I had been playing with various fractal apps in the early 2000’s and marveled shortly at the wild gyrations on the screen. But I rapidly started to feel uneasy with the machine like perfect outcome of those works. Such a perfection in my eyes appeared to be devoid of life characteristics such as the flaws that litter the path taken by the process of constant transformations that life is all about. In the case of a visual sign such flaws of life can take different forms like an imperfection in a line indicating an accidental move of the author’s hand. I always felt that the absence of such signs of life deny human craft qualities to these works. And this is how I started to experiment diverse methods to digitally transform photographs of my paintings. In 2006 I terminated 12 digital variations of each of the 40 acrylics in my acrylics “artsense” series. This gave 480 digital variations that were edited as very short limited edition prints. These works can be viewed here.
I later summed up my feelings about what I see are the main differences between fractals and digital transformations in the following posts:
Digital variations versus fractals
Science imaging: towards new patterns out of mathematical knowledge
The exponential rise in scientific imaging. (4) FRACTALS
Around 2010 approximately, I initiated a new approach in painting that I summarized in “The grand project” and I terminated its first 20 canvasses in 2015. Since then I concentrated mostly on writing my series of posts titled “From Modernity to After-Modernity”.
The grand project. 4 columns - 5 panels per column. More columns are in the work.
Each panel is 60 x 90 cm. The 4 columns reach 3 meters high and are 3.6 meters in width.
Each panel is 60 x 90 cm. The 4 columns reach 3 meters high and are 3.6 meters in width.
What is the grand project all about ?
- It’s all about thinking and dreaming. The target being to reconcile what we think of as certainties in our conscious mind with the visions and discoveries that our subconscious sometimes shares with its complementary counterpart the conscious mind. As you might expect our conscious mind does not take such a process easily. I wrote extensively about this process of reconciliation in “From Modernity to after-Modernity” book 3: Divination.
- Oil painting, to me, is like a process of meditation which allows me to slowly digest my thinking and dreaming. It is a process that is achieved outside of will and want that I consider to be deficiencies of the mind which were encouraged to grow during our societal domestication. What this implies is that my painting is like surfing on the waves of my subconscious’ visions and discoveries. In this process my conscious mind stays in the background limiting its interventions to solving the technical problems of the image rendering or of the execution of the craft of image making.
- Further thinking and writing then helps me to try to reconcile these subconscious visions and discoveries, that are partially illustrated in my paintings, with my conscious certainties. This sometimes results in the absorption and integration of these visions in my conscious mind. I say sometimes because all visions and discoveries, shared by the subconscious to the conscious mind, are not necessarily reconciled in my conscious and reasoning mind just after it perceives them. It sometimes takes years for my conscious mind to get a reasoned grip on a vision and thus be able to integrate it in the sum of its knowledge. Such a process, at its roots, had been refined over the tens of thousands of years that preceded the emergence of civilization by the tribal men of knowledge in their search to increase their consciousness. My own practice dates merely some 4 decades and is thus a sketchy and very poor approximation of the mastery attained by these men of knowledge of antiquity.
- Digital variations of photographs of my oil paintings are a convenient and fast way to dig deeper into my subconscious visions. They pull me deeper in the meaning that my subconscious had inscribed in my painting and in that sense digital variations expand the realm that becomes accessible to the mind. The other aspect that is so tempting with digital variations is the ease and speed in the execution of a multiplicity of images about these visions. While a painting (60 x 90 cm) takes me 150 hours of work on average a digital variation takes on average 1 and a half to 3 hours which is 1 to 2% of the time it takes me to terminate a painting.
Mining for images inside a digital variation
My past digital variations, I mean each one of them, emerged from the application of a series of algorithms to the photographs of my paintings. Such algorithms are also called filters. This time around I tested some 25 filters. After many trials, of different combinations of filters, my attention got suddenly caught by one of the transformations. It was the 7th trial in a row and the resulting image was so different from the other 6 that I felt like being pulled into it.
