Blender 5 Film Flash, Part 3: Meter and Match the Camera

Blender 5 film-flash series: 1 · Place and shape the light · 2 · Iterate lighting in real time · 3 · Meter and match the camera · 4 · Build and load the background · 5 · Style materials and the film look

This is the metering chapter of a five-part Blender exercise. The subject model and a plausible set already exist; now we will make the virtual camera behave like a known film-camera setup. The reference is a Fujifilm GW690III, 120 film metered at EI 80, and a manual electronic flash. The same stop-based method supports 135, 120, 4×5, and 8×10 film-format plans. Its extended planning envelope is f/1.0–f/32 for aperture and 1/125–1/1000 s for synchronized shutter speed, provided the real camera, lens, and flash system can actually use the selected values.

The important word is approximate. Blender can match framing, depth of field, relative illumination, shadow placement, and stop changes. It cannot infer a film stock’s sensitometry, a flash tube’s pulse shape, development, reciprocity behavior, or scanner decisions from an EI value. This chapter uses a measured reference so the virtual and real setups have a common anchor.

The result we are building

Start with this repeatable reference shot:

Camera body: Fujifilm GW690III
Frame: 56 × 82.6 mm
Lens: fixed 90 mm
Film: 120, metered at EI 80
Reference exposure: f/8 at 1/125 s
Flash: manual, metric GN 25 at ISO 100 and the chosen head coverage, full power
Flash distance: 2.8 m from flash head to subject
Color reference: neutral gray card at the subject

At EI 80, a direct GN 25 flash has an effective guide number of about 22.4. At 2.8 m, 22.4 ÷ 2.8 gives approximately f/8. That makes the reference internally consistent before we open Blender. A diffuser, bounce, flash-head zoom, or unusually reflective room changes the real answer, so the final set still needs a flash meter.

1. Pick a film format, then configure the camera

Set Scene Properties › Units › Unit System to Metric and leave Unit Scale at 1.0. One Blender metre now represents one real metre. Verify the subject with the Measure tool; unit labels do not repair a model imported at the wrong scale.

Choose one format before framing or placing lights. These are landscape starting presets; with Sensor Fit set to Horizontal, Blender derives sensor height from the output resolution’s aspect ratio:

The format names are not exposure settings. They change field of view and depth-of-field behavior for a given focal length and framing, but they do not change the guide-number equation. 120 is a roll-film stock, not one frame size: a 6×4.5, 6×6, or 6×7 camera needs its own gate dimensions and output ratio. The 135 row assumes the common 36 × 24 mm full frame. This series chooses the GW690III’s measured 6×9 gate for the 120 preset.

The 4×5 and 8×10 rows use nominal sheet dimensions as convenient planning frames. A real holder masks a slightly smaller image area, and different holders can vary. For precise field-of-view matching, measure the actual exposed gate or a processed negative and replace both the Sensor Width and output ratio. Keep the real lens’s focal length; the “normal” values above are starting points, not equivalence conversions or mandatory lenses.

For fast switching, duplicate the configured camera and name the copies CAM_135, CAM_120_6x9, CAM_4x5, and CAM_8x10. Do not change format inside an A/B lighting comparison: make one camera active, lock its transform, and keep its output ratio until that comparison is finished. Output resolution belongs to the Scene rather than the Camera object, so selecting a different camera does not switch the ratio automatically; enter the matching output dimensions from the same table whenever the format changes.

For the main exercise, select CAM_120_6x9 and enter these values in Camera Data Properties:

In Output Properties, use the negative’s aspect ratio, 82.6 ÷ 56 ≈ 1.475. A convenient preview is 2950 × 2000; a smaller preview may use 1475 × 1000. Pixel count changes render resolution, not framing, as long as the ratio stays the same.

Frame the subject in camera view with Numpad 0. Enable Lock Camera to View only while finding the position, then turn it off so normal viewport navigation cannot alter the shot. Keep the camera-to-focus-target distance at least 1 m because the real 90 mm lens does not focus closer. The Blender camera origin is an optical viewpoint rather than a modeled film plane, so treat this minimum as a practical match, not a mechanical measurement. Leave a little edge safety for the real rangefinder’s finder coverage and parallax.

The physical GW690III has these limits:

The static planning choices supported by this series are:

The broader f/1.0–f/32 aperture range and 1/125–1/1000 s shutter range belong to alternative film bodies, lenses, or a purely virtual experiment. They do not turn the fixed GW690III lens into f/1.0 or its shutter into 1/1000 s. For another body, replace the sensor, lens, focus limit, and sync limit with that camera’s specifications before trusting the preview.

