9

I'm preparing to go through thousands of photos and organize them and part of that is to give each one a rating of 1 to 5 stars. This has me wondering, has anyone developed a set of recommended standards for what qualifies a photo as 5 stars, 4 stars, etc. I'm looking for something that will make rating the images less subjective and give me specific photographic qualities to look for to reduce the effect of rating skew.

As an example that I just made up now, something like this:

  • 5 stars - Photo has perfect or near perfect sharpness of the subject, subject composition is as intended, light levels are perfect.
  • 4 stars - Photo composition is good, but sharpness or light levels are a bit less than perfect. Post processing may correct flaws.
  • 3 stars - Sharpness or light levels are good, but photo composition is not as intended.
  • 2 stars - photo is too blurry or light levels are . Cannot be corrected
  • 1 stars - photo is very blurred and/or has poor light levels. Cannot be corrected.

Ideally this would help me later find photos that could be corrected with additional post processing or future technologies such as advanced blur correction. I realize I could also tag such images as slightly blurry. Maybe there is a tagging standard? When I searched for these things online all I could find is how to use a 5 star rating system in a specific piece of software or how to implement one on your website, etc.

|improve this question
  • 2
    Yes, lots of people have developed sets of recommended standards. But why do you care what other people do? If a specific set of star ratings and/or tags work for what you're trying to do, then use it. – Philip Kendall Jul 24 '17 at 22:36
  • 1
    I reckon that the key is in consistency in how you rate your photos. I was once advised to use 4 and 5 stars very very sparingly...as my standard of photography improves over the years, I don't want to run out of 'available' stars when my photos improve in years to come, so that I can try and maintain consistency of ratings over the entire library – laurencemadill Jul 25 '17 at 8:09
  • 2
    I personally only rate the subject matter. An image with exceptional content gets one star. (The other stars don't really matter since I'm the only one who analyses my own pics. I know what I like. The star is just for record-keeping.) If the image is technically perfect and has no content, it gets no stars. Imperfect images with no content get the trash can. – unperson325680 Jul 25 '17 at 13:01
  • 1
    Not worth a whole answer, but could be considered interesting. In the VOIP world we do something called subjective testing where we place various people in front of a device be it video/audio/both they watch/listen to various snippets, where the quality has been changed in each one. They then have to rate the thing they've seen/heard. Using what has happened, e.g. horrible blocky video and the mean score of the subjects, we generate a rating score of the video quality! – Crazy Dino Jul 25 '17 at 21:36
  • 2
    Related: What is your workflow for “don't care” photos? What is an efficient workflow for picking photos from a large set? What's a good strategy for choosing which photos to keep? – Jere Kupari Aug 4 '17 at 21:16
16

There are no standards as much of it is subjective. Even though I consider myself to be very rigorous, I find that there is a drift over time. This the intention of my rating system:

The 3-star mark is what I reserve for a perfect photo: The subject is sharp and no major area is under or over-exposed. Framing is such that the subject is shown clearly and nothing detracts from paying attention to the subject.

A photo which is missing one of those qualities gets 2 stars. One that has all those qualities plus a compelling subject and no unwanted elements in it gets 4 stars.

For a photo to get 5 stars it needs to qualify for 4 plus offer no opportunity for improvement. Depth of field must exactly cover the intended subject, sharpness must be impeccable and framing must show the subject in a most flattering angle.

Despite a precise idea, there is some room for interpretation, so I usually give myself some calibration rounds where I rate a subset of photos over and over until I give each photo the same rating each time while going through the set.

One must consider the goal of a rating system but 5 stars is so coarse that I find myself often wishing Lightroom allowed for 1/2-star steps. When putting the ratings into goal-oriented terms, my ratings can be interpreted as follows:

  1. A photo kept only for sentimental value. Might never be shown (or seen again). Technical flaws can be numerous but not enough to warrant deletion (which happens for 70-80% or so of my images but that's another discussion entirely).
  2. Interesting photo with some technical flaws. Not quite good enough to show but interesting for souvenir and recollection.
  3. This is an almost technically perfect photo, potentially improved with straightening, cropping or dust removal but the adjusted results show good focus, sharpness, exposure and interesting compositions with a clean discernible subject.
  4. Photo that deserved to be printed. Every bit as good as 3 stars but with a strong and interesting point of interest, although it leaves room for improvement such as a less-than-perfect angle or not ideal depth of field.
  5. The perfect contest-worthy photo. Must be able to make an excellent print of a unique subject or original view of a known subject.

Over time as my skills and ability to analyze images improve, I find the the ratings I would give for any image goes slightly lower over the years.

