Wednesday 15 March 2017

Artefacts from Hell

At the end of Marko Riikonen’s recent post, we hear a profound crie de couer about artefacts which plague some camera makes and models. For those of you who are blessed with top of the range, state of the art equipment, you might be wondering what all the fuss is about. Can artefacts which plague some cameras really be that bad? The short answer is a resounding yes, they can be horrendous. Moreover, for the type of work serious halo observers are engaged upon, these cameras are virtually useless. Marko mentions that for him the problems are exacerbated when trying to photograph during daylight but ease somewhat at night when recording spotlight displays. From personal experience with my own particular model, I can say that it is equally bad day or night. Here is an example of how bad it can get even at night. The image is of a simple circumscribed lunar halo with a faint 9° odd radius.

The halo is almost obliterated by the artefacts. This is one reason why I very rarely post anything these days because the images are just so embarrassing. 

The consensus amongst seasoned halo observers and professional photographers is that full frame cameras are largely free of this level of artefact or at least they are not affected to quite the same degree. However, sadly for many, myself included, professional full frame cameras are just too expensive and so we have to settle for cheaper consumer models. The problem here is that by and large you select a camera model based on a review you either see online or in a camera magazine. None of the reviewers who write these are engaged in the type of work we do so rarely, if ever, is this problem of artefacts encountered, mentioned or highlighted. Once a budding halo enthusiast has selected his camera on the basis of one of these reviews and splashes out several hundred euros on the latest consumer model, disillusion can very quickly set in when he realises he has bought a complete and utter duff. More than likely, he or she is going to be chained to this millstone for the next four or five years until they can afford something better.



So then, we come to the crux of this post: what to do about this situation? To all those who are in this situation and are experiencing this problem, I want to ask our readers two things.

Firstly, can we compile a list of camera makes and models that are particularly prone to artefact problems? I would like to do this to alert everyone to the problems with those specific models and to try to stop people wasting their money buying them if they are predominantly going to be used for photographing halos. This is not to say that a particular model is “bad” per se, just that it is ill suited for halo photography. Conversely, if you are in the fortunate position of being in possession of a camera which is particularly free from artefacts or one which you could recommend in this respect, we would also like to hear from you. 

Secondly, I would like to ask everyone whether they have any tips or solutions for overcoming or lessening these artefacts if you are unfortunate to have a camera that is beset by them. These tips could either be in-camera or pre- or post-processing. 

After a suitable period of time, once we have received some feedback on this issue, perhaps we can collate the results and do a follow up post. I would also be interested in posting particularly horrific examples of artefacts in halo photographs which would act as a warning to the curious. Please send any images or private correspondence to thehalovault@gmail.com.

Update

To clarify following enquiries, the above image is 8-bit and is used for illustration purposes only. In my experience, 8-bit are always unusable for this type of work. 16-bit are variable and one cannot predict when or why artefacts will appear. When they do manifest, they are every bit as unusable as 8-bit, at least with my equipment.

11 comments:

  1. Interesting. I have my Nikon D60 (total low-end DSLR) for almost 8 years now and I have experienced no artefacts whatsoever when shooting in .raw (.nef) format, JPEG has some artefacts but that's due to the compression.
    I have started to get problems just recently, my shutter passed its shutter life of 50k cycles (mine has 65k) and it shows - when the sensor is "cold" and I snap a photo just after i turn on the camera, the first photo is totally unusable.

    ReplyDelete
  2. I was getting such concentric artifacts when I processed 8-bit stuff. So I think it is something associated with compression of data. Maybe these cameras just give dishonest raw files. And it seems to me, the artifacts do not depend on the grade level of a camera. DSLRs in any case are not designed to such difficult using.

    I have a little idea how get rid of it. Alec, Marko, can you send me any batch of raw files which you did for stacks?

    ReplyDelete
  3. I had Nikon D70 for a decade. No problems in daylight. But while D3200/3300 are ok in lunar and spotlight, they are useless in daylight. (I am talking about raw here of course). So it makes to think that it is about the intensity of the light. A very dark nd-filter (optical thickness 4.0) could be the poor man's way to solve the issue. Only that the mechanism to open my 8mm Sigma is broken and I can't put filters inside.

    ReplyDelete
  4. 20 raws zipped of the plate display stack shown in my post are here:
    https://www.dropbox.com/s/0o03zdbakfmse3u/platedisplay.zip?dl=0
    One image should be enough for testing, but artefacts become somewhat more pronounced in the smoother stacked image.

    ReplyDelete
    Replies
    1. So, I was wrong, the problem is serious indeed. I tried to remove these artifacts in several different ways, but unfortunately failed. Maybe the problem lies in cmos sensor or other hardware. Because I saw the artifacts even in debayer version.

      In my case I had similar artifacts when I used raw-converter from canon. Then i started using lightroom for conversion from raw. The artifacts disappeared when the camera profile called "adobe-standart" was applied.

      https://drive.google.com/open?id=0B1TsOrx5lNP6Nzd3d3NqS0x2NFU

      Delete
    2. Thanks for trying anyway. In your example the settings choice really makes a difference. Alec mentioned about in the camera there being two color space settings, so I might try if changing it from sRGB to Adobe RGB could bring a difference too.

      Delete
    3. I'm too affected this phenomenon, and I tested today, if color space has any difference. Unfortunately, it didn't have any effect. Adobe RGB and sRGB produced similar, hellish artefacts. I'll send a mail with examples and longer as a private email, like instructed.

      Delete
    4. What is your Nikon model, Sipinen?

      Delete
    5. I "forgot" to mention it, because I wanted to follow the idea of "receive feedback > collate results" but since results have started "oozing out" there's no difference :) My model is D5200. And of course sneak preview of affected models list(gathered from comments) can be very helpful for those haloholics, who are just about to purchase new equipment! So yes: D5200 is decent camera, but if You are going to use it for roughing halos out of faint displays, be warned! (I have no experience of other DSLR's)

      Delete
    6. Yeah, this is the frustrating thing. All these cameras may be great for almost every other type of photography except ours, they are not bad cameras per se. All the cameras mentioned so far, including my own, a D3100 have been Nikon. Anyone having problems with non-Nikon equipment?

      Delete
  5. 8-bit images are the worst, totally unusable. 16-bit images are better but even those can be susceptible and at times just as bad. Indeed, there doesn't seem to be any rhyme or reason why or when artefacts will appear. One thing is certain though, if artefacts are present, then usm will exacerbate the problem.

    ReplyDelete