Here, 1/1000 s means the camera’s shutter/synchronization target. Flash duration is a separate specification, usually reported as t.1 or t.5. A continuous Blender Light object does not simulate the flash tube’s pulse duration.

2. Understand which camera fields do not meter the render

Blender does not provide a coupled still-camera exposure triangle. In Blender 5.2:

• Camera Data Properties → Lens → Focal Length changes field of view.
• Camera Data Properties → Depth of Field → Aperture → F-Stop changes depth-of-field blur.
• Render Properties → Motion Blur → Shutter controls blur in fractions of a frame, not photographic exposure time.
• Render Properties → Color Management → Exposure applies a stop-based gain as part of the view transform.
• In Cycles, Render Properties → Film → Exposure is a separate brightness scale that changes scene-linear render data before the view transform; its neutral value is 1.0.
• Light Data Properties → Exposure multiplies a Light’s intensity by 2^Exposure; +1 doubles it and −1 halves it.
• World Properties → Surface → Strength is a multiplier, not an EV field.

Consequently, typing f/5.6 into the camera makes the blur wider but does not brighten the image. Typing a motion-blur Shutter value does not dim daylight. For this exercise, leave Color Management Exposure at 0 and Cycles Film Exposure at its neutral scale of 1.0; represent exposure changes with recorded Light Exposure and World Strength offsets. This keeps EEVEE previews and Cycles checks on the same baseline.

For a still portrait, turn motion blur off until lighting is approved. If motion matters, convert real shutter time to Blender’s frame-based value:

Blender shutter in frames = shutter time in seconds × scene frame rate

At 25 fps, 1/250 s is 0.004 × 25 = 0.10 frame; 1/1000 s is 0.025 frame. This only approximates continuous-light motion blur. A short real flash pulse may freeze a moving subject more sharply than a continuous Blender Area or Spot light predicts.

3. Build one calibrated exposure reference

Create two light collections:

LIGHTS_FLASH
└── FLASH_A
    └── Flash_Key

LIGHTS_AMBIENT
├── Ambient_Window
└── Ambient_Practical

Put a gray-card plane at the subject. Give it a diffuse neutral material and, when entering linear RGB values, use approximately 0.18 for red, green, and blue. This is a consistency patch, not a certified meter.

Use the following order:

1. Set the virtual camera to 90 mm, f/8, and the correct focus target.
2. Under Render Properties → Color Management, set View Transform to AgX, Look to Medium High Contrast, Exposure to 0, and Gamma to 1. Keep this identical to Part 2. When using Cycles, also set Render Properties → Film → Exposure to its neutral scale of 1.0.
3. Set the flash color or temperature for the real unit; daylight electronic flash usually starts near the stock’s daylight balance, but use the flash manufacturer’s measurement when known.
4. In the Outliner’s View Layer display mode, exclude LIGHTS_AMBIENT and disable the World contribution used as room light.
5. Place the virtual flash at the real direction, height, modifier size, and 2.8 m distance.
6. Leave the flash object’s Exposure at 0 and adjust Power until the gray card, face, shadows, and specular highlights match the real f/8 reference. Record this Power as the flash baseline; call its current Exposure Flash offset 0 EV.
7. Exclude FLASH_A from the View Layer. Restore LIGHTS_AMBIENT, leave each ambient Light Exposure at 0, and adjust Light Power and World Strength to represent the real 1/125 s meter reading. Record those values as Ambient offset 0 EV.
8. Enable both collections and render a reference frame. Save a still and the .blend; never recalibrate this baseline while comparing later variants. Bypass any compositor film grade during measurement, because the fixed AgX view is the comparison transform. Apply the creative grade only after the lighting decision.

The robust real-world version is the reverse: arrange and meter the real flash at EI 80, make a gray-card and skin test, process or scan it consistently, then recreate its measured geometry in Blender. The calibration absorbs unknown differences between Blender watts and real flash output. Without that photographed reference, the result is still useful for relative planning, but its absolute film density and scan brightness are not predictions.

4. Translate guide number, distance, and manual power

For direct flash, use metric GN and distance in metres:

effective GN = GN at ISO 100 × √(target EI / 100)
             × √(manual power fraction) × 2^(-modifier loss in EV / 2)

working f-number ≈ effective GN ÷ flash-to-subject distance

At EI 80, full-power GN 25 becomes 25 × √0.8 ≈ 22.4. These full-power distances give useful starting apertures:

Reducing power by one stop multiplies GN by √1/2, not by one half, because guide number is proportional to the square root of light output:

The last two results fall outside the supported f/1.0–f/32 aperture range. They are still useful when a flash offers 1/128 or 1/256 power and you need to solve for a shorter distance, higher EI, less modifier loss, or a higher guide-number flash.

