Firefox and corrupted PNG images, colour management and ICC v4 support

Jan 122013

Introduction

I’d previously reported that I had experienced corrupted PNG images in Firefox if I enabled ICC v4 support within the browser:

Corrupted PNG images, Firefox colour management, ICC v4 profile support, and Windows 7
http://www.metalvortex.com/blog/2012/09/17/1008.html

That was way back in September 2012 and there have been many releases of Firefox since then, and my Windows machines have been through a number of Service Pack and revision updates since then including some changes to my implementation of monitor calibration.

I’ve also noticed this PNG image corruption is logged as a bug at https://bugzilla.mozilla.org/show_bug.cgi?id=791456 and there is suspicion there that the profile embedded in the PNG image may be corrupted in some way. I’ll have to do some separate checks on that suspicion but I could not detect any embedded profile in the PNG image and it appears to be a simple greyscale-formatted image. In the meantime, I have undertaken some quick checks on ICC v2 and ICC v4 support and seeing the effect on viewing the xkcd PNG images.

Tests

With the release of Firefox 18 I thought it was about time to test again. I undertook five tests, with configuration and results detailed below. Generally, I used the ICC test page at http://www.color.org/version4html.xalter to confirm ICC version compatibility, and I also checked the PNG image at http://xkcd.com/1108/ for PNG image corruption. My system is a Windows 8 Pro 64-bit machine with all the latest updates as of today (12th January 2013) in a dual-monitor configuration using two Dell UltraSharp 3008WFP 30-inch wide-gamut monitors.

1. Test 1 – ICC v4 monitor profile (matrix-based), ICC v4 enabled in Firefox

Config: The ICC v4 monitor profile was created with x-rite “i1 Profile” using an i1 Display Pro colorimeter. The profile type was matrix-based.

When I checked at http://xkcd.com/1108/ the PNG image there was not corrupted.

When I checked at http://www.color.org/version4html.xalter the image colours were shown in an unexpected manner:

icc4-test1

Clearly something was very wrong.

I then checked the integrity of this monitor profile using “ICC Profile Inspector” and CHROMIX “ColorThink”; the profile seemed fine, so I suspect that the error is within Firefox. I would like to have used ArgyllCMS to create an ICC v4 profile to help in validating if the error is within Firefox and not within the profile but, at this time, ArgyllCMS does not support ICC v4.

2. Test 2 – ICC v2 monitor profile (LUT-based), ICC v4 enabled in Firefox

Config: The ICC v2 monitor profile was created with ArgyllCMS + dispcalGUI using an i1 Display Pro colorimeter. The profile type was XYZ LUT + “swapped” matrix.

When I checked at http://xkcd.com/1108/ the PNG image there was corrupted.

When I checked at http://www.color.org/version4html.xalter the image indicated full ICC v4 compatibility:

icc4-test2

ICC v4 support and LUT-based support is excellent here but the downside is that the PNG image was corrupted. Oh dear.

3. Test 3 – ICC v2 monitor profile (LUT-based), ICC v4 disabled in Firefox

Config: The ICC v2 monitor profile was created with ArgyllCMS + dispcalGUI using an i1 Display Pro colorimeter. The profile type was XYZ LUT + “swapped” matrix.

As soon as I loaded Firefox I knew there was a problem; all of the Firefox app window colours were messed up. And, of course, so were the colours on the Internet.

When I checked at http://xkcd.com/1108/ the PNG image there was not corrupted but other elements had incorrect colours.

When I checked at http://www.color.org/version4html.xalter the image indicated ICC v2 compatibility but all the colours were incorrect:

icc4-test3

What does all this mean? Well, the ICC v2 profile was LUT-based, to which was added a deliberately wacked matrix. Firefox was clearly not using the LUT data when ICC v4 is disabled and instead used the matrix data as shown by the wacked colours. So the lesson here is not to use LUT-based ICC profiles (at least with ICC v2 profiles, as I was not able to create LUT-based ICC v4 profiles).

To clarify the above, the Firefox setting “gfx.color_management.enablev4″ enables ICC v2 LUT support support as well as enabling ICC v4 support; you can’t have one without the other.

4. Test 4 – ICC v2 monitor profile (matrix-based), ICC v4 disabled in Firefox

Config: The ICC v2 monitor profile was created with ArgyllCMS + dispcalGUI using an i1 Display Pro colorimeter. The profile type was matrix-based with curves. I used a matrix-based profile as it was clear from the previous test that Firefox was not reading correctly reading LUT-based profiles.

When I checked at http://xkcd.com/1108/ the PNG image there was not corrupted.

