ISO Invariance: Which Cameras Benefit Most from Pushing Exposure

The question ISO invariance answers is specific and practical: if you shoot at ISO 100 and then brighten your image 4 stops in post-processing, do you get the same result as shooting at ISO 1600 in the first place?

For some cameras, the answer is yes. For others, the in-camera ISO setting genuinely reduces noise — brightening a dark raw file in post reveals more noise than exposing at the equivalent higher ISO would have. The gap between these two outcomes determines whether base-ISO shooting with shadow recovery is a viable strategy for your camera.

This gap matters most for astrophotography, high-contrast landscapes, and mixed indoor/outdoor portraits — situations where the subject brightness range exceeds what a single correct exposure can hold. Understanding where your camera lands on the ISO invariance spectrum determines whether you should expose for the highlights and recover shadows, or whether you should expose the shadows correctly and risk clipping highlights.

---

What ISO invariance means in plain language

Most camera specs don't mention ISO invariance — but it shapes how you should expose in the field.

The practical version: a fully ISO invariant camera lets you shoot everything at base ISO (ISO 100), underexpose if needed to protect highlights, and then brighten the raw file in Lightroom or Capture One without penalty. The result is indistinguishable from having shot at the equivalent higher ISO. A non-ISO-invariant camera punishes that approach — brightening a base-ISO file reveals banding, color shifts, and elevated noise in shadows that would not have appeared if you had raised ISO in camera.

In practice: ISO invariant cameras reward ETTR (expose to the right) workflows in astrophotography and landscape. Non-invariant cameras reward getting ISO right at capture. Neither is wrong — they are different optimal strategies for different sensors.

---

The technical problem

When a camera records an image, two noise sources compete for dominance:

Photon shot noise: noise from the random arrival of photons at the sensor. This scales with the square root of the photon count — more light = more noise in absolute terms but better signal-to-noise ratio. You cannot reduce photon shot noise in post; you can only reduce it by collecting more photons (longer exposure, wider aperture, larger sensor pixels).

Read noise: electronic noise introduced during the analog-to-digital conversion of the sensor signal. This is fixed regardless of exposure — it is present whether you collect many photons or few. At low ISO, read noise can constitute a significant fraction of the total signal from shadow areas. Raising ISO — amplifying the signal before digitization — buries read noise deeper under the amplified signal, reducing its relative contribution.

ISO invariance describes the relationship between these two sources. A camera is ISO invariant when its read noise is already so low relative to the amplified signal that applying additional gain (raising ISO) adds no meaningful advantage. In that state, the photon shot noise dominates across all ISO settings, and brightening in post is mathematically equivalent to shooting at a higher ISO.

A camera is not ISO invariant when raising ISO in-camera produces measurably cleaner shadows than raising ISO in post — because the in-camera gain reduces the relative contribution of read noise before digitization.

The boundary between these states is not binary. Most cameras have an ISO level above which they become effectively ISO invariant. Below that threshold, shooting at a higher ISO and underexposing slightly produces cleaner shadows than shooting at base ISO and brightening in post. Finding that threshold for your camera is the practical outcome of this analysis.

---

The formula and why it matters

ISO invariance can be tested with a simple procedure:

Test A: Shoot a scene at ISO 100, 1/100 s, f/5.6. Bright the RAW file by 4 stops (+4 EV) in your raw processor.

Test B: Shoot the same scene at ISO 1600, 1/100 s, f/5.6 (same shutter speed and aperture — 4 stops underexposed). Examine the file at its native exposure.

If the shadow noise in both files is visually identical, your camera is ISO invariant between ISO 100 and ISO 1600. If Test B shows noticeably less shadow noise than the brightened Test A file, your camera benefits from raising ISO in-camera.

The Photons to Photos Shadow Improvement chart (photonstophotos.net/Charts/PDR_Shadow.htm) quantifies this relationship for hundreds of cameras by plotting the read noise reduction gained from ISO amplification across the ISO range. A flat line on this chart means ISO invariant. A rising slope means the camera benefits from in-camera ISO amplification at that setting.

Shadow Improvement (stops) = log₂(read noise @ ISO n / read noise @ base ISO)

If Shadow Improvement → 0 as ISO increases: camera approaching invariance
If Shadow Improvement stays high: camera benefits from in-camera ISO raise

The inflection point — where the slope of the Shadow Improvement chart levels off — is the effective ISO invariance threshold.

