Thoughts on Composites
For November 2015 Stardust
Composites: the good and the bad by Alister Ling
Obviously I am a fan of composites. I create them when it is not possible with a single image to capture a scene with high dynamic range or to convey the passage of time. But sadly the technique is often misused to artificially hype the "wow". At some point, the image is no longer astronomy but art - and since artistic taste is in the eye of the beholder, it's hard to define the line. Luca Vanzella and I have traded quite a few messages about Moon composites over the last few years, giving thumbs up to the excellent ones, and thumbs down for the bad ones. It is fair to say that Luca and I squarely land in the camp that composites should create "a human's perspective of the event", in a word, one that is "representative".
It's fine to create a movie-poster summary (art!) where proportion is deliberately adjusted to highlight one's memory or emotional experience. But it's another when the general public, who don't have the experience to know better, are led to think that it is a faithful capture of a wonderful event.
Image by Mike Taylor, reproduced at earthsky.org, "Best photos of the total lunar eclipse".
It's definitely art, but it's not a photo! I could accept it as a commemorative T-shirt/poster of a festival, but as a photo of the eclipse, it's way past the line. There is *some* astronomy in there, but there is so much "wrong" here that it really rankled me (Moon too big, shadows not right, curved trajectory, changing size of the Moon, daylight illumination of the ground).
We (and others) get quite annoyed when trickery is used to give the impression that it is representative, such as this recent example from APOD (Astronomy Picture of the Day) for September 29, 2015:
by Jose Antonio Hervás. The giveaway is how "postcard" the shot is. The odds of this being a true once-in-a-lifetime event is nearly astronomical. I was ready to send APOD a wag of the finger when I noticed that others already had, in the APOD General Astronomy Discussion Forum "Re: APOD: Supermoon Total Lunar Eclipse and Lightning Storm (2015 Sep 29) "http://asterisk.apod.com/viewtopic.php?f=9&t=35221".
It is interesting to see how some people defend this. One supporter noted "It was stated as being a composite, which is a legit astrophotography technique. There was a lot of work going into making this image, but I'd still call it astrophotography not art imagery." A follow-up comment that summarizes my perspective was "I don't have anything against composites but I'm certainly against art imagery presented as astrophotography."
A different reader made an important point: "Composite images are real images of phenomenon as they occur through time. Many composites exist for many reasons. All composites are either stitched mosaics covering areas larger than the original imager system allows to create a large panorama, or stacked composites which allow for gathering more light without having the shutter open long enough to create Blur or Star Trails. Why do you consider this image to be 'un-real' ?" Excellent point: no one sees star trails when they look at the sky or through a telescope.
Back came the reply from a subscriber to the forum, explaining the core of my disappointments:
"Let's summarize. The goods:
1) I personally like the image as well.
2) The Moon is real.
3) The landscape and the lightning are real.
The not so goods:
1) You can see the stars in front of the Moon in some cases. This is an indication that the Moon it is just copied and pasted in this position.
2) The Moon is too low. The moon is real but it's position is false.
3) The distance between the Moon's phases is also not indicative of the reality because if it was, the sequence wouldn't fit into this nice frame.
Consider this similar example. You turn the camera north and shoot some startrails. Then you turn your camera south and you shoot the Moon. Then you go to photoshop, copy the Moon and paste it in the centre of the startrails in front of Polaris.
The Moon is real. The startrails are real, but honestly, do you believe that this would be a real astrophoto? If the answer is yes, I respect your opinion but I disagree."
Bingo. But wait, there's more that's wrong: the angle of the Moon's position against the stars is about 30 degrees off of the ecliptic's angle against the background star field. And where is the reflection of the Moon in the water? Where is the reflection of the hill in the water? Ever seen artificial illumination of terrain from that far away that is perfectly balanced against the background stars? How come the Moon does not get increasingly yellow lower down due to atmospheric extinction like it does in all of Alister's and Luca's shots?
Of course it's easy to preach when one of my shots has not made APOD. Where do you draw the line between retouching and conjuring?