On the DPR Medium Format Forum, a poster has advocated for use of the term “image fidelity” as a scalar description of photographic technical quality. He has referenced the term “high fidelity” in audio as an analog and an exemplar.
I am unconvinced that this new terminology offers any advantages over the conventional term “image quality” to describe the same thing. Further, based on the history of consumer audio reproduction, I think it invites confusion.
I have a deep background in audio. I’ve done research on speech bandwidth compression, speech recognition, and constructing models of the human hearing system. I have designed speech communications systems. I have designed and built music reproduction systems for many years, including preamps, amps, active crossovers, and room environment simulators. I have owned and used high-end gear, including exotica like three-piece DSC digital processors, Tact equalizers, Levinson preamps, studio CD systems, moving-coil cartridges, minimalist Scottish turntables with active line conditioning, ribbon tweeters, planar speakers, linear tone arms, electrostatic headphones, 200-pound mono amplifiers, single ended triode amp, Nelson Pass class A amps, and Wilson Audio speakers in concrete enclosures (don’t try to move one of those alone). I’ve been a keen observer of hi-fi progress and the attitudes and convictions of self-professed audiophiles since the sixties.
Here’s what I’ve seen.
Originally the term high fidelity had an unambiguous, if difficult-to-assess, meaning. The word faithful means “true to the facts or the original.” The word fidelity means “the degree of exactness with which something is copied or reproduced.” So, the ultimate objective of high-fidelity audio was to reproduce the original sound field. However, this ship immediately foundered on the rocks of practicality. To reproduce the sound of a symphony orchestra playing Beethoven’s Ninth Symphony would require power levels and dynamic range great enough to make such systems unaffordable. The number of speakers necessary to reproduce an orchestra’s three-dimensional sound field would be staggering.
So the search for high fidelity became constrained by economics. Also, a critical mass issue immediately made itself felt. Most people who wanted to listen to faithful reproductions of musical performances didn’t want to make their own recordings, but instead they wished to purchase those made by others. For there to be an economically viable market for such recordings, there needed to be a lot of people with compatible playback equipment. That meant that you couldn’t buy 64-channel recordings of orchestral performances, because there weren’t enough quality-obsessed folks out there to make producing those a money-making proposition.
Since a reproduction indistinguishable from the original event was not practical, the people making the recordings and the people listening to them both had to compromise. And – surprise, surprise – not everybody could agree on what compromises should be made. Indeed, as time went by, the high-fidelity world splintered into different camps, depending on what compromises were prized by various groups of listeners.
Then things got worse. Some of the camps that had many vociferous advocates, such as single ended, low-feedback vacuum tube amplifiers, were objectively much worse than other alternatives. It appeared that the search for high fidelity had become replaced but the search for aurally pleasing departures from a faithful reproduction. It was at that point that objective testing became despised by many who considered themselves audiophiles.
Objective testing is repeatable, quantitative, and, done right, authoritative. However, before you can test for something, there needs to be agreement that testing for that thing is aurally meaningful. Over time, more and more things with aural impact were discovered. Static frequency response, hum, noise, and static distortion, both single-frequency and intermodulation, were the bedrock tests, but crossover distortion, transient intermodulation distortion, jitter, and a host of other degraders of fidelity were discovered, researched, and added to the pantheon of objective tests.
The additional tests did not bring peace and agreement to the audiophile community. Magical thinking arose, with some people paying thousands of dollars for two conductors insulated from each other and assembled into a cable, in the belief that those would elevate their listening experience.
Then arose another attempt to bring some unity to the assessment of fidelity. In many psychological tests, double blind protocols had proven useful. Some groups, led, as I remember, by the Boston Audio Society, promoted a type of testing called ABX testing. People who claimed to hear immense differences among products when they knew what they were listening to found that in many cases they couldn’t tell the difference when they didn’t know which products were used.
This finding did not produce a kumbaya moment. The people who wanted to hear, but couldn’t hear, differences said that the test was flawed, and even advanced the scientifically strange notion that the stress induced by not knowing what they were listening to destroyed their ability to hear what they could hear if they knew what gear was being used.
