Fuji/Panasonic sensor 26.2 stops organic sensor

[sanking]

http://petapixel.com/2013/06/11/fuji-and-panasonics-new-organic-sensor-boasts-insane-14-6-stops-dynamic-range/

Would love to have that kind of range in my current digital work flow!

Sandy

[Lachlan 717]

Whilst I have no idea about what it all means, Thom Hogan set out to debunk this article's findings:


"The Reason You Read This Site and Not the Others
June 11, 2013 (commentary)--More than one Web site is now running information that claims that Fujifilm/Panasonic's just previewed organic sensor is twice as good as the D800 sensor. The basis of that claim is to take measured DxO numbers that are post shot noise, post dark current noise, and post read noise and compare them with numbers that are theoretical engineering dynamic range pre shot.

The theoretical number for the new organic sensor is 88dB. Sounds great, doesn't it? Almost 30 stops (each 3 dB represents a doubling, or what we call a stop in photography). I guess those other sites didn't look to see what other sensor companies report their theoretical dB numbers as. I'll give you one for comparison: the Aptina AR1011HS sensor, which is a 10mp 1" sensor that's mostly the same as the sensor used in the original Nikon 1 cameras (missing the AF component): 84dB. Don't believe me? Click here (http://www.aptina.com/products/image_sensors/ar1011hs/), look in the right column. I used Aptina for a couple of reasons. First, they're one of the few sensor makers whose stated specifications are available to anyone on the Web (Sony keeps them mostly private these days). Second, I wanted a sensor that was in a camera that everyone could agree doesn't perform as well as the one in a Nikon D800.

So don't go putting your D800's on eBay because someone threw a faux stone at the ego you've got vested in your sensor. Your D800 still outperforms almost everything else on the market."

As I wrote, I have no idea what this all means, so I'd be interested to read what his contention means to your link, Sandy.

[sanking]

As I wrote, I have no idea what this all means, so I'd be interested to read what his contention means to your link, Sandy.

The Fuji/Panasonic sensor being discussed will not be in any camera for at least one or two more generations, the Nikon D800 is now over a year old, so I don't understand the reason to compare what is with what may be?

How this plays out is anyone's guess, but I believe it indicates that there is research and development going on that is constantly upping the game in terms of digital sensor capabilities.

In any event I am definitely holding on to my Nikon D800. But I am also holding on to my Fuji GSW690III and my 5X7 outfit because B&W still rocks with medium format and larger negatives!

Sandy

[Andy Eads]

Correct me if I'm wrong, but a sensor with that kind of dynamic range will need a huge number of bits to record the image in TIFF or RAW. This may be pointing us to some practical limits for how much dynamic range you really need. This would apply to how many bits you need to assure good tonal separation and how many your camera has to swallow every frame. Has anyone published numbers for the dynamic range that is really captured and used when folks apply HDR software? My brief search yielded claims as high as 14 stops. I can't remember shooting anything like that sort of range for landscape photos.

[asnapper]

All the major camera manufacturers have patents for none Bayer sensors, and Fuji have been developing the organic sensor for several years, and we could see a camera featuring one within 18 months

http://www.43rumors.com/ft4-new-sensor-to-be-ready-for-end-2014-early-2015-only/#comments

[Leigh]

A high-range sensor is useless unless the presentation medium can support it.
I don't know of any that can.

Transmissive media would have the best chance of displaying the full range of tonality, but an image bright enough
to distinguish between the darkest tones would be bright enough to burn your eyes out.

HDR processing just compresses that range down to what the media can support.

A true 1-stop variation in subject illumination is dumbed-down to 1/2 stop or some such, so you really
accomplish nothing that couldn't be done by a regular sensor

I'm sure we'll see interminable marketing hype about these long-range sensors when cameras hit the market.
Marketing hype is marketing hype.

- Leigh

[Brian Ellis]

Whilst I have no idea about what it all means, Thom Hogan set out to debunk this article's findings:


"The Reason You Read This Site and Not the Others
June 11, 2013 (commentary)--More than one Web site is now running information that claims that Fujifilm/Panasonic's just previewed organic sensor is twice as good as the D800 sensor. The basis of that claim is to take measured DxO numbers that are post shot noise, post dark current noise, and post read noise and compare them with numbers that are theoretical engineering dynamic range pre shot.

