oh!

197g commented Dec 4, 2025

Uh oh!

RunDevelopment commented Dec 6, 2025

Uh oh!

197g commented Dec 7, 2025

Uh oh!

fintelia Dec 6, 2025

Choose a reason for hiding this comment

Uh oh!

197g Dec 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

197g Dec 21, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

Uh oh!

fintelia Dec 7, 2025

Choose a reason for hiding this comment

Uh oh!

fintelia Dec 7, 2025

Choose a reason for hiding this comment

Uh oh!

197g Dec 8, 2025 • edited Loading Uh oh! There was an error while loading. Please reload this page.

Uh oh!

Choose a reason for hiding this comment

Uh oh!

fintelia Dec 7, 2025

Choose a reason for hiding this comment

Uh oh!

197g Dec 16, 2025

Choose a reason for hiding this comment

Uh oh!

Uh oh!

This comment was marked as outdated.

This comment was marked as resolved.

This comment was marked as resolved.

197g commented Dec 16, 2025

Uh oh!

197g commented Dec 22, 2025

Uh oh!

Reviewers

Assignees

Labels

Projects

Milestone

Development

Uh oh!

5 participants

197g commented Nov 26, 2025 •

edited

Loading

mstoeckl commented Nov 27, 2025 •

edited

Loading

197g Dec 8, 2025 •

edited

Loading

197g Dec 21, 2025 •

edited

Loading

197g Dec 8, 2025 •

edited

Loading