---

How we measured

We did not conduct original shadow improvement measurements. The data in this article derives from Photons to Photos Shadow Improvement charts (for cameras that appear in their database), published ISO invariance analyses from Capture the Atlas, Improve Photography, and Lonely Speck's astrophotography ISO guides, and general review consensus from DPReview and Space.com's astrophotography camera comparison.

Specific ISO invariance threshold values are estimated from these sources. Most are DATO_SIN_FUENTE_PENDING_VERIFICACION and require direct verification against Photons to Photos charts before being published as factual claims. The relative ordering (Sony > Nikon > Fujifilm > Canon > Panasonic for ISO invariance quality) is consensus across multiple independent sources; the specific threshold ISOs are estimates.

---

Master table

Cameras ranked by estimated ISO invariance threshold — the lowest ISO at which the camera becomes effectively ISO invariant. Lower threshold = better (the camera is ISO invariant over a wider range). Confidence: Confirmed = verified from published analysis; Estimated = derived from review consensus; PENDING = requires direct verification.

Brand	Model	Format	Sensor type	Pixel pitch (µm)	Est. threshold ISO	ISO invariance quality	Source
Sony	a7 V	Full-frame	Partial-stacked	5.10	~ISO 160	Excellent	Estimated
Sony	a7R V	Full-frame	BSI	3.76	~ISO 160	Excellent	Estimated
Sony	A7 IV	Full-frame	BSI	5.10	~ISO 400	Excellent	Confirmed (capturetheatlas.com)
Sony	a6700	APS-C	BSI	3.76	~ISO 400	Excellent	Confirmed (multiple sources)
Sony	ZV-E10 II	APS-C	BSI	3.76	~ISO 400	Excellent	Estimated (same sensor as a6700)
Sony	A1 II	Full-frame	Stacked BSI	4.16	~ISO 400	Excellent	Estimated
Nikon	Z8	Full-frame	Stacked BSI	4.35	~ISO 800	Very good	Confirmed (improvephotography.com)
Nikon	Z9	Full-frame	Stacked BSI	4.35	~ISO 800	Very good	Confirmed
Nikon	Z6 III	Full-frame	Partial-stacked	5.92	~ISO 800	Very good	Estimated
Nikon	Z5 II	Full-frame	BSI	5.92	~ISO 800	Very good	Estimated
Nikon	Z50 II	APS-C	BSI	4.20	~ISO 800	Very good	Estimated (EXPEED 7 architecture)
Fujifilm	X-T5	APS-C	X-Trans BSI	3.04	~ISO 640	Good	Confirmed (dual conversion gain)
Fujifilm	X-H2	APS-C	X-Trans BSI	3.04	~ISO 640	Good	Confirmed
Fujifilm	X-S20	APS-C	BSI	3.76	~ISO 640	Good	Estimated
Fujifilm	X-H2S	APS-C	Stacked BSI	3.76	~ISO 1250	Moderate-good	Estimated (stacked architecture)
Canon	EOS R5 II	Full-frame	BSI	4.39	~ISO 1600	Moderate	Estimated (improved over R5)
Canon	EOS R6 II	Full-frame	BSI	5.98	~ISO 1600	Moderate	Estimated
Canon	EOS R7	APS-C	BSI	3.20	~ISO 1600	Moderate	Estimated
Canon	EOS R1	Full-frame	Stacked BSI	6.00	~ISO 3200+	Limited	Estimated
Sony	A9 III	Full-frame	Global shutter	5.94	Not comparable	Special case	See note

Sony A9 III: the global shutter architecture changes the read noise model fundamentally. ISO invariance analysis as described here does not apply cleanly to global shutter sensors. Treat as a special case; consult camera-specific analyses before applying this framework to A9 III shooting decisions.

---

Reading the table

Sony's early dual-gain architecture is the current benchmark. Sony's BSI CMOS sensors in the A7 series have used dual-gain readout circuits since approximately the A7 II generation (2014–2015). The second gain circuit activates at or before ISO 400 on most Sony bodies — meaning at ISO 400 and above, these cameras are essentially ISO invariant. Below ISO 400, there is a small measurable benefit to raising ISO in-camera rather than brightening in post; above it, the two approaches produce the same result. This is why astrophotographers who use Sony bodies commonly shoot at ISO 400 or ISO 800 regardless of sky conditions: the higher ISO improves sky SNR while the strong ISO invariance above threshold ensures no shadow penalty in the foreground.

