Canvas, an HTML API, renders 2D graphics and animation on a webpage. Moreover, the function of Canvas adds entropy for browser fingerprinting. As stated from the study of Englehardt and Narayanan (2016) from Princeton University, it revealed that above 5% of websites are using Canvas for browser fingerprinting.
To summarize, canvas fingerprinting requests browsers to make hidden canvas images. These images are moderately different on several computers. However, these images are the same as identical computers. Following the image is finished, browsers will transform into a hash string. After this, the features are further utilized as an additional entropy for identification.
GoLogin allows user to have the ability to control the canvas fingerprints on their browser profile. It has 3 modes of operation: Noise, Block, and Off.
Website will send an appeal from the browser about the Canvas function readout. After this, Canvas will create an algorithm on Noise mode block in the middle and an additional random but persistent noise to the readout. It is the best method to learn how voice modifier works. A voice modifier is placed at a particular preset, the voice will change and it will create a significantly different sound from the original voice and it is consistent over time.
After users apply the random noise to the readout, applying statistical analysis will recognize the fingerprint with 100% uniqueness.
For users who want to disable the website ability in reading a canvas, he or she needs to enable block mode. Website will perform the readout on the browser profile. Users who configured the canvas to block mode will notice that the returned value turns blank.
Solution treatment on this matter depend entirely on the website discretion. However, there are some cases that happened to users who did not intentionally try to hide their canvas fingerprints. In this case, browser error occurred during the process of data retrieval for the canvas object.
The feature will disable canvas mask and websites can recognize the actual canvas fingerprinting of the computer. It is an advantage to few scenarios when websites will react negatively on canvas readouts that are blocked or has 100% uniqueness.
It is best to put in mind that the actual canvas fingerprinting hash are common. Several computers that has a copy of it and it exists everywhere in the world. Therefore, when the real canvas fingerprint is revealed, to be the same as those computers who have the same hardware setup. Also, altering fingerprints increases the website entropy by allowing browser profiles to have separate identities.
There is another method used for decreasing browser profile entropy. Therefore, it allows a better blending to normal distribution of users in GoLogins on Mac computers. Mac computer have the same webGL fingerprints because they have the same build in nature. Therefore, some instances models will have identical hash.
Launching Browser Profiles on Several Computers
Users have to keep in mind when making browser profile with Canvas, configure the Noise mode. After which, launch the software on several computers that has different hardware. Websites will recognize the Canvas hash is inconsistent on several launch.
Added noise will persevere. However, there is an additional filter at the top of the existing computer fingerprint. Therefore, when a computer changes, the readout will also alter.
Some few resolutions for several computers that need non-changing readout:
Open GoLogin via VPS (Virtual Private Server) or VM (Virtual Machine), which has similar configurations to Hardware fingerprints that are on noise mode. Computers should have similar setups, and the masked WebGL fingerprint should be constant on several computers.
Open GoLogin through computers with the same model and uses the same driver, hardware, and operating system setup. Marked fingerprints remain constant to several computers as their setup in hardware are the same.
Open GoLogin via Mac computers. Above mentioned situations should be applied, it is best if it blends in.
Still not clear? Watch this video: