The for loop in apply_preset_5 was not including 9.
apply_preset_6 now builds now calls apply_preset_4 instead of
apply_preset_5, and adds all compression levels from 1 to 9.
Co-authored-by: Nino Burini <nburini@jabra.com>
For RGB(A) images that contain gray colors, this reduction can achieve
significant space savings. However, in the absence of gamma correction
data, some PNG decoders assume more exotic color spaces for grayscale
images instead of the ubiquitous sRGB. This results in gamma
miscorrection, and for the end user this means that colors will look
wrong, like "washed-out". Java's ImageIO class, which is popular in the
JVM world to read PNG files, uses rather unconventional defaults, as
explained in this StackOverflow question: https://stackoverflow.com/questions/31312645/java-imageio-grayscale-png-issue
Gamma miscorrection problems aside, OxiPNG currently tries hard to
reduce RGB(A) images to grayscale, because it expects that reduction to
be quite effective. However, in some cases, OxiPNG generates smaller
PNG files when reducing grasycale RGB(A) images to paletted color than
actual grayscale color. For example, let's say that "~/gray.png" is a
256x256 RGBA image entirely filled with (119, 119, 119, 255) pixels.
OxiPNG, by default, reduces this image to grayscale and achieves a
68.23% decrease:
$ cargo build --release && target/release/oxipng -omax --out ~/out.png ~/gray.png
Processing: /home/user/gray.png
256x256 pixels, PNG format
4x8 bits/pixel, RGBA
IDAT size = 604 bytes
File size = 661 bytes
Reducing image to 1x4 bits/pixel, Grayscale
Trying: 144 combinations
Found better combination:
zc = 6 zs = 0 f = 0 153 bytes
IDAT size = 153 bytes (451 bytes decrease)
file size = 210 bytes (451 bytes = 68.23% decrease)
Output: /home/user/out.png
However, if the --ng option that this commit adds is used to skip the
grayscale reduction step, OxiPNG reduces to a single color palette
instead, which is much more efficient, achieving a 84.42% decrease:
$ cargo build --release && target/release/oxipng -omax --ng --out ~/out.png ~/gray.png
Processing: /home/alejandro/gray.png
256x256 pixels, PNG format
4x8 bits/pixel, RGBA
IDAT size = 604 bytes
File size = 661 bytes
Reducing image to 1 bits/pixel, 1 colors in palette
Trying: 144 combinations
Found better combination:
zc = 3 zs = 3 f = 0 31 bytes
IDAT size = 31 bytes (573 bytes decrease)
file size = 103 bytes (558 bytes = 84.42% decrease)
Output: /home/alejandro/out.png
While OxiPNG should arguably be made smarter to better handle these
cases, in the meantime, adding an option to manually skip that grayscale
reduction can't hurt. In fact, it may even help users achieving the most
out of current versions of OxiPNG, and developers reasoning about what
makes a grayscale-like RGB(A) image compress better with a color
palette.
Due to the reasons stated above, this adds a simple "grayscale_reduction"
option to the Options struct, and a "no-grayscale-reduction" command
line switch, that makes OxiPNG not try this problematic grayscale
reduction on RGB(A) images.
* Switch to crossbeam-channel + rayon::spawn
* Remove thread_spawn for evaluation altogether
This allows to avoid a deadlock when there is only one Rayon thread, and doesn't sacrifice performance, since the caller of .get_result() had to always block on the iterator to be finished anyway, and all the messages are already sent from separate threads.
* Fix `verbose_mode` test
This one is easier to "fix", since we're in control of the rayon pool - just adding one extra thread to the default number to make sure we always can one root "spawn" call without stealing from the actual pool.
* Update --help to exclude -a
* Add a deprecation warning to level 4 constructor
* Initialise logger earlier
* Add warning for level > 3 for non-zlib
It's not obvious immediately that these levels don't have any effect on libdeflater and Zopfli, since they don't iterate over zlib-specific fine-tuned options.
Hence, show warning so that user knows they're getting "downgraded" to level 3.
* Add "max" level alias; more level warnings
* Update --help trial numbers for non-zlib
* Fix incorrect trial numbers
This allows to configure or compile away logging in the library from a single place in Rust apps.