Nikon's Z-series trades a slightly higher threshold for extremely low read noise above it. Nikon's Z sensors become effectively ISO invariant around ISO 800, slightly later than Sony's threshold. The trade-off is that once above the threshold, Nikon's shadow recovery from underexposed files is extremely clean — some reviewers describe it as the cleanest among all mirrorless systems. For astrophotographers who track the sky and can control exposure time, the difference between ISO 400 (Sony fully invariant) and ISO 800 (Nikon fully invariant) rarely changes the practical outcome.

Fujifilm's X-Trans sensors introduce dual conversion gain at the film simulation layer. Fujifilm's X-T3, X-T4, X-T5, and X-H2 implement dual conversion gain at approximately ISO 640. Below that threshold, in-camera ISO amplification reduces noise more than post-brightening. Above it, they behave similarly to Sony in ISO invariance terms. One important nuance: Fujifilm's film simulations modify the JPEG output based on the ISO setting applied in-camera — the film simulation behavior is tuned to the ISO-amplified signal. If you shoot Fujifilm JPEGs rather than raw, ISO invariance is less directly relevant; the camera's color and tonal rendering are applied before delivery.

Canon has improved significantly from its historical baseline but has not closed the gap. Older Canon sensors — DSLR-era bodies from the 5D Mark III era and before — were among the least ISO-invariant sensors in mainstream photography. Brightening a base-ISO Canon RAW file in post revealed visibly more noise than shooting at the equivalent higher ISO. Canon's R5, R6, R7, and R5 II represent a genuine architectural improvement. The threshold ISO for current Canon R-series bodies is estimated at ~1600 — considerably higher than Sony or Nikon, but reaching ISO invariance eventually. For photographers who take exposures well above ISO 1600, the gap becomes irrelevant. For those who want to shoot at base ISO and recover shadows heavily, Canon bodies require more careful exposure discipline.

---

Edge cases

Most ISO-invariant: Sony a7 V, a7R V, a6700, A7 IV (threshold ~ISO 160–400). These cameras allow aggressive shadow recovery from base-ISO RAW files with results comparable to shooting at ISO 800. For Milky Way foreground exposures, where the foreground is metered for shadow detail and the sky is exposed separately, this enables strong composite work from base-ISO capture.

Least ISO-invariant in this table: Canon EOS R1 (estimated threshold ISO 3200+). The R1's stacked sensor, optimized for 40 fps burst at $6,299, trades read noise performance at base ISO for readout speed. Shooting the R1 at ISO 100 and brightening 4 stops in post will produce noticeably noisier shadows than shooting at ISO 1600. For the R1's intended use case — sports and action in controlled light — this trade-off is typically invisible. For any use case where base-ISO shadow recovery matters, the R1 is not the right tool.

Global shutter exception: Sony A9 III. The global shutter in the A9 III changes how read noise is distributed across the image. The standard ISO invariance model — where raising ISO in-camera reduces read noise from the digitization stage — does not apply cleanly to global shutter architectures because the readout structure is fundamentally different. The A9 III's shadow behavior at various ISO settings requires a camera-specific analysis from Photons to Photos' Shadow Improvement chart rather than applying the general framework described here.

Fujifilm X-H2S stacked APS-C. The X-H2S's stacked sensor enables 40 fps electronic burst, but like the Canon R1 and Sony A9 III, speed comes with a read noise penalty at base ISO. The X-H2S has an estimated threshold ISO of ~1250 — higher than the X-T5 (~640) despite sharing a similar 3.76 µm pixel pitch. For X-H2S users shooting astrophotography or high-contrast landscapes, shooting at ISO 1250 or above rather than base ISO is a more reliable strategy than heavy shadow recovery from ISO 160.

---

Which camera would we buy for ISO invariance today

If ISO invariance across the full shooting range is the primary criterion — for landscape, astrophotography, or any work that relies on base-ISO shadow recovery — the Sony lineup provides the earliest threshold and strongest invariance: a7 V, a7R V, a6700, A7 IV, all crossing the invariance threshold at or before ISO 400. The a6700 ($1,399) is the best-value entry point in the Sony ecosystem for this specific characteristic.