The original resolution of the photograph of painting 1-01 gives a printable size of 60 x 90 cm (10630 x 7087 pixels). I decided to keep the same ‘resolution-printable size’ for each digital transformation.
Here is an image of the 7th digital transformation of this photograph of painting 1-01 that resulted from the application of 15 successive filters.
This image is in JPG format. It sizes to 1200 x 800 px and is optimized for the web at a JPG quality of 60 and it still has a weight of 174 Kb. While totaling roughly 0.1% of the total pixels, in the original image of the transformation, it is still possible to discern the utter complexity that is at play here.
I had applied some 15 filters in succession to the original photograph of the painting and this transformation confronted me with something that I was not prepared for. The result was indeed totally unexpected. I mean the resulting image was so utterly complex and its substance appeared to be so rich that I felt compelled to zoom in it. It was like an irresistible attraction that I had never experienced earlier. And totally out of the blue that attraction was soon engaging me in new adventure that later would prove to be very time consuming indeed.
Here is what happened when I zoomed in this digital transformation:
1. Confronted with its complexity and richness I fell compelled to blow up the size of the image by a factor of 2 vertically and 2 horizontally, which is 4 times the initial printable size of 60 x 90 cm, but the complexity was still there. This resulted in a division of the image in 4 quarters ABCD (red lines on the graph here under). So the name of any given image starts with the letter indicating what quarter it is located in.
2. Each quarter was then zoomed in by the same zoom factor of 2 vertically and 2 horizontally. Keeping in mind my intention to keep the original size of 60 x 90 cm for each new image this zoom resulted in a total of 4 images horizontally that multiplied by 4 images vertically for a total of 16 images. But the complexity and richness were still too great to possibly appreciate what was going on in each part of the transformed image. Each quarter ABCD is divided in the 4 quarters 1234 (in blue on the graph). So the name of any given image in A, for example, becomes A1, A2, A3 or A4.
3. Seen the non usability of these 16 images I decided to zoom still deeper by the same zoom factor of 2 vertically and 2 horizontally. The result was 8 images horizontally that multiplied by 8 images vertically which gave a grand total of 64 images. But the richness of content was still such that I could not grasp what was really going on in each of these images. Taking as example an image in quarter A1 its name now has expanded to A1-1, A1-2, A1-3, or A1-4
4. I ventured to zoom deeper once more by the same zoom factor of 2 vertically and 2 horizontally. The result was 16 images horizontally that multiplied by 16 images vertically which gave a grand total of 256 images. At that level of zoom the one image, among these 256 potential images, that I extracted was revealing a baffling subject matter that simply blew my mind. The image names than expanded to their final name. A1-1, for example expanded into A1-11, A1-12, A1-13, or A1-14
Here follows image A1-24 that is indicated on this graph. The name of the image is in reality 07A1-24.JPG. The 07 stands for my 7th essay at applying a combination of filters. It was only after seeing this image that I decided to concentrate all my energy on this particular transformation. The other essays are being kept in storage for eventual future uses.
This image sizes to 1200 x 800 px and is optimized for the web at a JPG quality parameter of 60. The original gives a printable size of 60 x 90 cm (10630 x 7087 pixels).
This was the first image that I extracted from this 4th zooming episode and I was mesmerized. First I was baffled that this image was so well balanced in all its technical aspects. Then I sat there for hours observing what was going on in it. And that’s when the magic set in. I fell over on my magic carpet and it transported me in the world of my subconscious visions. What a trip. I was hooked.