None of the sturm und drang surrounding double blind testing surprised oenophiles. In that community, of which I am a long-time member, there is a concept called “drinking the label.” This describes a situation in which the drinker knows what wine they are consuming and finds that how the wine tastes is influenced by how they think it should taste. Some people call this “confirmation bias.” In wine tasting, it is routine to deal with this by conducting all serious comparisons double-blind. That consensus has never been accepted in the audiophile world.
So now we have the situation in high-fidelity that the once-agreed goal of reproduction of the original performance is not universally accepted, and the now-diminished audiophile population is no closer to agreement on what’s right and what’s not than they ever were.
There are many parallels between photographers’ search for visual nirvana and audiophiles’ quest for auditory enjoyment. It is not economically feasible to create a replica of the original scene in a photograph. A photograph is two-dimensional; the world in general is not. Photographic reproduction can’t handle the range of brightnesses in real-world scenes. Producing a high-resolution image with the spectral characteristics of the original scene throughout would cost a fortune. Diffraction limits sharpness. Perfect lenses don’t exist at prices that consumers are willing to pay.
So the search for images that are faithful to the original scene, like in audio, boils down to pick your poison. How much of what departures from perfect reproduction of the original scene are acceptable to you? Don’t expect that your answer to that question and my answer to that question will be the same.
It gets worse. Most photographers, like a great many makers of audio recordings, don’t want to accurately reproduce the original stimuli. They want to put their own stamp on things. So fidelity to the original isn’t what is valued.
Given all that, I think that we photographers should stick with image quality rather than invent a new term for something that is, if not the same, is very similar.
From Greg Johnson:
It was an honor to be the instigator of yet another impressive Kasson editorial and thread. There have been so many…. As all of you know, I am the one on the DPR MF forum that started this and likes to use the term image fidelity vs image quality. I don’t do a good job with the definition of image fidelity since a lot of it is what we see and perceive, is opinion based and therefore can be somewhat subjective, but I prefer the term and would like to use it unless it offends the forum. I’m not even sure I could properly define IQ. I just prefer Image Fidelity when I talk about MF’s biggest current advantage.
Last week I briefly and in passing related the term image fidelity and what photographers see to hi fidelity and what audiophiles hear and their arguments about both, but it is probably not really a good comparison and I only mentioned it once as part of a post on some unrelated thread.
I have used the term image fidelity for a while now and never really equated it in my mind to hi fidelity until late last week when someone asked me what I meant by image fidelity. I said in one post (sort of in passing) that image fidelity was sort of like the term hi fidelity because photographers argue about IQ in terms of what they see and audiophiles tend to argue with each other about what they hear or what sounds the best to them.
I wasn’t thinking of it in terms of reproducing the “original” accurately. When I say “image fidelity” I don’t mean that it has reproduced the exact original scene better or worse. When it comes to IQ or image fidelity in photography, there is a lot of room for corrective and other manipulations, just as in the use of the term high fidelity as heard via sound recordings.
When I considered myself an audiophile as a younger man and spent far too much money that I didn’t have on turntables, amps, revievers, pre amps, styluses and especially very expensive and large speakers, I didn’t care about how well it depicted whatever the original instrument, voice, or mix was, I just cared about how it sounded. We argued endlessly over that. It was subjective and for the most part, pretty ridiculous. But I sure liked my sound system gear, sort of like I like my camera gear now and the superb image fidelity of MF.
When we “audiophiles” argued about and described what we heard, the guys I ran with didn’t argue over graphs or measurements. We sounded like the Leica Forum when some of those guys described the look of their Leica images. We talked about entering another dimension of sound, a deep richness, a tonal beauty and delicate balance of luminous vibrosity and other silly descriptions describing what we heard. We used terms like balanced, detailed bright, muddy, veiled, bassy, tinged, bleed, richness, peppery, salty, delicate, bloated, burped, hi, low, fast, focuses, tingy, weepy, glassy, watery, blurred, dark, flat, deep, dimensional, boomy, boxy, breathy, crisp, dull, sharp, and 50 other terms we would throw in to describe what we thought we heard.
I remember on those forums that some tester would pop in and say, you can’t measure boxy! And we would laugh and start arguing about which speakers have the boxiest sound.