One principle which I believe in is that the rating of each photo must be independent of each other. It makes it easier to rate but poses some difficulty when selecting. For example you can have several equally perfect images of a certain subject, all given 5 stars, but then just filtering for 5 stars makes for a repetitive set.

|improve this answer
  • I don't think perfect means what you think it means. – Caleb Jul 25 '17 at 18:56
  • 1
    The point is that a perfect photo is a technicality which I consider as the starting point to making a good photo. One can have a perfect photo of a napkin and it probably wouldn't interest me much. – Itai Jul 25 '17 at 19:14
17

There is no absolute standard, nor is there any standard that is generally applicable or generally accepted. There are certainly certain situations where specific systems are used — your example list could be one. Or you could use this:

xkcd

Crucially — see the discussion in the comments below! — any one-dimensional scale can't possibly include every important way to think about a photograph. Any general system would at least need multiple axes.

|improve this answer
  • 2
    Raising the question: unless the photos are crime scene documentation, why keep anything less than four stars? – user50888 Jul 25 '17 at 1:23
  • 1
    @benrudgers Sometimes you shoot one that's a technical train wreck but the subject is compelling enough to make it worth giving it enough arbitrary bonus stars to push it above the throwaway line. – Blrfl Jul 25 '17 at 3:20
  • 4
    @ben, as Itai's answer puts it, sentimental value. The first time I went shooting with my manual focus 500mm I missed focus on every single shot. I kept one out of about 500, because it was almost in focus and it was my first gray wagtail. Curiously it's one of my most liked photos on Flickr, possibly because people mistake the subject and think it's a decent photo of boat. – Peter Taylor Jul 25 '17 at 7:11
  • 1
    @benrudgers I've shot photos I thought were useless but later were of value to customers looking for a rare subject matter and couldn't be picky about the technical quality of the shot. – jwenting Jul 25 '17 at 13:44
  • 1
    @jwenting Potential salability is another criterion missing from the scale. If I were to critique the question, rather than answer it, the limited dimensions which the proposed ratings consider would be the heart of that critique. The Kent State photo in my link, might be an example of the reasons technical criteria might miss the point...all test chart photos would get four or five stars under the proposed system. But that's a critique not an answer. – user50888 Jul 25 '17 at 13:50
2

As others said, there are no standards and you need to find a system that makes sense to you. The challenge is in "compressing" multiple dimensions of evaluation into a single scale and do so consistently.

Being aware of those dimensions (and the different ways of using them might help. So here we go:

  • Does the rating reflect the image an aesthetic or a technical evaluation?
  • Does the rating reflect the image in its current state or its potential? (i.e. can the rating be improved by editing the image? Perhaps another way of thinking about this is via an anslogy from the film era: are you evaluating the negative as a negative or are you evaluating a particular print from that negative? (The problem, of course, is that, in the digital world, you can only ever look at (preliminary) prints, never at the negative itself.).
  • Is the rating about about how suitable the image is for a particular purpose? If so, what is the purpose? (Exhibition? Photo book? Social media? Slideshow on your TV? Document an event?) Or can the purpose change from picture to picture?
  • Are you going to use the rating later on in your workflow (e.g. to decide which images to prioritise when editing or where you gradually increase your rating in the editing process) or is it merely for filtering later on when looking for the best images about a particular topic?
  • Does your rating also imply a keep vs. reject decision or are ratings independent? If they are independent, a zero star rating just means "unrated" whereas it would otherwise mean "rejected". Likewise, if four or five stars imply "keep", then you will usually have to make sure that you only give those high ratings to one image in a series of similar images. (I think this alone is a reason to keep ratings separate from keep/reject decisions.)

It seems to me that the best way of consistently combining the multiple dimensions involved is by evaluating the picture's potential usefulness. But then again that's just my perspective.

The advantage of this "utalitarian" approach is that it also forces you to answer the question "Why did I take this photo?" or "What do I want to do with this photo?"

Of course (as with any rating system) there are some edge cases where following this approach may seem counterintuitive. For example, if you are a photography instructor and want to demonstrate the effect of very high ISO, the extremely noisy image of a brick wall will do the job perfectly and hence the image should be rated with four or five stars, even though it is clearly a bad picture that would normally be rejected. If you don't like this "incoherence" (which really is coherence, actually) nobody is going to stop you from making exceptions to your rule, but I would argue it's better to stick to the good rating because it will also prevent you from accidentally deleting the image. In order to prevent the image from showing up when you are searching your database for an excellent wall picture, just don't tag it as "wall" but rather "photography 101".

|improve this answer
0

You forgot one technical aspect... The emotional one. Oops, that is not technical.

Sometimes there are more important factors than sharpness, than exposure or dynamic range.

Is the photo the only one of an historical event?

Is the photo very emotional?

Is the photo outstanding in concept, in composition, regardless of some technical flaw?

I would reserve a 5 star if the photo is one that I would hang on my wall.

But this obviously makes a rating system subjective.

One interesting thing that @Christoph commented is that if a picture can improve with some editing. Some white balance, contrast or a bit of sharpening; some Tone mapping. Inclusive a slight reframing.

Again this is a bit subjetive.

|improve this answer
0

Here's my personal way to use the stars (probably not standard but probably not unique either).

If I have 300 photos to sort through, to output around 30 good ones:

  1. I will do a first pass looking at them fullscreen, and mark as 1 star all the usable ones.

  2. I then filter to only show 1 star or above, and pass through this list fullscreen again. There are often a few photos of a same subject in a row, so if i see three different 1star photos of the same subject, i'll add 2 stars to the best of them.

  3. I keep raising the star level and comparing photos that are of the same quality until i reach the desired quantity of photos to output.

This is my favorite way to sort through high quantities of photos taken at en event (wedding, party ..).

I tend to reserve 5 stars for photos that'll need specific work, like stitching a panorama together or combining faces from 2 group shots.

|improve this answer

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.