The subject matter of that image was baffling me and so I rapidly extracted a couple more. Each one of these images had the same well balanced character while the subject matter of each one of them was very distinct. I called on the intuition of my wife Xiaohong to confirm that I was not dreaming. She confirmed that these images had something special in them that reflects beauty while questioning the mind. I knew enough. I was convinced that each one of these images was suggesting the impression of a scene in the future unfolding of the predicament of humanity. But let me cut short on my interpretation of the subject matter of these images. First I did not chose the subject matter. The digital transformation imposed it on me. Secondly this post is intended to illustrate the process that gave way to this digital transformation and then of my handling of the extraction of images from zooming deeper and deeper which in my mind appeared as if I was mining for imaging resources.
The next day I came back to my senses and was forced to recognize that I had in fact really set foot into a minefield. I had spent some 5 hours letting the computer calculate the parameters of each filter. Not bad for a transformation of such a complexity. But now I felt that I was literally sequestered in a mine and that the extraction of the next 250 images was going to keep me in the mine for a few months… My guesswork was right on the mark. Today I have terminated the extraction of the 128 images of quarter A and quarter B. This took me nearly a month. I still need to extract the 128 images of quarter C and D. One more month work at 8 hours a day on average. Yay the extraction of one image costs me just under 2 hours.
As any seasoned expert in imaging will attest the process I just described here above is not for the faint of heart. Blowing up an image at such high zooming factors degrades indeed their quality and their print-ability. It is generally admitted that there is a technical limit to blowing up images and that limit is most often considered to be around an order of magnification of 10. What this means is that magnifying an image over 10 times its original size is generally considered to be impossible for reason of degrading quality.
But, as I already mentioned, I was engaged in an adventure of discovery and so I did not mind these rules of thumb. I was not sure that I would continue to get printable images after each level of magnification. But I did not mind. My curiosity was tricking me deeper and deeper. Once you are in the middle of a jungle there is no guarantee that you’ll come out of it unscathed. But you simply never think about it. Your attention is just absorbed by what you discover with each step forward. Your absorption in the process lets you forget all the barriers that the conscious mind otherwise never misses to remember you about. This complete absorption in the job procures you momentarily a complete freedom, some people call this madness, that allows you to keep going and going on a path that otherwise would be considered impossible.
At each level of magnification I applied 5-6 enhancement filters to the image before blowing it up. And then after extraction of each magnified image I applied another set of different enhancements. To my surprise the extracted images kept remained crisply clear and sharp and even stayed so when zoomed at 200%. But I did not blindly trust the quality of the images on screen. I kept indeed testing their quality by printing A4 size extracted parts. Crisply sharp prints is what encouraged me to venture always deeper. Let’s not forget that the final images that I extracted are resulting from an impossible magnification level of 16 horizontally x 16 vertically. These prints were reassuring...
Thinking about it I believe two things made sharp printed images possible even after such high levels of magnification:
Having summarized the process lets now view some samples extracted from my mining adventure.
Some Samples extracted from the initial transformation
(Click on an image to blow it up at its maximum size)
The subject matter of that image was baffling me and so I rapidly extracted a couple more. Each one of these images had the same well balanced character while the subject matter of each one of them was very distinct. I called on the intuition of my wife Xiaohong to confirm that I was not dreaming. She confirmed that these images had something special in them that reflects beauty while questioning the mind. I knew enough. I was convinced that each one of these images was suggesting the impression of a scene in the future unfolding of the predicament of humanity. But let me cut short on my interpretation of the subject matter of these images. First I did not chose the subject matter. The digital transformation imposed it on me. Secondly this post is intended to illustrate the process that gave way to this digital transformation and then of my handling of the extraction of images from zooming deeper and deeper which in my mind appeared as if I was mining for imaging resources.
The next day I came back to my senses and was forced to recognize that I had in fact really set foot into a minefield. I had spent some 5 hours letting the computer calculate the parameters of each filter. Not bad for a transformation of such a complexity. But now I felt that I was literally sequestered in a mine and that the extraction of the next 250 images was going to keep me in the mine for a few months… My guesswork was right on the mark. Today I have terminated the extraction of the 128 images of quarter A and quarter B. This took me nearly a month. I still need to extract the 128 images of quarter C and D. One more month work at 8 hours a day on average. Yay the extraction of one image costs me just under 2 hours.