When I checked at http://www.color.org/version4html.xalter the image indicated ICC v2 compatibility:

icc4-test4

5. Test 5 – ICC v4 monitor profile (matrix-based), ICC v4 disabled in Firefox

Config: The ICC v2 monitor profile was created with ArgyllCMS + dispcalGUI using an i1 Display Pro colorimeter. The profile type was matrix-based.

When I checked at http://xkcd.com/1108/ the PNG image there was not corrupted.

When I checked at http://www.color.org/version4html.xalter the image indicated ICC v2 compatibility:

icc4-test5

Recommendations

These tests are by no means exhaustive or even as rigorous as I would have liked but I can recommend the following:

Avoid the use of ICC v4 monitor profiles and, instead, use ICC v2 profiles whenever possible.
Avoid the use of LUT-based monitor ICC profiles and, instead, use matrix-based ICC profiles whenever possible. A mixed LUT+matrix profile is acceptable.
Ensure that ICC v4 support within Firefox is disabled (gfx.color_management.enablev4 = false). Note that this will also disable ICC v2 LUT-based profile support.
Wait for Mozilla to confirm and fix colour-management issues with ICC v4 profiles!

Article by Kulvinder Singh Matharu – 2013

One of the evils of Hinduism

Uncategorized Comments Off

Jan 082013

Most of my arguments and pot-shots against religion have concentrated on Christianity and Islam simply due to the level of exposure that these two religions have here. Other religions, of course, cannot escape from the exposure of their own injustice and barbarity. Here’s one article on Hinduism and the terrible caste system:

My Take: Hinduism’s caste problem, out in the open
http://richarddawkins.net/news_articles/2013/1/5/my-take-hinduism-s-caste-problem-out-in-the-open

What can I say about Hinduism? Trash, unjust, evil, barbaric. Yup, this can be equally well to Christianity, Judaism, and Islam.

Buddhism is another on the list as its core beliefs are ingrained with injustice. But that’s a topic for another article.

Article by Kulvinder Singh Matharu – 2013

Will a Lance Armstrong confession be sufficient?

Uncategorized Comments Off

Jan 072013

I asked in August 2012 that if Lance Armstrong cheated, could he be forgiven?

However, Armstrong has consistently and ruthlessly gone after anyone who crossed him, ruined some of them, and has continually denied cheating. Well, if Armstrong is found be a cheat then he has likely lost his window of rehabilitation, despite his charity work, and I won’t be able to forgive him; he has burnt too many people. But I suppose it depends on the manner of his confession, his humility, and any restitution done to heal and mend everything that he has destroyed.

Since I wrote that article, Armstrong has been found to be a cheat and a liar, but he has still not admitted to anything. But now there are reports that Armstrong is considering making a public confession. I’m afraid, at least in my mind, that any confession now will be far too little, far too late, and may perhaps be driven purely by a selfish need to partake in competitions. His route to rehabilitation and forgiveness cannot be allowed to be an easy one, and certainly it cannot be a quick one; Armstrong needs to prove the sincerity of his confessions and of any forms of restitution that he undertakes. And when have we ever seen a humble Armstrong?

Are my thoughts too punitive? I’m not sure. But there does certainly need to be a punitive measure imposed on Armstrong as part of his rehabilitation.

Article by Kulvinder Singh Matharu – 2013

Jerry Coyne discusses Susan Jacoby and atheism

Uncategorized Comments Off

Jan 072013

Jerry Coyne has an article on Susan Jacoby and her recent work, plus a mention of her soon-to-be-released book on Robert Ingersoll.

Susan Jacoby: her new book and “strident” editorial
http://whyevolutionistrue.wordpress.com/2013/01/07/susan-jacoby-her-new-book-and-strident-editorial/

I’d not heard much about Jacoby before which is quite surprising as I’ve been doing a lot of reading on Ingersoll over the last couple of years or so; I definitely need to read more of her work.

Article by Kulvinder Singh Matharu – 2013

What is atheism?

Uncategorized Comments Off

Jan 062013

Someone asked me what values “atheism” had, what philosophies it had. That’s an interesting question but the answer is rather simple.

I am an atheist; simply a person who is convinced that there is no god. What makes me so sure that there is no god? Well, I certainly have no method to prove the non-existence of a god, and I do allow for the possibly that a god may exist, but I’ve not found any credible evidence for the existence of a god, and the world looks exactly how it should if there was no god. I therefore conclude that the existence of a god is unlikely. This just happens to align quite closely with Richard Dawkins’s own position of “6” along the spectrum of theistic probability.

To elaborate, I claim the non-existence of a god with the same confidence that I claim that there are no fairies living at the bottom of the garden and that there are no invisible dragons living in my garage. An atheistic viewpoint should be the default position, and other viewpoints taken when evidence allows.

