Camera Simulation vs Filters: What's the Difference?
A filter changes how a photo looks. A camera simulation also changes how it's taken. Here's what that distinction actually means — and where each is the right tool.
A working vocabulary
A quick vocabulary before the argument, because most of the confusion in this category is a vocabulary problem.
- Filter. A visual treatment applied to a finished image. Same input pixels produce same output pixels. Every VSCO preset is a filter.
- Preset. A saved bundle of edit values — exposure, contrast, curves, split tone. A preset is a filter stored as a recipe.
- LUT. A colour lookup table. A precise mathematical mapping of every input colour to an output colour. Powerful, still a filter.
- Film simulation. A look modelled on a specific film stock's colour response. A good one respects real film science; a bad one is a saturation slider with a Kodak name on it.
- Camera simulation. A model of an entire capture system: sensor behaviour, motion artifacts, flash, exposure metering, shot count, framing constraints, viewfinder chrome, and workflow. A camera simulation is a filter plus everything that happens before the shutter closes.
Six things a filter cannot do
The honest test is upstream of the shutter. Every one of these ships in Afterhours today, enforced by code, not marketing.
1. Shot count
SNAP-27 gives you 27 photos, then the roll is done. Not "27 photos before a filter unlocks" — 27 photos, then you develop. The number lives in the camera's spec (rollCapacity: 27), and imports from your photo library are explicitly refused. The internal docs are blunt about it: "an imported photo must not bypass the develop-into-a-roll film fiction."
A filter can't say no. It runs on whatever you point at it.
2. Delayed development
POLAROID-90 makes you wait 20 seconds for the develop. You capture instantly, get a blank white square, and a chemical reveal animates from a seed frozen at capture. Same seed, same develop, byte-identical on replay.
Shake to 3× speed — you cannot skip. A filter cannot make you wait, because a filter doesn't know what "wait" is.
3. Slow-sync flash fusion
G8X MIII fires an ambient frame ~200 ms before the flash burst, then fuses subject-from-flash over background-from-ambient at develop time. This is the sunset-behind-subject shot every pocket compact used to do.
Slow-sync is a capture-time technique. You need two exposures, at two times, taken by the camera. A filter arrives at a finished frame with the ambient one already lost.
4. Bimodal capture grade
CZ-30 uses one colour table when the flash fires and a different one when it doesn't. Warm-and-punchy for ambient, cool and pale for flash. The live preview always shows the ambient grade, because the era's LCDs did too. The internal comment is blunt: "the flash look only exists in the captured photo, exactly like the era's LCDs."
A filter doesn't know whether the flash fired. It doesn't know there was a flash. It just gets pixels.
5. Signal-chain modelling
VHS-83 is a direct port of ntsc-rs — real composite-video math, not overlaid noise. RGB is remapped to FCC YIQ. The chroma channel is bandlimited. Luma is smeared with a specific filter. QAM demod produces cross-colour rainbow crawl on fine detail. A head-switch tear draws itself at the bottom of every frame.
When VHS-83 renders, the digital-tape stage is disabled. The doc puts a rule on it: "the two systems must never stack." A VHS filter that also renders MiniDV macroblocks isn't a VHS filter. It's an overlay pack.
6. Format-native chrome
CRT-9 draws the tube, not just the tint. Rounded-corner bezel around the frame, curvature, phosphor mesh, bloom on the glass, boot-static roll before the first frame, signal-decay tail on every video export.
The chrome renders at the same 480×640 grid as the picture, then soft-blurs so it inherits the pixel character of the format. The wordmark and REC dot and 14-segment HUD all belong on that tube because they went through the same signal. A filter puts branding on top of your photo. A camera simulation makes the branding part of it.
Where a filter is the right tool
Filters are the right tool when you already have the shot and just want a look. Fast, non-destructive, five presets in ten seconds, strip it off next year and get your original back. Instagram, VSCO, Lightroom mobile — nothing in this guide is a case against them.
