Reinhard Matz – Faces Without People

Review by Gerhard Clausing

The process of depicting simulated people in fake ‘portraits’ has reached new levels of perfection. The software programming known as GAN (Generative Adversarial Network) combines features from known human portraits to create endless new simulated ‘portraits’ depicting nonexisting pseudo-people, looking quite real. When I first came across this process in 2019, reviewing Tyka’s Pictures of Imaginary People, such images still looked somewhat rough and more artificial, a sort of Neanderthal version of what is now possible. Now they look perfect, at least most of the time, and such computed ‘portraits,’ based on real photographic input, raise a series of issues.

I was pleased to discover this update and curation by Reinhard Matz, via his article in Photonews, “Individuals in the Realm of Possibility” (Hamburg, June 2022, p. 16), an essay which is also translated and printed in this photobook of his. Here he has collected and published a series of newly perfected pseudo-portraits, gleaned from the website https://thispersondoesnotexist.com/, where you can generate endless perfect or not-so-perfect simulations, free of copyright, based on StyleGAN2. Matz also makes a series of valuable observations about the meaning and impact of such visuals in this current age of visual and other fakery.

The ‘portraits’ look real, and yet we know that they show individuals that don’t exist. If we want to call these pictures photographs, that would call for a revision of our definition of that term. These are feature composites, not unlike the old morphing procedure, which created a fictional portrait that was an interpolation between two actual portraits. Here the process combines the facial characteristics from many existing portraits of existing individuals to make ‘unlikenesses,’ forming pictures of non-existing, fictional individuals under software algorithms. No reality, no copyright. As Reinhard Matz points out, such fictional characters have always been acceptable in painting, but are a novum in photography, which still has the old concept of documentary truth value attached to it. Maybe this can be one of the factors that contribute to the further elevation of photography as an art form. The challenge to us is the level of perfection that these generated images present to us.

So here is a bit of a footnote on the current state of algorithms concerning GAN processes: most of the generated ‘portraits’ are perfect – yet ‘glitches’ occasionally find their way into the generated images. I was so fascinated by this process that I generated at least a thousand iterations on the website. Some weird finger combinations and other intrusions have been observed. Most interesting from the point of view of machine learning are those that show intrusions of seemingly alien other companion creatures on the margin, snuggling up to the primary figure depicted. Of course, you need to generate a huge number of iterations to get some marvelous and astonishing glitches, and I am sure that the algorithms will undergo further improvements to approach even greater perfection in the future. But this also demonstrates that our natural instincts in evaluating photos are still more sophisticated than these algorithms. The same website cited above now has new links where you can also try your hand at generating cats, horses, and abstract art that never existed. For the sake of completeness, I also am showing some less perfect images here and an example of abstract pseudo-art (these are not part of Matz’s book and were all generated during the last several weeks):

Reinhard Matz’s photobook sampler of perfect non-entity portraits, shown below, of people who never existed, is an important curation of what is now possible. I still detect a slight look of artificiality about these images. For instance, why is there a big lock of hair hanging on one side in figure 7 below? Would the ‘lady’ in the last image below, if she existed, only wear one earring for a portrait; would her eyebrows and eyelashes be that exaggerated and symmetrically perfect?

Simulated photographs are just one piece of a bigger picture. Fake videos also have reached levels that represent discernability challenges to us as human viewers; these have also been in existence for some time now, and I expect the creation of motion pictures with actors no longer alive will soon be upon us as well, as these various technologies and software improvements merge and undergo further ‘improvements.’ Only yesterday I heard on the news that Amazon’s Alexa “could soon be speaking with the voice of your dead relatives”! A DeepFake Klitschko simulation apparently ‘talked’ with the mayors of Berlin and Madrid this week …

O brave new world that has such art-like fakery in it!

____________

Gerhard (Gerry) Clausing, Associate Editor of the PhotoBook Journal, is an author and photographer from Southern California

____________

Reinhard Matz – Faces Without People

Author/Editor:  Reinhard Matz (born in Bremen, lives in Cologne and Berlin, Germany)

Publisher:   Hyper Focus Books, Cologne (Köln), Germany (self-published); © 2022

Text:  Reinhard Matz

Languages:  English and German

Hardcover, illustrated, sewn; 64 pages, unpaginated, with 28 color plates; 6 x 8.5 inches (15.5 x 21.5 cm); printed and bound in Germany by WirmachenDruck, Backnang; ISBN 978-3-948040-07-9

Photobook Designer:  Martin Zahder, Köln

__________

Articles and photographs published in the PhotoBook Journal may not be reproduced without the permission of the PhotoBook Journal staff and the photographer(s). All images, texts, and designs are under copyright by the authors and publishers.

One thought on “Reinhard Matz – Faces Without People

Add yours

Leave a Reply

Fill in your details below or click an icon to log in:

WordPress.com Logo

You are commenting using your WordPress.com account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )

Connecting to %s

This site uses Akismet to reduce spam. Learn how your comment data is processed.

Website Powered by WordPress.com.

Up ↑

%d bloggers like this: