Skip to content
#small language model Open access

The "Depth" of Depth of Field Is a Convention, Not a Measurement ── With the same 50 mm lens at the same f/2.8, the hyperfocal distance is 29.8119 m and also 207.6912 m ── What changed is not the equipment but the circle of confusion, which is an agreement ── One set of equipment holds four answers ── [Paper 358]

Sep 2026 · Zenodo (CERN European Organization for Nuclear Research)
Image Processing Techniques and Applications

Abstract

Depth of field is printed in tables and looks like a property of the lens. This paper asks where that number comes from──the answer is that it comes not from the equipment but from an agreement about when blur becomes visible. No new mathematical theorem and no new law is claimed. Scope of this paper (scope note): No new mathematical theorem and no new law is claimed──the circle of confusion, hyperfocal distance, and the depth-of-field formulae are all standard. We do not build optics──the theory of the diffraction limit, aberrations, and the point spread function is not entered. It is named and no more. Vision is not treated──the limits of human acuity and the psychophysics of viewing distance are not entered. Four provenances for c are simply set side by side. We evaluate no equipment──the performance of particular lenses and cameras is not the subject. We give no photographic advice──which c one ought to choose is not addressed. Diffraction is not included──the diffraction blur of a small aperture is not in the numbers. The computation is within the geometric-optics approximation. Relation to earlier papers: Paper 352 showed a transition temperature moving with the measurement’s time scale──there what moved was the observation time; here it is an agreement about the limit of what can be told apart. Paper 355 showed restitution not to be a material constant──there the condition was impact speed. Paper 354 showed length ceasing to be the quantity that measures a curve──there too the choice of ruler fixed the answer. Paper 300 showed whether two things share a root is decidable──by that test, focal length and circle of confusion have distinct roots: the first is measured, the second is decided. What is added is setting four provenances for c side by side with the resulting hyperfocal distances, showing the answers spreading by a factor of 6.9667 on identical equipment, showing that changing the aperture does not remove the four, confirming the formula by matching the traditional value against the accepted figure, and putting the separator on c. First, the formula contains the circle of confusion c, and c is not a measured value (Section 2). Second, this is the core of the paper. At the same 50 mm and f/2.8, changing c from 0.0300 to 0.0043 mm takes the hyperfocal distance from 29.8119 m to 207.6912 m, a factor of 6.9667 (Section 3). Third, not one thing about the equipment has changed (Section 3). Fourth, stopping down to f/8 does not remove the four answers; they line up from 10.4667 to 72.7244 m (Section 3). Fifth, the traditional c=0.030 mm at f/8 gives 10.4667 m, matching the accepted “about 10 m” (Section 4). Sixth, the separator is c, not the lens and not the aperture (Section 5). Depth of field is printed in tables and looks like a property of the lens──but the number is not fixed by the equipment alone. The denominator of H=f^2/(Nc)+f holds the circle of confusion c──f and N are engraved on the lens and anyone measuring gets the same value, while c is the limit at which blur becomes recognisable as blur, and where that happens depends on the size viewed, the distance viewed from, and who is viewing. So c is not a property of the equipment but a stipulation about viewing conditions. And there is more than one stipulation──the traditional 35 mm 0.030 mm, the stricter 0.020 mm, the high-pixel-count 0.015 mm, and the pixel pitch 0.0043 mm. All four have grounds and none of them is wrong. So one set of equipment holds four answers──at f=50 mm and f/2.8 the hyperfocal distance is 29.8119 m, 44.6929 m, 59.5738 m, and 207.6912 m. A spread of 6.9667, with not one thing about the equipment changed. Stopping down to f/8 simply shortens them all, to 10.4667 through 72.7244 m, leaving the spread ratio unchanged: the aperture is no substitute for the agreement. In landscape work, “where do I focus” answered as 29.8 m or as 207.7 m is a different way of shooting. The formula itself is right──the traditional c=0.030 mm at f/8 gives H=10.4667 m, agreeing with the accepted “about 10 m”. A smaller c lengthens H, and the f/2.8-to-f/8 ratio 2.8483 falls just short of the f-number ratio 8/2.8=2.8571 because of the +f term. The trouble is not the formula but that c arrives from outside it. One thing separates them──the circle of confusion; not the lens and not the aperture. c acts on H with the same inverse dependence as N, so inside the formula they look alike ── but one is a measurement and the other a convention. Fix a rule computing c from viewing conditions and it moves toward being measured, yet the agreement about what acuity to assume remains: an agreement changes place, it does not vanish. To be said honestly──this paper does not say depth of field is meaningless. State c and the number is completely determined. What is said is that one must not report a depth number without reporting c. One last thing──if even one agreement sits inside a formula, the number coming out is a function of that agreement. Depth of field looks as though it is set by f and N, and is in fact set by c. Unless one keeps track of which quantities are measured and which are stipulated, a stipulation gets mistaken for a measurement. On the making of this work: The ideas and content of this work stem from the author's own considerations. Assistance from an AI (a large language model) was used for structuring, English translation, and checking the algebra. Any remaining errors or misinterpretations are solely the author's. Feedback and corrections are sincerely appreciated. ----- 被写界深度は表に載っており、レンズの性質のように見える。本稿が問うのはその数がどこから来るかである──答えは機材からではなく、「これ以上ぼけたら分かる」という約束から来る。新しい定理も法則も主張しない。 本稿の射程(射程注記):新しい定理も法則も主張しない──許容錯乱円・過焦点距離・被写界深度の公式はすべて既知である。光学を作らない──回折限界・収差・点像分布関数の理論には立ち入らない。名前を挙げるにとどめる。視覚を扱わない──人間の視力の限界や観察距離の心理物理には立ち入らない。 c の由来を四通り並べるだけである。機材を評価しない──特定のレンズやカメラの性能は主題ではない。撮影法を論じない──どの c を選ぶべきかという実務上の助言はしない。回折を含めない──絞りを絞ったときの回折ぼけは数値に入れていない。幾何光学の近似の上での計算である。既刊との関係:論文352 は転移温度が測定の時間尺度で動くことを示した──そこでは動かすのが観測時間だった。ここでは「見分けられる限界」という約束である。論文355 は反発係数が材料の定数でないことを示した──そこでは条件が衝突速度だった。論文354 は長さが曲線を測る量でなくなることを示した──そこでも物差しの取り方が答を決めた。論文300 は同根か別根かが判定できることを示した──焦点距離と許容錯乱円は、その基準で別根である。前者は測れるが、後者は決めるものである。加えたのは、c の由来を四通り並べて過焦点距離を数に出したこと、同じ機材で答が 6.9667 倍に開くことを示したこと、絞りを変えても四通りが消えないことを示したこと、伝統的な値が通説と合うことで式が正しいことを確かめたこと、分離子を c に置いたことである。 第一に、被写界深度の式には許容錯乱円 c が入っており、c は測定値ではない(第2節)。 第二に、これが本稿の芯である。同じ 50 mm・F2.8 で、c を 0.0300 から 0.0043 mm に変えると過焦点距離は 29.8119 m から 207.6912 m へ、6.9667 倍になる(第3節)。 第三に、機材は一つも変わっていない(第3節)。 第四に、絞りを F8 にしても四通りの答は消えず、10.4667 から 72.7244 m に並ぶ(第3節)。 第五に、伝統的な c=0.030 mm・F8 の答 10.4667 m は、通説の「10 m 前後」と合う(第4節)。 第六に、分離子は c であって、レンズでも絞りでもない(第5節)。 被写界深度は表に載っており、レンズの性質のように見える──だがその数は、機材からだけでは決まらない。式 H=f^2/(Nc)+f の分母には許容錯乱円 c が入っている──f と N はレンズに刻印されていて誰が測っても同じ値になるが、c は「これ以上ぼけたら、ぼけていると分かる」という限界であり、どこで分かるかは見る大きさ・見る距離・見る人によって変わる。つまり c は機材の性質ではなく、鑑賞の条件についての約束である。そして約束は一つではない──35 mm 判の伝統的な 0.030 mm、厳しめの 0.020 mm、高画素機の 0.015 mm、画素ピッチの 0.0043 mm。四つとも根拠があり、どれかが誤りというわけではない。だから同じ機材が四通りの答をもつ──f=50 mm・F2.8 で、過焦点距離は 29.8119 m・44.6929 m・59.5738 m・207.6912 m。開きは 6.9667 倍で、機材は一つも変わっていない。絞って F8 にしても 10.4667 から 72.7244 m へ全部が短くなるだけで、開きの比は変わらない。絞りは約束の代わりにならない。風景写真でどこにピントを置くかの答が 29.8 m と 207.7 m では、別の撮り方になる。式そのものは正しい──伝統的な c=0.030 mm・F8 で H=10.4667 m となり、通説の「10 m 前後」と一致する。 c を小さくすれば H は伸び、F2.8 から F8 への比 2.8483 は絞り値の比 8/2.8=2.8571 にわずかに足りない。問題は式ではなく、c が式の外から来ることである。分けているものは一つ──許容錯乱円であって、レンズでも絞りでもない。 c は H に対して N と同じ反比例で効くので式の中では同格に見えるが、片方は測定量で他方は規約である。鑑賞条件から c を計算する規則を固定すれば測定量に近づくが、「視力をいくつにするか」という約束は残る。約束は場所を変えるだけで消えない。正直に言えば──本稿は「被写界深度は無意味だ」とは言っていない。 c を明記すれば数は完全に確定する。言っているのは、c を書かずに深度の数だけを報告してはいけない、ということである。最後に一つ──式の中に一つでも約束が入っていれば、出てくる数は約束の関数である。被写界深度は f と N で決まるように見えて、実は c で決まっている。どの量が測定でどの量が規約かを分けておかないと、規約を測定と取り違える。 作成にあたって:本稿の着想と内容は、著者自身の考察に基づくものです。文章の構成整理や英訳、数式の確認には AI(大規模言語モデル)の助力を得ました。最終的な内容の解釈や誤りがあれば、それらはすべて著者の責に帰します。お気づきの点があれば、ご教示いただければ幸いです。

