Slides for a keynote on G'MIC

Hello everyone,
I’ve been asked to prepare a 45mn keynote for an international conference, called Cyberworlds 2021 (link to the program).
Of course, I’d like to talk about G’MIC, and more generally about what I call “Artistic Imaging”, that is the research field where new algorithms can be designed to eventually help artists producing their artworks or allowing them to be even more creative.

So far, here are the slides I’ve produced:

(there are a lot of slides, but see them as something that is mainly animated, the plan is not to spend 3 mn on each slide).

I’d be interested by your feedback on this. It’s a keynote, so for a general audience (scientists anyway), not specially intended to be highly technical, just something that summarizes the concepts and the work I’ve done on G’MIC, as well as a focus on a few algorithms I’ve been able to produce while collaborating with artists (David Revoy and @patdavid to name them).
I’d really like to improve the presentation, if possible, so feel free to comment. Particularly interested by what the regular G’MIC contributors think about it! (@grosgood , @Reptorian , @afre , @garagecoder, @Iain, @Joan_Rake1 , @rich2005 , @KaRo and others I forget).

Thank you in advance to the people who will take the time to watch this and give me their opinion!



There’s a couple of filters that I have made that are more into science part. That would be Henon’s Phase Diagram, Chirikov-Taylor, and finally Lyapunov Fractal. Not sure how much these will apply to your audience, but the existence of them do show that G’MIC can be used for scientific application.

It would be admittedly hard to explain what I did in those as they’re a special case of using string and converting them into math parsing code.

It’s not yet a replacement for python for that usage yet, but there are things to do regardless.

1 Like

With each new filter, the list of things to cover gets even longer!

I create almost no artistic filters, so perhaps not the best person to ask. However, the reason is that the creativity is instead spent on algorithms. It’s partially covered by saying it’s good for prototypes, but actually it’s very good for “messing about” with algorithms to see what works too. I lose hours to that sometimes! There are few languages powerful enough to do that easily (shaders perhaps?).

It’s also incredibly easy to recreate what would be multiple steps in a GUI as a single-click g’mic filter (think layer blending etc.).

Another thing I mentioned in the past is that g’mic could lend itself as an audio synthesizer. One day I’ll get around to it!

1 Like

I already made a thread on that quite a while ago. You could continue my work.

Found it → 1D processing thread

1 Like

I actually experimented with GMIC and Photosounder.

Photosounder is a program that converts sound files to spectrogram images and also spectrograms to sound files. A spectrogram image is easily manipulated in GMIC.

The problem is there is the round trip to get feedback on what you are doing. You can easily change how a spectrogram looks, but you have to convert it back to sound to hear how your changes sound.

Trim anyway. Quality ideas are never served through excess quantity.

  1. An illustrator’s stock-in-trade is Enchantment. Endowing drawings with such is a crap-shoot in which every illustrator must partake — with endless experiment, for there is no formula for Enchantment. G’MIC’s extraordinary array of filters increases one’s casts of the die. Perhaps one throw, one filter, suddenly — abruptly! — renders an image enchanting. No telling what leads to it, but a large number of possibilities always helps.
  2. An illustrator’s life is filled with the mundane. Experiencing finished Enchantment gives no idea of the steps leading up to it. G’MIC’s practical tools in filtering noise or auto-completing counters and spaces for color filling liberates the illustrator from much of the mechanical bits of illustrating. That opens up time for the further search of Enchantment, wherever it may be.

Right, I practiced my presentation today, and it does not fit in the schedule.
I’ve removed 44 slides (mainly about Filter Showcase). Still a bit long I think, so I’ll try to reduce even more.

Concerning enchantment :
I attach you an image which I used sometimes in talks. Looking at it, one half of the auditory does not see what is shown and the other half cannot believe that anybody don’t see what it shows! The latter is true for me, still sometimes I am looking at the image, thinking, what was it, there was something…

Maybe it could serve as an example image in the @grosgood sense. I don’t know what filters applied could show. (1.6 MB)

It is obviously (?) a photograph from Lascaux caves!

Sorry, didn’t have the time to take a look. Based on your previous presentations, I bet it will be great.

I probably don’t have any advice that you haven’t heard of before or have learned from experience. What I would say is that I generally prepare enough content for half the time. Usually, that leaves me breathing room for a comfortable full session.

That’s about the right place to cut. Perhaps I thought it overly long about there – the pace slowed for me in the long march through filters.

Don’t get me wrong - It is a good presentation! But you have your work cut out for you. The technical discussion concerning G’MIC, to my mind, is quite straightforward: are you improving the signal and cutting the noise? The artistic discussion is challenging, however.

Ditto here. My earliest recollection of G’MIC was play — “oh, what might happen if I do this?” To my mind, offering an artistic playpen for illustrators is not something that can be readily discounted. You cannot, directly, wave a magic wand and make an illustration enchanting, but to play with possibilities is likely the best indirect approach — “abruptly, using that filter, this magic happened!” So, for artistry’s sake, G’MIC is a tool that offers a great many avenues for exploration. If you can leave your audience with that sense of wonder — child-like but not childish — you have won the day, methinks. Good luck!

For me, all technical filters are one step from being artistic ones. I couldn’t tell you how many times I used one of my own base filters for 10 very different effects. That is why I focus on the mundane. But I suppose presentations require the pizzazz. Off to lunch: and guess what? On the menu is pizza without the zz!

Good stuff.

The final slide, final line, has “reproductible”. The English word is “reproducible”.

1 Like

Needs Need for tools to visualize / explore data, convert image formats, apply classical IP operators (filtering, geometric transformations, frequential frequency analysis, …) for very generic image data, sometimes on thousands of imagesat the same time.

Existing existing software/libraries were…

We had done like others : We had developed generic libraries for image processing :CImgandPandore(open-source, in C++) :

One quick thought is that the filter example slides could be reused in the reference pages because they are a quick visual of what is possible with little cues on how to get there.

1 Like

Many thanks for your feedback!

Keynote done. I had to do it online because my daughter got sick and couldn’t go to school.
It went relatively well, at least I was on time.


Stay safe out there and hope your family get better.