As any seasoned expert in imaging will attest the process I just described here above is not for the faint of heart. Blowing up an image at such high zooming factors degrades indeed their quality and their print-ability. It is generally admitted that there is a technical limit to blowing up images and that limit is most often considered to be around an order of magnification of 10. What this means is that magnifying an image over 10 times its original size is generally considered to be impossible for reason of degrading quality.
But, as I already mentioned, I was engaged in an adventure of discovery and so I did not mind these rules of thumb. I was not sure that I would continue to get printable images after each level of magnification. But I did not mind. My curiosity was tricking me deeper and deeper. Once you are in the middle of a jungle there is no guarantee that you’ll come out of it unscathed. But you simply never think about it. Your attention is just absorbed by what you discover with each step forward. Your absorption in the process lets you forget all the barriers that the conscious mind otherwise never misses to remember you about. This complete absorption in the job procures you momentarily a complete freedom, some people call this madness, that allows you to keep going and going on a path that otherwise would be considered impossible.
At each level of magnification I applied 5-6 enhancement filters to the image before blowing it up. And then after extraction of each magnified image I applied another set of different enhancements. To my surprise the extracted images kept remained crisply clear and sharp and even stayed so when zoomed at 200%. But I did not blindly trust the quality of the images on screen. I kept indeed testing their quality by printing A4 size extracted parts. Crisply sharp prints is what encouraged me to venture always deeper. Let’s not forget that the final images that I extracted are resulting from an impossible magnification level of 16 horizontally x 16 vertically. These prints were reassuring...
Thinking about it I believe two things made sharp printed images possible even after such high levels of magnification:
- for one this level of magnification was reached in incremental steps – 1 to 4, – 4 to 16, –16 to 64, – and finally 64 to 256.
- secondly the application of enhancements filters
- to the images before their magnification
- to the images following their extraction.
Having summarized the process lets now view some samples extracted from my mining adventure.
Some Samples extracted from the initial transformation
(Click on an image to blow it up at its maximum size)
Some initial thoughts about these images
I already wrote here above that I was immediately taken aback by the well balanced technical aspects of the first image that I extracted and also by the mystery popping out of the image that questioned my mind. Waves of doubts were certainly regularly pushing down my enthusiasm but an idea sprout in my mind that ever grew stronger. I was finding these images far more interesting than the original painting they were derived from. Was it because I had spent all these hours in the painting and so they had no secrets for me any longer? I soon found out that there was something else at work. These parcels extracted from transformation #7 are like extensions of the original painting. But they go further than the painting. They actually suggest the impression of a scenery where life is thriving. But where does this scenery come from? Are we to believe that the algorithms creates this? No way! The only possible explanation that makes sense in my mind is that, even if my painting does not literally contain such scenery, its spirit is nevertheless infused throughout the painting. And the magnifications and enhancements applied during the transformation process somehow suggest an impression of what is latent throughout the painting.
Technically, while being derived from the application of algorithms, what is striking is that these images don’t really look like being machine made. It seems that the multiple magnifications and enhancements reproduced truthfully the imperfections caused by my human hand in the original painting and so these imperfections have been passed down to the final images. The second technical aspect that baffles me is that each of these images is balancing so well all the pictorial elements. This kind of stuns me. Is this also due to truthful reproductions through magnifications and enhancements of something that was there in the original painting? Or are we to believe that the algorithms automatically produce such a balanced outcome? That question is easier to answer. After having initiated many hundreds, or perhaps thousands of algorithm combinations, I know for a fact that the outcome of algorithms is not automatically harmonious nor well balanced.
Are we thus left with the only reasonable explanation that the outcome
of algorithmic transformations is given or is contained in the original
image? I would argue that this is the case indeed.
No comments:
Post a Comment