View source

Similar papers

#small language model Dataset Open access Oct 2026

Socratic guiding questions in synthetic arithmetic data: matched LoRA runs (revision v2)

Supporting data, adapters, predictions and code for the article *Low-Cost LoRA Fine-Tuning of Small Language Models for Multi-Step Arithmetic Reasoning* by Jake O'Grady, Asena Isik Gürhan, Chee Fong Ting and Effirul Ramlan (University of Galway). We generated 20,000 GSM8K-derived arithmetic problems with step-by-step s...

O'Grady, Jake, Gürhan, Asena Isik, Chee, Fong Ting et al. · 465 citations
#computer vision Open access Jun 2016

Software Development in Startup Companies: The Greenfield Startup Model

The results are packaged in the Greenfield Startup Model (GSM), which explains the priority of startups to release the product as quickly as possible, and the need to shorten time-to-market, by speeding up the development through low-precision engineering activities.

Carmine Giardino, Nicolò Paternoster, M. Unterkalmsteiner et al. · 178 citations · ⚡14
#computer vision Open access Oct 2016

Software Startups - A Research Agenda

Software startup companies develop innovative, software-intensive products within limited timeframes and with few resources, searching for sustainable and scalable business models.

M. Unterkalmsteiner, P. Abrahamsson, Xiaofeng Wang et al. · 157 citations · ⚡17
#machine learning Review Open access Oct 2016

“Failures” to be celebrated: an analysis of major pivots of software startups

This study conducts a case survey study based on the secondary data of the major pivots happened in 49 software startups, and demonstrates that customer need pivot is the most common among all pivot types.

Sohaib Shahid Bajwa, Xiaofeng Wang, Anh Nguyen-Duc et al. · 127 citations · ⚡15
#computer vision Review Open access May 2015

A survey study on major technical barriers affecting the decision to adopt cloud services

The comparison of adopter and non-adopter sample reveals three potential adoption inhibitor, security, data privacy, and portability, which underlines the importance of the technical and security perspectives for research investigating the adoption of technology.

Nattakarn Phaphoom, Xiaofeng Wang, S. Samuel et al. · 111 citations · ⚡8
#computer vision Open access Feb 2018

Lean Internal Startups for Software Product Innovation in Large Companies: Enablers and Inhibitors

This study investigates how Lean internal startup facilitates software product innovation in large companies and identifies its enablers and inhibitors, and shows the potential of the method-in-action framework to investigate the Lean startup approach in non-startup context.

Henry Edison, Nina M. Smørsgård, Xiaofeng Wang et al. · 78 citations · ⚡6

Related blog posts

MIT News · Artificial Intelligence Oct 15, 2026

AI model decodes the language cells use to communicate

A new computational approach reveals how cells communicate to assemble into tissues. This could accelerate stem cell engineering, improve disease models, and inform therapies.

Microsoft Research Blog Oct 15, 2026

Proof-based general-purpose computing with Magnetar

Zero-knowledge proofs can verify computing results without rerunning workloads or exposing sensitive data. Magnetar helps bring them into the mainstream with support for standard programming languages, major speed gains, and quantum-resistant security. The post Proof-based general-purpose computing with Magnetar appeared first on Microsoft Research.

We use cookies to run the site and, with your consent, for analytics and to show ads. See our Cookie Policy.