For APS-C astrophotography specifically: the Sony a6700 and ZV-E10 II share the same sensor and the same ISO invariance characteristics. The 3.76 µm pixel pitch means Milky Way exposures are NPF-limited at shorter intervals than the Nikon Z50 II's 4.20 µm pixels — but the a6700's ISO invariance and IBIS offset this for many shooting scenarios. If you need longer untracked exposures, the Nikon Z50 II's larger pixels are worth the slightly higher ISO invariance threshold.

If you shoot Nikon and accept a threshold of ISO 800: the Z8 ($3,999) and Z5 II (~$1,399) both deliver exceptional ISO-invariant performance above that threshold. For tracked astrophotography where the sky exposure is always above ISO 800, the Nikon Z-system's read noise above threshold is among the lowest available.

If you shoot Canon: accept that shadow recovery from base ISO will cost you relative to Sony and Nikon. Expose the R7, R5 II, or R6 II for adequate shadow brightness in-camera rather than relying on post-processing recovery from the noise floor. This means slightly higher ISOs for shadow areas — but the camera's performance above ISO 1600 is strong and the resulting files are clean.

If you shoot Fujifilm raw: the X-T5 and X-H2 cross the ISO invariance threshold at ISO 640. For landscape work and slow astrophotography, shooting at ISO 640 base and recovering underexposed shadows in post works reliably. The X-Trans color filter array introduces complexity in raw processing (some demosaic algorithms produce slightly different results), but the ISO invariance itself operates the same as on any other BSI sensor above the threshold.

---

Affiliate block

Cameras we would buy if ISO invariance for astrophotography or landscape work is the primary criterion:

• Sony a6700 (APS-C, 3.76 µm, threshold ~ISO 400, $1,399) Buy on Amazon →
• Sony a7 V (FF, 5.10 µm, threshold ~ISO 160, $2,899) Buy on Amazon →
• Nikon Z50 II (APS-C, 4.20 µm, threshold ~ISO 800, $906) Buy on Amazon →
• Nikon Z8 (FF, 4.35 µm, threshold ~ISO 800, best above threshold, $3,999) Buy on Amazon →
• Fujifilm X-T5 (APS-C, 3.04 µm, threshold ~ISO 640, $1,699) Buy on Amazon →

---

Sources

Photons to Photos Shadow Improvement chart: photonstophotos.net/Charts/PDR_Shadow.htm — the primary reference for per-camera ISO invariance data. Direct verification of all threshold ISOs in this article should use this chart.

ISO invariant camera lists and definitions: Capture the Atlas, "What is ISO Invariance? Find out if your Camera is ISO-less" — general lists of ISO-invariant cameras by brand. Improve Photography, "ISO Invariance: What it is, and which cameras are ISO-less" — threshold ISO estimates by brand. Photography Life, "ISO Invariance Explained" — technical framework used in this article.

Astrophotography ISO optimization: Lonely Speck, "How to Find the Best ISO for Astrophotography: Dynamic Range and Noise" — the most detailed astrophotography-specific treatment of ISO invariance in practical use.

Sony a6700 ISO invariance: Brian Smith Photography, referencing Photons to Photos test data, confirming Sony a6700's top APS-C performance. Consistent with general Sony BSI CMOS architecture characterization.

Fujifilm dual conversion gain threshold (~ISO 640): multiple Fujifilm-focused review sources confirm the dual conversion gain activates near ISO 640 on X-T3/X-T4/X-T5/X-H2 generation cameras.

Canon ISO invariance improvement with RF generation: DPReview review notes for Canon EOS R5, R6, R7, R5 II — all note improved base-ISO read noise performance versus DSLR predecessors. Specific threshold ISOs remain estimated.

---

Related reading

• Best APS-C Cameras in 2026 — Sony a6700 and Nikon Z50 II in broader APS-C context
• Best Wildlife and Sports Cameras in 2026 — how ISO invariance threshold affects choice between A9 III and a7 V for mixed action and low-light wildlife
• Best Portrait Cameras in 2026 — ISO invariance and natural light portrait shadow recovery latitude
• Dynamic Range in Cameras: What the Numbers Actually Mean — PDR explained; how base-ISO DR and ISO invariance interact
• Best Full-Frame Cameras in 2026 — full-frame coverage across use cases

---

Astrian Light is the photography vertical of Astrian, powered by NASA JPL DE441 astronomical data. We write technical, no-bullshit guides for photographers who plan their shots.