FAQ

Dithermark was originally inspired by Tanner Helland’s excellent introduction to image dithering. While it was easy to find example images for different dithering algorithms, there were very few, if any applications that let you interactively compare them using your own photos. After an initial MVP and many visits from the feature creature, Dithermark was born.

If you have a version of either Google Chrome, Mozilla Firefox, Apple Safari or Microsoft Edge release in the second half of 2017 or later, you are most likely good to go. More specifically, Dithermark has been tested to work with at least Chrome 63, Firefox 59, Safari 11.1 and Edge 16, though older versions might work as well. For absolute best experience, a laptop or desktop computer with Google Chrome or Mozilla Firefox is recommended.

We are committed to your privacy, which is why there is no tracking, cookies, telemetry or ads. No user data is stored or sent to our servers and everything that is saved is saved locally in your browser. Any images you open from your device remain private and are not sent to our servers. Once the page fully loads no additional web requests are made, the only exceptions being when you open images from the web. When opening an image from a URL, a normal web request is made, the same as if you opened the image url in your browser. When opening a random image, in accordance with Unsplash’s API Guidelines a request is sent to Unsplash for the image and when you export it a network request is sent to Unsplash to let them know you downloaded it. This request does not contain any identifying information about you or how you have modified the photo, as it only sends the photo id so the photographer can get credit. Dithermark is open source, so if you are excessively concerned about your privacy you can audit the source code or host it locally.

Yes! Dithermark is open source and released under the MIT license. You can get the source code at the Dithermark GitHub page.

First, read through this FAQ and see if it addresses your problem (particularly that you meet the minimum system requirements). If nothing helps, please open an issue on our GitHub issues page. Please make sure to include your operating system, operating system version, web browser and web browser version in your issue. While we appreciate you taking the time to help improve Dithermark, please be aware that the project is a free resource maintained by a single person, so it might be some time before your issue is addressed.

The random images are generously provided by Unsplash, courtesy of their free photo API.

Opening very large images on mobile devices has been found to crash the browser. Make sure the ‘Shrink large images’ option is checked in the ‘settings’ tab. If this still does not help, try opening a smaller image, such as by selected ‘Random image’ from the ‘open’ tab, and then pixelate the image by selecting the option in the ‘image’ tab (this has the same effect as shrinking the image). Now try reopening the image that caused your browser to crash.

If the web server for a website has not enabled CORS (Cross-Origin Resource Sharing) your browser will not let you open the image for security reasons. In that case, you should try downloading the image to your device and opening it from there.

WebGL is enabled by default as it leverages the power of your GPU to complete tasks 30-100 times faster. Dithermark uses WebGL (when enabled) for everything except the error propogation dithers and optimize palette functionality. The only times when it might be advisable to disable WebGL is if you are noticing a large amount of artifacts in the random dithers, or you are opening images that are too large to fit in a single WebGL texture unit (you will see a warning if this is the case). If you do not see an option to enable WebGL, that means your device/browser doesn’t support it. In that case all the dithering and color quantization features will still work though some of the image filters will be disabled.

The random dithers require random numbers, however without expensive, specialized hardware computers can only generate pseudo-random numbers. When using WebGL, the random numbers are even less random than usual since WebGL lacks a native random function, and so randomness is simulated using the sin() function. If you are seeing artifacts, the best solution is to re-run the dither by changing a parameter and changing it back. If you are seeing an excessive number of artifacts that are not fixed by this process you might want to consider disabling WebGL.

Fully opaque and fully transparent pixels should always look correct, however for browser that don’t support the WebGL premultipliedAlpha attribute, semi-transparent pixels can be rendered in weird colors. One workaround is to to disable WebGL in the settings tab or to use Google Chrome as your web browser. Alternatively, you can open the image in your image editor of choice and make all pixels either fully opaque or fully transparent. For example, to do this in GIMP you would open the image and then choose Layer→ Transparency→ Threshold Alpha.

Dithermark saves your settings and palettes locally in your browser using the browser localStorage API. That means if you clear your browser cache, your settings and palettes will be deleted as well. Also, if you open an incognito tab or change browsers, you will not be able to access your saved settings or palettes.

While the color limit is one sense arbitrary, it does have technical, artistic and user experience justifications. From a technical perspective, increasing the color limit decreases performance, as the the dithering algorithms and some of the color quantization algorithms get slower the larger the output palette is. From an artistic standpoint, it is generally easier (especially for beginners) to create interesting images with a limited color palette. From a user experience perspective, it gets increasingly more difficult to manually edit palette colors as the palette size increases, particularly on mobile. Nevertheless, Dithermark is open source, so if you would like to increase the color palette you can edit the code yourself, and the project provides documentation on how to do this.

It is recommended you use the png file type when exporting your images, as it avoids undesirable compression artifacts. However, in certain situations (such as when exporting a large, unpixelated image or using post dither image filters), exporting as a jpeg may result in a smaller file in exchange for a negligible loss in quality.

The resources page contains links where you can learn more about image dithering, color quantization and other related topics.