To solve for power rather than aperture:

required effective GN = target f-number × distance

required power fraction = (required effective GN
                         / full-power GN after EI and modifier correction)²

For EI 80, f/5.6, and 2.8 m without a modifier, the required GN is 5.6 × 2.8 = 15.68. The corrected full-power GN is 22.4, so (15.68 ÷ 22.4)² ≈ 0.49: choose 1/2 power. If the result is greater than 1, full power is insufficient; move the flash closer, open the lens, raise EI, remove loss, or use a higher-GN source.

After calibrating the virtual Light at real full power, add the table’s EV value to Light Data Properties → Exposure to represent a manual power reduction. If a softbox costs 1.3 stops and its geometry is not modeled, subtract another 1.3 from Light Exposure. If the virtual modifier already blocks and spreads light, do not subtract the loss again. Likewise, when you physically move the Blender light to a new measured distance, let Blender calculate that distance change; do not apply the GN distance change a second time through Exposure.

Guide number is unreliable for bounce and close, large modifiers because distance, coverage, and surface efficiency are ambiguous. Use it to reach a starting neighborhood; use a real flash meter to decide the film exposure.

5. Translate camera settings into Blender stops

Relative to the calibrated reference, calculate three values. Use shutter time in seconds, so 1/250 is entered as 0.004:

C = camera change
  = log₂(target EI / reference EI)
  + 2 × log₂(reference f-number / target f-number)

S = shutter change on ambient
  = log₂(target shutter time / reference shutter time)

P = real flash-output change
  = log₂(target manual-power fraction / reference manual-power fraction)

C affects flash and ambient. S affects only continuous ambient while the complete electronic flash pulse occurs during the synchronized opening. P describes a deliberate change on the real manual flash. This separation does not apply unchanged to HSS, flash bulbs, a pulse longer than the shutter opening, or a mistimed trigger.

The EI term represents the effective speed of the film-and-process combination used for the comparison. Merely changing the number on a handheld meter does not make film more sensitive and must not be added as Blender gain unless the resulting aperture, shutter, lighting, stock, or processing also changes.

Keep the Color Management and Cycles Film exposures at their baseline, then configure Blender:

If flash distance, modifier geometry, ambient fixtures, or the set itself changes, reproduce that change physically in Blender first. Add only the remaining numeric offsets; never count the same change in geometry and Exposure.

The examples below round conventional marked apertures and shutter speeds to whole stops. The logarithmic formulas preserve the small difference when exact values are required.

Example A: open the lens while holding exposure

Move from EI 80 · f/8 · 1/125 s · full flash to EI 80 · f/5.6 · 1/250 s · half-power flash:

C = +1 EV
S = -1 EV
P = -1 EV
Flash offset = C + P = 0 EV
Ambient offset = C + S = 0 EV

Set Camera F-Stop to 5.6; leave flash Exposure, ambient Exposure, and World Strength at their recorded baselines. At 25 fps, use Motion Blur Shutter 0.10 frame if motion blur is required. The wider aperture offsets half flash power, while the faster shutter offsets the extra ambient exposure. If the real flash instead stays at full power (P = 0), set the Blender flash to +1 EV: the flash-lit subject becomes one stop brighter while unchanged ambient stays at its reference brightness.

Example B: stop down at the 1/125 lower limit

Move from EI 80 · f/8 · 1/125 s to EI 80 · f/11 · 1/125 s:

C ≈ -1 EV
S = 0 EV
P = 0 EV
Flash offset ≈ -1 EV
Ambient offset ≈ -1 EV

Set Camera F-Stop to 11, the flash Light to approximately −1 EV, and unchanged ambient Lights to approximately −1 EV; multiply the baseline World Strength by about 0.5. Maintaining the original flash exposure would require about P = +1 EV: because the reference was already full power, use a more powerful source or move a direct flash from 2.8 m to about 2.0 m. Move it in Blender rather than also adding the distance gain to Exposure. Maintaining the original ambient exposure requires one more stop of real continuous light, because 1/125 s is the lower shutter boundary for this planning scenario; do not solve it with an unsupported slower shutter.

Example C: an extended f/1 and 1/1000 study

For a different synchronized camera, moving from f/8 at 1/125 s to f/1 at 1/1000 s at the same EI gives C = +6 EV and S = −3 EV. Reducing the real flash to 1/64 power gives P = −6 EV, so the flash offset is 0 EV; set Camera F-Stop to 1. Unchanged ambient still has a +3 EV offset, so set its Lights to +3 EV and multiply World Strength by 8. At 25 fps, Motion Blur Shutter is 0.025 frame.