The theoretical number for the new organic sensor is 88dB. Sounds great, doesn't it? Almost 30 stops (each 3 dB represents a doubling, or what we call a stop in photography). I guess those other sites didn't look to see what other sensor companies report their theoretical dB numbers as. I'll give you one for comparison: the Aptina AR1011HS sensor, which is a 10mp 1" sensor that's mostly the same as the sensor used in the original Nikon 1 cameras (missing the AF component): 84dB. Don't believe me? Click here (http://www.aptina.com/products/image_sensors/ar1011hs/), look in the right column. I used Aptina for a couple of reasons. First, they're one of the few sensor makers whose stated specifications are available to anyone on the Web (Sony keeps them mostly private these days). Second, I wanted a sensor that was in a camera that everyone could agree doesn't perform as well as the one in a Nikon D800.

So don't go putting your D800's on eBay because someone threw a faux stone at the ego you've got vested in your sensor. Your D800 still outperforms almost everything else on the market."

As I wrote, I have no idea what this all means, so I'd be interested to read what his contention means to your link, Sandy.

Can't tell you what it means but as you perhaps already know, Thom Hogan is affiliated with Nikon in some manner or fashion. I don't know and don't care to check on exactly what the connection is but he writes extensively, and perhaps exclusively, about Nikon products (used to do the same for Canon I believe). Which doesn't mean he has nothing worthwhile to say, just that I'd read anything he writes about any Nikon product relative to any potential competing product with that in mind.

[paulr]

A high-range sensor is useless unless the presentation medium can support it..

Why? I don't think there's any relation between the two at all. More dynamic range just means you can capture a scene with more brightness variance. How you map it to the output medium is just as arbitrary as it is with film.


A true 1-stop variation in subject illumination is dumbed-down to 1/2 stop or some such, so you really
accomplish nothing that couldn't be done by a regular sensor

What you accomplish is being able to capture detail in both the highlights and the shadows. If you don't, then it's gone.

What you call dumbing down is what photographer's have been doing since the beginning—representing a very high contrast world on papers that have just a few stops of luminance range.
I get about 12 stops of useful dynamic range out of 4x5 black and white film, doing n- development. This is almost always enough for my needs. My current digital camera gives me about 10. This is more than what used to be possible, and is plenty most of the time. But having another couple of stops would be great.

[paulr]

My brief search yielded claims as high as 14 stops. I can't remember shooting anything like that sort of range for landscape photos.

As an example, I'm doing work now in old, empty industrial spaces (http://paulraphaelson.com/phototext/#2). In many cases they're lit just by daylight coming in the windows. Many of the scenes include both unlit shadows, and the bright windows themselves. With 14 stops I could probably do this in a single exposure. With my current camera, I'm a few stops short, so this situation forces me to blend multiple exposures.

It works, but I don't find it ideal. The upside is that you have as much dynamic range as you need, and you can completely eliminate ANY visible noise. The downsides are tedium, huge volumes of data, headaches if anything in the frame is moving (even things like cobwebs blowing around in a draft cause problems), and occasional artifacts around high contrast edges that need to be retouched out.

I would much rather do this work in a single exposure every time.

[sanking]

Why? I don't think there's any relation between the two at all. More dynamic range just means you can capture a scene with more brightness variance. How you map it to the output medium is just as arbitrary as it is with film.

What you accomplish is being able to capture detail in both the highlights and the shadows. If you don't, then it's gone.

What you call dumbing down is what photographer's have been doing since the beginning—representing a very high contrast world on papers that have just a few stops of luminance range.
I get about 12 stops of useful dynamic range out of 4x5 black and white film, doing n- development. This is almost always enough for my needs. My current digital camera gives me about 10. This is more than what used to be possible, and is plenty most of the time. But having another couple of stops would be great.

You are absolutely right. More dynamic range simply means that you can capture a scene with a higher range of subject luminance.

Current top digital sensors have a dynamic range that is much greater than color transparency film, about the same as color negative film with the caveat that there is a lot more useful dynamic range in color negative film if you scan and print digitally than if you print optically.

With specialized development it is possible to get extremely high dynamic range with B&W film, 15-18 stops or even more. Pretty easy with a two-bath developer like Pyrocat-HD or Diafine.

As you say, how one maps the recorded image to output is a separate issue.

Sandy