Calling owners of both X and GF systems - tonal range

UPD: added a few words on WB, missed it in the 1st edition of this comment.

pictureAngst wrote:

Hi X and GF owners, I have a question about tonal range.

I remember when digital was becoming mainstream, medium format digital cameras had lower MP counts than their DSLR cousins, but had noticeably greater tonal range because of their larger pixel size.

I think you are probably confusing the dynamic range of the sensor with the tonal range of the resulting image.

To my opinion, the technology of silicone sensors is so mature already, that there wouldn't be any significant difference in dynamic range between any two sensors you take. Maybe the new top-class cinematic ARRI sensors stand out, according to rumors and tests. You can google it yourself, there are plenty of solid scientific studies of what the real DR of different sensors is today.

Just for example, here isJim Kasson's blog about Engineering Dynamic Range of GFX 50S he measured, his result is approx. 13.7 EV (read: stops) at ISO 100.

The Nikon Z9 data from the same author (single source of data is good for comparisons thanks to identical methodology and procedure used) are also available, Z9 shows somwhat lower EDR values, maybe due to smaller size of each individual photosite on its sensor?

Generally, measurement of sensor characteristics is a pretty complex topic, and there is a difference between Engineering DR and Photographic DR , and so on, but I think it's safe to assume that any sufficiently modern camera sensor provides at least 13 EV of DR, often 14, some as much as (almost) 15 EV of dynamic range, and the value of luminosity, recorded by each individual photosite, is encoded into a 16-bit binary integer value than. 16 bit is enough precision for that.

But back to the tonal range. Let's start with a simple fact: RAW acquired from the sensor, is not an image at all. And if no "image" exists until the RAW is developed in software (either in-camera or on the external computing device), then no notion of any "tone" does exist either. Conclusion: when we are speaking of "tonal range", we are not in fact comparing sensors and/or RAWs - we are comparing the abilities and settings of different software RAW processors.

When I say ‘tonal range’ I’m referring to the resolving of subtle changes in colour and brightness, for example in the graduated shades of a sky, feather or flower. In my experience the smaller the sensor the more ‘crunched’ these graduations become.

Now that the X-H2 has a similar pixel count to the GFX50, potential buyers like myself will be weighing up the pros and cons of each system - I’m thinking specifically of portraiture and landscape applications.

So my question is, is this difference in tonal range between the systems still noticeable, particularly in JPEGs?

Yes, the in-camera engines that produce images from RAWs are definitely different - they use different demosaicing algorithms, first of all. They apply different input profiles to demosaiced data. And they apply different tone curves than, and/or LUT(s) btw you can tune these to some point using camera settings for film simulation, color, shadows and highlights etc. (some cameras have more various OOC JPEG settings, some have less).

What exactly am i leading to? To my opinion, if you are using an external computer equipped with a RAW development software, with two RAWs from different cameras, taken at identical conditions at a single scene and exposed to the same level, then - after application of sensor-specific demosaicing algorithm at its best settings and other sensor-specific things - input profile, WB and exposure corrections - you may apply an identical further workflow to both RAWs (i.e. tone curve and LUT) and achieve visual results, which are close to identical.

This implies (indistinguishably) close tonal range characteristics.

(I appreciate no one has an X-H2 yet, but I would expect a comparison between say an X-T3 or 4 and a GFX50 would produce similar conclusions.)

This is a different target, I think. If you want to acquire data which is comparable, you need to apply the procedure analogous to the one described here, but adapted to the case of two different camera bodies:

• one single scene,
• identical lighting,
• to compare tones (and tonal ranges) WB consistency between two bodies is critical; the best way is to set WB on both bodies "by example", with the same "gray card" target before taking any shot,
• you may include some kind of colorchecker target into your frame, i.e. the one offered by DataColor, this will allow some minimal quantitative comparison of tone reproduction,
• identical AOV of the lens (ideally - one and the same lens for both bodies),
• identical exposure settings (aperture and SS and ISO), with exposure corrected by shutter speed adjustment until both bodies will show the same under- or overexposure level (or well-exposed image with 0 EV EC),

than you acquire RAW + OOC JPEG FINE picture from each of your two bodies, and you may compare OOC JPEGs side-by-side than, either in a qualitative or even in a quantitative way.

Also given both bodies are Fujis, you may like to set the in-camera RAW developer and OOC JPEG generating engine to as close settings to each other, as possible - I mean, the same film simulation, shadows/highlight, color, minimal possible values for sharpening and NR, etc.

Otherwise any meaningful comparison is probably impossible. How do you compare tonal range of two images, taken at different scenes, at different lighting, and with different settings of different camera bodies?

many thanks for any replies.

Best, Steve.