For the CLI side, the usage and output remained the same, except it's now colour-coded.
Fixes#217.
Follow-up to #210.
I haven't noticed / forgotten that clap has own mechanism for conflicts between arguments, and it's probably best to use it instead of custom checks.
Rayon uses a singleton global pool.
By default it's set to a regular spawn handler with number of logical CPU cores, but it can be overridden by Rust applications to customize number of threads, spawn handlers, exit handlers and other options.
Such customization should be usually done at the app level, because if a single library initialises the global pool, then Rayon will prevent any further overrides and they will error out. This can cause conflicts between libraries or library and user code and make them impossible to use together.
Hence, I've removed the `threads` option from the `Options` struct and instead moved initialisation to the CLI part of the codebase (main.rs).
Users of the library that didn't depend on custom `threads` number can keep using it as before - they'll still get same number of threads as number of logical CPU cores, while users who need fine-tuning, can do that by customizing rayon pool themselves at the top level of the app.
Note: another alternative to keep the option could've been to use `ThreadPoolBuilder::build` + `ThreadPool::install` to use a local pool just within OxiPNG, but that would ignore any customizations made by users in top-level pool and would prevent usage on targets that require custom spawn handlers like WebAssembly. As such, I've decided to avoid it.
Co-authored-by: Josh Holmer <jholmer.in@gmail.com>
In order to avoid problems with mutable references,
the approach is to use the eval sender for non-parallel.
Computations will still be done synchronously
with the fake synchronous rayon module.
* Skip baselines earlier in the evaluator
We know that we'll throw them away anyway, so there is no point in even sending them to Sender and comparing them with others once we used them for the best candidate size.
* Simplify comparison logic
* Improve performance consistency
Switch from HashMap / HashSet to IndexMap / IndexSet for consistent iteration order of various options and, as a result, more predictable performance.
libdeflater is a Rust wrapper around
[libdeflate](https://github.com/ebiggers/libdeflate) - an alternative
heavily optimised library for deflate/zlib/gzip compression and
decompression that is intended for situations where upper bounds of the
output are well-known.
In my benchmarks on test files in the repo it has shown to be usually
both slightly faster and providing better compressed output than
cloudflare-zlib, but in some cases showing the opposite, so rather
than swapping defaults, it's currently provided as another option,
similarly to zopfli.
Since it's not strictly better in all cases, I'm not providing median
numbers, but you can check distribution histograms for time and size
differences here (all using `oxipng -o 6 -t 6 -P`):
https://docs.google.com/spreadsheets/d/1WOKgeYZBhLkQvMGAC36snN4azilElzOFhx63RJu0EZY/edit?usp=sharing
* Avoid using time API when we don't need it
This avoids a syscall to the time API when the result is ignored later anyway.
This allows to use the library with default options on wasm32-unknown-unknown, where the unimplemented syscall would panic otherwise.
* eval_send doesn't need to be an Option
We can drop the value manually, thus avoiding unwrap on each access.
* Keep single `use rayon::prelude::*`
If either `rayon` is already imported, then `rayon::prelude::*` should always resolve.
* Extract comparator
* Fully enable non-parallel mode
* Fix verbose message
* No cloning when restoring original data
* Make reductions return a new uncompressed image
Partially fixes#145
* Async reduction evaluator
* Assert
* Faster bit depth check
* Also try 4-bit depth for small-depth images
* Skip test when using miniz
* Ensure palette is trimmed after depth reduction
Fixes#159
* Fudge factor for reductions to prefer better reductions even if gzip estimation says otherwise
* Move reductions to a module. Make copy instead of changing in-place.
* Alpha reductions
* Immutable color reductions
* Immutable interlace reductions
* Compress alpha with strategy sensitive to repetitions
* Update reductions flag when alpha changed
* Avoid needlessly cloning out-of-date idat_data and temp raw_data
This reverts commit 2702145f3f.
cfzlib seems to have a MASSIVE memory leak in it. Trying to use it
when processing many files at once will cause your system to run out
of memory.
Closes#125