This is valid stop arithmetic but impossible on a stock GW690III. The alternative camera, lens, trigger, and flash must genuinely synchronize at 1/1000 s. A short flash duration by itself does not make an incompatible shutter synchronize.

6. Estimate camera settings from a Blender variation

An arbitrary render has no unique ISO, aperture, shutter, or GN. Many combinations produce the same pixel value, and an uncalibrated Blender light supplies no absolute real-world meter reading. The reverse estimate works only relative to the saved reference and only after known light-output changes are separated from simulated camera exposure.

For an unchanged flash position and modifier, let F be its total Blender Exposure offset and let P be the real manual-power change already represented. Recover the camera change with:

C = F - P

If EI stays at 80, convert C to an aperture relative to f/8:

target f-number = 8 × 2^(-C / 2)

Thus +1 EV corresponds to about f/5.6, +2 EV to f/4, and −1 EV to f/11. If aperture stays at f/8 instead, sensitivity is:

target EI = 80 × 2^C

So +1 EV can be described as an equivalent EI 160 film-and-process response. Relabeling the same EI 80 film as EI 160 does not create that gain; without compensating processing it instead gives the film one stop less exposure.

For unchanged ambient geometry, let A be the total ambient Light offset. Separate its shutter component from the camera change:

S = A - C
target shutter time = reference shutter time × 2^S

If C = +1 EV but A = 0 EV, then S = −1 EV: 1/125 s becomes approximately 1/250 s. Round the result to an actual marked shutter speed, then recalculate the small difference when precision matters. If you moved the flash, changed a modifier, or changed the set, use the GN calculation and a meter instead of pretending that the Light Exposure field contains the whole answer.

Use this reverse worksheet for every approved render:

Reference: EI 80 · f/8 · 1/125 s
Known real flash-power change P: ______ EV
Total Flash Light offset F: ______ EV
Recovered camera change C = F - P: ______ EV
Total ambient/World offset A: ______ EV
Recovered shutter change S = A - C: ______ EV
Equivalent aperture at EI 80: f/______
Equivalent shutter after camera change: 1/______ s
Real flash GN, power, modifier, and distance: ____________________
Settings inside the physical camera/flash limits? yes / no

7. Meter and validate on the real set

The GW690III has no built-in exposure meter. Transfer the virtual plan in this order:

1. Set the handheld meter to the chosen working EI, here EI 80.
2. Place an incident/flash meter at the subject. Measure the key alone, then ambient alone.
3. Recreate the virtual flash direction, head-to-subject distance, modifier, and coverage.
4. Set the flash power suggested by the calibrated stop difference or GN table.
5. Meter again; change real flash power or distance until the meter returns the intended aperture.
6. Set shutter speed from the ambient decision while remaining inside the actual sync range.
7. Photograph a gray card and, when practical, bracket the first controlled test before committing an important eight-frame roll.

Back in Blender, temporarily select Render Properties → Color Management → View Transform → False Color to locate extreme highlights and deep shadows. Then restore AgX and Medium High Contrast before judging the intended picture. False Color is a diagnostic display, not proof that a negative, transparency, lab scan, or print will hold the same tonal range.

Configuration checkpoint

Before proceeding to backgrounds or materials, confirm:

• Scene scale, the selected format’s Sensor Width and output ratio, and the intended physical lens are locked. For the GW690III reference these are 82.6 mm, 82.6:56, and 90 mm.
• Camera f-stop matches the planned capture and is used only for depth of field.
• Flash and ambient occupy different collections and have recorded 0 EV reference values.
• Color Management Exposure, Cycles Film Exposure, and Gamma remain at their calibration values; camera equivalence is recorded on Light Exposure and World Strength.
• GN is metric at ISO 100 and has been corrected for EI, manual power, and unmodeled loss.
• Every proposed camera value is marked either GW690III-valid or extended experiment.
• A handheld meter, gray card, and real flash test remain part of the final plan.

Continue with Part 4: Build and Load the Background, or return to Part 2: Iterate Lighting in Real Time when the stop changes are correct but the preview is too slow.

Sources

• Fujifilm GW690III / GSW690III / GW670III owner’s manual
• Blender 5.2 Manual: Cameras
• Blender 5.2 Manual: Light Objects
• Blender 5.2 Manual: Motion Blur
• Blender 5.2 Manual: Cycles Film exposure
• Blender 5.2 Manual: Color Management displays and views
• Blender 5.2 Manual: Color Management
• Kodak Ektar 100 technical data: 135 and 120 negative-size examples
• Ilford processing data: nominal 4×5 and 8×10 sheet-film sizes
• Nikon SB-500 manual: guide number, aperture, distance, and ISO factors
• Sekonic: incident flash-meter placement