Camera simulations are the right tool when the taking of the photo is part of the point. When you want to shoot 27 frames and stop. When you want to wait for the reveal. When you want a fluorescent restaurant to metabolise the way a 2004 digicam actually metabolised fluorescent restaurants. When the point of the picture is that it was taken on the thing it was taken on.
If the taking doesn't matter to you, use a filter. It's built for that.
How Afterhours draws the line
Afterhours has two shelves that behave slightly differently.
The AFTERHOURS shelf — video-artifact cameras and disposables like NS-8, SNAP-27, CZ-30, HDV-2000, VHS-83, G8X MIII, DV-90, CRT-9, POLAROID-90 — are camera simulations in the strict sense. Each has its own format, aspect ratio, flash behaviour, chrome, capture flow. Some have shot counts. Some have reveal animations. Some are photo-only or video-only.
The FILM shelf — MONO 100, MORNING 01, STILL 02, FADE-200, VERDE 50, NOCT-800, HIBI-200, DAILY 01 — sits between. Each is technically a camera (per-body format, controls, flash style, live shader baked at capture), but the device grammar is intentionally sparser. Film-forward shooting is about the material more than the body, so the cameras get out of the way.
Both shelves run through the same live-in-viewfinder pipeline. Both bake at the shutter. The difference is how much camera the design puts in front of you. If a shelf ever ships that's a filter, Afterhours will call it a filter.
What this doesn't mean
A camera simulation is not a hardware clone. NS-8 is not a Sony Hi8 Handycam. HDV-2000 is not a Canon VIXIA. VHS-83 is not a JVC GR-45.
Real Hi8 tape hisses in ways no shader has faithfully reproduced. Real MiniDV macroblocks stochastically in ways an on-device model can only approximate. Real VHS heads sputter differently on the third pass of the same cassette than on the first.
Afterhours simulates a reading of each camera — the specific behaviours that make it feel like that camera when you use it. The chroma bleed of DV. The 27-shot discipline of a disposable. The wait of a Polaroid. The head-switch tear of a VHS-C at the bottom of every frame.
If you want the object, buy the object. That's the honest boundary of any simulation.
Sources: internal engineering docs and camera specifications for Afterhours Camera, verified at the reviewed date above. Corrections welcome.
Frequently asked questions
Is film simulation the same as a filter?
Sometimes. A film simulation applied to a finished photo — the kind you pick from a menu in Lightroom — is a filter. A well-authored one, but a filter.
A film simulation running live in a camera app's viewfinder, driving exposure metering, freezing a grain seed at capture time, and baking into the file the moment the shutter closes, is a camera simulation. The term covers both. What matters is where in the pipeline it lives.
Can a LUT recreate VHS?
No. A LUT can recreate the colour of VHS — the YIQ chroma bleed, the pale warmth, the way saturated reds pool. It cannot recreate the head-switch tear, which is a spatial artifact tied to frame boundaries. It cannot recreate dot crawl, which is a demodulation artifact. It cannot recreate the tracking errors that only exist because a rotating drum reads a moving tape. Those live in the signal chain, not the colour space.
Does camera simulation make an iPhone identical to the original hardware?
No. Pretending otherwise is where most vintage camera apps lose credibility — they promise identity, can't deliver, then compensate by turning up the filter until every photo looks the same.
A camera simulation should promise a reading: the specific behaviours that make a camera feel like itself. Anything more is marketing.
Why does any of this matter if the final photo looks the same?
The final photo usually doesn't look the same. But that's not the point.
It matters because how you take a photo shapes what you take. You shoot differently when you have 27 frames than when you have infinite. You wait differently when a reveal is coming. You point the camera differently when the chrome is part of the picture. If a look is the only thing you want, a filter is enough. If the shooting is part of what you want, a filter isn't.
Does Afterhours ever use filters?
Afterhours does not apply post-capture filters to imported photos. The look is baked at capture from the camera you're using. If a shelf ever ships that's a filter — a treatment you apply to a photo you've already taken — Afterhours will call it a filter, not a camera.
Published by Afterhours Camera. Written by Peter Garcia. Corrections and questions: bricksandcanvas@gmail.com.