Is the X-H2S “dynamic range improvement” a myth?

[PoorAlchemist]

---

Is the X-H2S “dynamic range improvement” a myth?

Since the release of the Fujifilm X-H2S, Fuji has heavily promoted:

• the sensor’s internal 14-bit readout,

• the new F-Log2 profile,

• and a supposedly tangible improvement in video dynamic range compared to previous generations (X-T4, X-S20, etc.).

But when you look at actual laboratory measurements, the narrative starts to fall apart.

---

What the numbers actually show (measured data, not marketing)

Based on IMATEST / SNR-style measurements:

• X-H2S

  • ≈ 12.2 stops at SNR=2

  • ≈ 13.6 stops at SNR=1

  • Measured in ProRes HQ, high native ISO (1250)

• X-T4

  • ≈ 11.8 stops at SNR=2

  • ≈ 13.4 stops at SNR=1

  • Measured in H.264 / H.265, lower native ISO (800)

The real-world difference is about 0.2 to 0.4 stop, depending on the threshold used.
This is nowhere near a generational leap.

---

The core question: where did the 14-bit promise go?

If the X-H2S sensor is truly read internally at 14-bit, a simple question arises:

Why does this extra bit depth not translate into a measurable increase in usable dynamic range?

Because:

• the final recorded signal is still 10-bit,

• read noise appears to cap the signal before those extra bits can matter,

• SNR curves remain very close to those of the X-T4.

In short:
14-bit upstream, same ceiling downstream.

---

And what about F-Log2?

F-Log2 is supposed to:

• extend highlight latitude,

• better exploit the sensor signal.

Yet in practice:

• measured dynamic range barely increases,

• what we mostly see is curve redistribution, not actual expansion,

• shadow noise rises earlier.

This raises a legitimate concern:
Are we just looking at a remapping of the same dynamic range, rather than a true physical gain?

---

Provisional conclusion (but an uncomfortable one)

Based on the available data:

• the X-H2S “dynamic range improvement” appears largely overstated,

• the 14-bit readout looks more like a theoretical talking point than a measurable benefit,

• F-Log2 seems primarily like a grading convenience, not a sensor-level breakthrough.

---

Open but serious question

Is the internal 14-bit sensor readout and F-Log2, in practice,
a damp squib with no truly palpable impact on real-world video dynamic range?

If anyone has:

• independent measurements showing otherwise,

• or a demonstrable gain beyond ~0.3 stop,

I’m genuinely interested.
But for now, the numbers simply do not support the narrative.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[jerryy]

It seems you are dragging up old stuff that has been known for some time. It is stated in the specs the video is recorded in 10 bits:

X-H2S @ 4:2:2 10-bit internal recording

X-T4 @ 4:2:0 10-bit internal recording

The 4:2:2 gives better color space results.

The big video advantage for the X-H2S has been faster readout, being able to use external hdmi capture devices and 6K video capture.

There is a fairly good video that has been around for a while discussing F-Log2 vs F-log:

https://www.youtube.com/watch?v=8kDlhiEOAus

p.s. Welcome to the forum.

 

Edited December 31, 2025 by jerryy