There is no need to add any other philosophies or viewpoints to the word “atheism”; atheism does not dictate or imply a way of life. Recent attempts to redefine and expand what atheism means, well-meaning though they are, cause confusion and create division. Whilst it is a good idea for atheist communities to exchange thoughts, re-branding atheism into a set of philosophies is not the way forward and I distance myself from such re-branding exercises.

For myself, I recognise the special time that I am alive, and I intend to make the most of what little time I have. I empathise with those who suffer and I do what I can to alleviate suffering. I seek joy in myself and in others. But these traits are not unique to any single philosophy; the religious experience these as well. But I am also rather interested in explaining to others why belief in the existence of a god is not rational, and why religions with their unjust teachings should be abandoned; such debates can be rather interesting but mostly unsatisfactory as not many people are able to abandon a framework that has been indoctrinated into them since childhood. But again, none of this is required to be an atheist.

So, the answer is rather simple. Atheism is simply the position that there is no god. That’s all.

Article by Kulvinder Singh Matharu – 2013

Religion is a barrier to the understanding of the true nature of things

Uncategorized Comments Off

Jan 062013

I was perusing the Discovery Magazine blogs and came across an article from Nov 2012 by Razib Khan which seeks to describe some of the tensions experienced between religion and science:

Mine eyes have seen the glory; we are as gods!
http://blogs.discovermagazine.com/gnxp/2012/11/mine-eyes-have-seen-the-glory-we-are-as-gods/

Within the framework of Western civilisation, past endeavours at seeking knowledge have, in large part, been motivated and instigated through religious inspiration. However, the religious increasingly did not like what was being found, and their early attempts to smother the truth were heavy handed, unjust and brutal. But these prohibitions on seeking and disseminating knowledge on the true nature of things were ultimately doomed to failure as the methodology for discovering truth matured into what we call science. Religion, today, has no part to play in the search for truth, be they of a scientific nature or of a moral nature.

Khan’s blog is quite interesting as it piques my interest in human origin, so I intend to spend a bit of time there catching up on all of his articles.

Article by Kulvinder Singh Matharu – 2013

Why science and religion are not compatible

Uncategorized Comments Off

Jan 042013

A presentation by Jerry Coyne, author of “Why evolution is true”, explaining why religion is not compatible with science:

Jerry Coyne on the Odd Couple: Why Science and Religion Shouldn’t Cohabit

Article by Kulvinder Singh Matharu – 2013

Is it acceptable to use the manufacturer’s colour profile for my monitor?

Uncategorized Comments Off

Jan 042013

I’ve previously explained the importance of calibrating and profiling the monitor to ensure correct colour management. But some people have queried if it’s acceptable to use the manufacture-supplied colour profile for the monitor. In most cases this is a bad idea. Look at the diagram below. This is a 2D CIE colour space chromaticity diagram where I’ve plotted two colour profiles for my wide-gamut monitor; one profile is the standard profile supplied by the manufacturer, the other profile is one I’ve created after calibrating the monitor.

The manufacturer-supplied colour profile is actually not too far from an sRGB profile. However, it’s pretty clear that the manufacturer-supplied profile does not even remotely characterise my monitor and is therefore completely unsuitable for use in any colour-management workflows. So, next time you try to use colour-managed software such as colour-enabled web browser, make sure that your monitor has been calibrated and profiled otherwise your user experience will suffer. If you haven’t calibrated and profiled your monitor, please, disable any colour management features in your software.

I actually have strong feelings that colour-management in web-browsers need to be turned-off by default, and that web-browsers need to give the user a prominent and clear choice on whether to turn-on colour management. That’s why I think Firefox has got it wrong be enabling colour management by default. But that’s a rant for another day.

Article by Kulvinder Singh Matharu – 2013

1:12 scale model of 2012 Harley Davidson FLSTF Softail Fat Boy motorbike

Uncategorized Comments Off

Dec 292012

A 1:12 scale model of a 2012 Harley Davidson FLSTF Softail Fat Boy motorbike. Photo taken using three flashguns bounced off the ceiling for even lighting. There was some ambient tungsten lighting but its contribution was minimised somewhat by using a 1/250 second shutter speed and the flashguns at full power. I used a Nikon D800 camera for this at an aperture of f/8 and ISO 100 so there should be plenty of detail there for a very large print; I may do a silver print at some point.

http://www.metalvortex.com/images/Misc%20Photos/slides/_DSC2196c.html

Article by Kulvinder Singh Matharu – 2012

Photo of mini replica guitars

Uncategorized Comments Off

Dec 242012

I’ve just uploaded a photo of mini replica guitars. Each guitar is about 25 centimetres (or 10 inches) in length. The guitars were elevated about 2 centimetres from the background lying horizontally on a flat surface. I used three flashguns bounced off the ceiling for even lighting.