I guess I could have made an analogy to wine tasting when I describe the term image fidelity. Don’t get me started on that because I spend way too much on good bottles of wine and was part of a wine collective for many years that met weekly to describe what we tasted in a 15 dollar bottle vs a 115 dollar bottle of some particular wine type using hundreds of terms that would split your head. Want me to list them? If I ever again hear someone say (when tasting a zin) that it is “peppery and has a hint of blackberry” I think I will scream and run away.
So, I prefer the term image fidelity to image quality. But it is hard for me to explain to you why. It has a nice ring to it. I like it…. So does Fuji because I think they are going to use it on their next advertising push. Should I make them give me a lens to give them the rights? LOL.
https://www.thewinecellarinsider.com/wine-topics/wine-educational-questions/davis-aroma-wheel/
I somewhat resonate with this.
We as consumers (and producers) of image media have grown used to tropes of images delivered by real cameras and lenses.
There are good things (sharpness, contrast?), bad things (softness, smearing, CA, OOF PSF texture, field curvature, lens flare), personal preference (OOF PSF edge brighness), things that are difficult to even describe in words why some lenses are good and some are bad (transition between in-focus to out-of-focus).
But a little bit of a bad thing can be a good thing. For example, in computer graphics they use the fact that real lenses are modeled in renders to add “realism”. Lens flare is added to video games to help convey the brightness of the sun or streetlights or explosions, even though manufacturers fight to remove lens flare from real lenses.
I use classic Zeiss lenses from the C/Y system, and my favorite is the 85/2.8, which is quite sharp by most measures, but it isn’t “biting sharp”. Still, I like it more for landscapes than the 100/2.8 which is so much sharper that things look… harsh? I don’t know how to convey it. It doesn’t help that the 100/2.8 has edge-heavy background OOF PSFs while the 85 has really smooth edges and a nice transition (again, impossible to articulate why).
Likewise, I have the Tessar 45/2.8 and one modern lens of the same focal length, the Tamron 45/1.8 VC. I’ve compared them on the same subjects and always prefer the Tessar images despite it being very definitely being flawed in technical ways. The extreme corners are soft, the overall sharpness isn’t anything to write home about, and the bokeh has a weird bright spot in the middle. But the Tamron, which is sharp, fast aperture, image stabilized, weather sealed, autofocus (though I rarely use that), and has very neutral OOF PSFs (at least when far out of focus) just always looks unpleasant to me, especially when stopped down. I don’t know why, but every time I use the Tamron for its aperture or weather sealing I always wish I had the Tessar instead.
What is it that makes some lenses good despite flaws, or unpleasant despite lacking technical flaws?
And why do some lenses seem to have it all (my Hasselblad 180/4) and others neither (old aps-c kit zooms)?
How can lens designers teach their software to optimize for these qualities that can barely even be articulated in words, let alone math? Sharpness is easy, at this point, with computer-controlled manufacturing. You need more than just sharpness though.
With all due respect, Greg, your approach to language here seems a lot like Humpty Dumpty’s: ‘“When I use a word,” Humpty Dumpty said in rather a scornful tone, “It means just what I choose it to mean—neither more nor less.”’
I recall those audiophile days; people argued about $100/foot pure silver speaker cables and other such nonsense. Listening is passive, photographing is active. Audiophiles produce nothing when they listen, photographers, with the exception of commercial and journalists are often attempting to produce “art.” To quote one of my favorite people, Charlie Cramer, “there are no wrong answers, it’s all art.” If I want my images soft, that’s my privilege.
I’ve found that in over 60 years of photography, content is what matters. In all the photographs I’ve shown to people, “regular” people and photographers, no one has commented about the “bokeh.” Of course people can see it if one points it out and perhaps there are subconscious factors at play when a person admires or criticizes an image, but in the end, it’s content.
Arguing about terminology eg “image quality,” or IQ, versus “image fidelity,” is just a semantic exercise, ultimately meaning nothing. I like high quality cameras and lenses because they give me a great starting point for my art. I can make images all sharp with focus stacking or with selective focus, it’s my choice. You can like either or neither.
My basic point is that unless one is being paid to produce something specific, it should be at the discretion of the artist and it should be fun!
A village in my home country is famous for its “healing” honey. Many people paid good money for that and we heard all kind of medical success stories by those who used that (the only thing the damn honey couldn’t fix was amputated leg!). Turned out the real production of that village is less than ten percent of the total sales. The rest was provided by beekeepers far from that village.