Panasonic Researching Sensor That Captures HDR in One Shot

Megalith

24-bit/48kHz
Staff member
Joined
Aug 20, 2006
Messages
13,000
I am not the biggest fan of HDR photos, as most of them look overly surreal to my eyes, but Panasonic is working on new tech that will make it much simpler to capture images with much greater dynamic range. Traditionally, you’d use the bracket function on your camera to get a series of differently exposed shots and combine them in Lightroom or similar software, but the sensor described here could allow for ten times higher saturation and over 120dB of DR with one click. While there are DSLRs and smartphones out there with an “HDR mode,” I think those rely more on in-device software processing and trickery---Panasonic’s results are probably more genuine.

The organic photoelectric conversion film layer is stacked on top of a traditional CMOS chip, according to a Google Translate interpretation of the research. That layer structure allowed developers to divide the pixel’s electrodes into large and small areas. The sensor can then change the voltage applied to the first layer, essentially adjusting how sensitive the sensor is to light on a per-pixel basis. The effect is a wider dynamic range that exceeds the 120dB standard limitation. The sensor was also paired with a global shutter for eliminating motion distortion.
 
I have HDR software and with it I can control how much saturation and HDR are in the final photo. Doubt you can adjust much in-camera.
 
I have HDR software and with it I can control how much saturation and HDR are in the final photo. Doubt you can adjust much in-camera.
The way HDR works usually is by having the camera sensor take two photos, one at a low exposure time, one at a slightly higher exposure time, and then you merge the two images together.

That causes problems if someone is moving very fast, this is actually really good.
 
The way HDR works usually is by having the camera sensor take two photos, one at a low exposure time, one at a slightly higher exposure time, and then you merge the two images together.

That causes problems if someone is moving very fast, this is actually really good.

Well, it's usually 3 or more exposures, the more the better the effect, but yes this camera will help with motion blur. Still can't make many adjustments like I can with my software. People say HDR looks too fake, etc. but that is due to settings, you can make it barely noticeable if you choose. With the software I have I can make pseudo HDR too. I can create multiple exposures from one exposure, etc.
 
I am not the biggest fan of HDR photos, as most of them look overly surreal to my eyes, but Panasonic is working on new tech that will make it much simpler to capture images with much greater dynamic range. Traditionally, you’d use the bracket function on your camera to get a series of differently exposed shots and combine them in Lightroom or similar software, but the sensor described here could allow for ten times higher saturation and over 120dB of DR with one click. While there are DSLRs and smartphones out there with an “HDR mode,” I think those rely more on in-device software processing and trickery---Panasonic’s results are probably more genuine.

The organic photoelectric conversion film layer is stacked on top of a traditional CMOS chip, according to a Google Translate interpretation of the research. That layer structure allowed developers to divide the pixel’s electrodes into large and small areas. The sensor can then change the voltage applied to the first layer, essentially adjusting how sensitive the sensor is to light on a per-pixel basis. The effect is a wider dynamic range that exceeds the 120dB standard limitation. The sensor was also paired with a global shutter for eliminating motion distortion.
Ignoring those who are new to HDR, the reason they look surreal is because the person who took the image wanted to make it look surreal. Trey Ratcliff is perfectly capable of taking normal pictures. What he does in post processing is often more interesting. And clearly photographers agree, since he makes a ton of money by teaching others photography.

https://stuckincustoms.smugmug.com/
 
This is relevant to my interests, one of the biggest advantages that the super expensive medium format backs have is that they have 14 stops of dynamic range as opposed to 11 or so (and usefully, lower) on a good SLR. If that can increase it will definitely improve the possibilities, though it'd be useful to make the range selectable (not a problem with raws obviously)
 
HDR makes for a recognizable marketing (and headline-grabbing) term, but it's really meaningless here:

The concept in application is essentially higher dynamic range, and what it *really* is, is more highlight headroom; as in, a sensor with this technology could be set to expose far lower (or far longer) such that the darkest tones (shadows) get exposed above the noise floor, while the brightest tones (the highlights) are also correctly exposed.

You'd still be noise floor limited, i.e., you'd still need longer exposures, but you'd only need one (or fewer) than with current sensor designs.

As far as the typical HDR 'look' that is rightly derided- HDR is a means of presenting more information, in photography meaning more color and tone information, and it brings more processing freedom to photographers after the shot or shots are taken.

[the talk about the global shutter is cool, but while it may be technically related, it's really a separate discussion]
 
Well, it's usually 3 or more exposures, the more the better the effect, but yes this camera will help with motion blur. Still can't make many adjustments like I can with my software. People say HDR looks too fake, etc. but that is due to settings, you can make it barely noticeable if you choose. With the software I have I can make pseudo HDR too. I can create multiple exposures from one exposure, etc.

Well, first you can do the exact same editing with this. I would really doubt the put some DRM in to prevent you from editing the raw images... That would guarantee that no professional would buy it, and they are the ones with the money for this.
And second, Would you rather start your edits with a more accurate photo (that is what this camera is doing for you) or less accurate? It really means that you have to add less fakery to get the HDR level/contrast that you are trying to create.

Your position is basicly the same as saying "You don't need a sensor with high sensitivity... I can do the same thing by adjusting the exposure in (insert editing program of choice)."
 
What I consider good HDR you wouldn't even know it while looking at it. My current camera records just over 2 stops in raw I believe which is enough most times. I've done a few bracketed ones but it's not to the point I'd like to buy software for it so I just tone map with what I have.

However, the thing is this may not mean much without a background in photography but this is excellent as it would grab more data per shot allowing for better post processing - which is one of the reasons I use a DSLR in raw anyway. Sony was reportedly working on something like this years ago. I wonder what happened seeing as everyone and their mother (besides cannon) uses a Sony sensor. It could be that why we are seeing higher amounts of light data in newer sensors I think they are well above 3 now.
 
Well, it's usually 3 or more exposures, the more the better the effect, but yes this camera will help with motion blur. Still can't make many adjustments like I can with my software. People say HDR looks too fake, etc. but that is due to settings, you can make it barely noticeable if you choose. With the software I have I can make pseudo HDR too. I can create multiple exposures from one exposure, etc.
You do understand that if you increase the exposure in software on 1 image you will get additional noise correct? This will not only get rid of motion blur but the picture should be less noisy as well.
 
Most better camera sensors are taking 14 bit images. Compared to 8 bit jpeg, you can do single image HDR. how many pictures you bracket impacts how broad your dynamic range is.

The unrealistic stuff is just harshly mapped HDR, or possibly not even HDR on the input side. You can use HDR sources to feed into reasonably executed tone mapping parameters and get stuf that is not surreal looking.
 
Back
Top