GH5 Dual ISO or Large Pixels?

B

Bassman2003

Veteran
Hello,

As many are, I am excited to see the developments with noise and lower light shooting in the m4/3rds world. Looking at the GH5s, I wonder which aspect might be most responsible for the great low light, the Large pixels or the Dual ISO?

Makes me wonder what a GH6 might look like with a 20MP sensor along with Dual ISO? From my point of view, I do not need to see in the dark, just want to have a camera that is super clean up to ISO 3200-6400.

Seems Dual ISO would improve any camera or does it need fewer pixels in order to have the processing horsepower to complete the task?

Up for discussion...
 
Both. Mostly larger pixels / photo sites.

Dual ISO helps maintain the DR at high ISO which is very important on a limited DR camera such as this already. Also helps with noise.

The GH5 noise processing was already very very good - GH5s is next level.

A GH6 will almost certainly have GH5s lowlight + IBIS. But you'll be waiting quite a while I'd assume. Unless Panny really feels pressure from BMD, but then, I'm guessing they'd actually release another "5" variant with a letter monicker if it was next year. Unlikely I think.

The 6 may or may not implement the new organic sensor that is in development. Whenever that comes around, M43 should take a big step upwards - even into S35 territory (of course that goal post will keep moving too). But i see good things for future of M43 as the sensors mature, because the form factor and mount is brilliant, and once DR is 14+ stops and lowlight is super clean @ 6400, it's an extremely compelling system.
 
So if the current GH5 sensor happened to have Dual ISO, what would we see? would it be as clean as the GH5s up to certain levels like ISO 3200 or 6400? Because it seems that noise and DR are more important in the mix.
 
It isn't as easy as that. Dual ISO depends on a sensor that is capable of a certain range of sensitivity. So just adding dual ISO to a GH5 type sensor might not yield very impressive results. Maybe a one stop difference vs the almost three stops of the GH5S. I don't exactly how hard the GH5 sensor could be pushed with dual ISO but part of what makes dual ISO work so well on the GH5S is the ability to suck in more light with the larger pixels. I think we have seen evidence of this on the EVA1 where the dual ISO is only about 2 stops better because the EVA1 packs a lot of pixels on that sensor. So a pixel packer sensor like the GH5 might also struggle to push the dual ISO very high. It could still offer a benefit however and to some users maybe even that one stop would be a very welcome improvement. If we actually get dual ISO on a consumer camera is another story and it is possible Panasonic is saving that for their more pro line of cameras.
 
Thanks for your reply. The technology seems logical to me. I do not know how difficult or expensive it is to increase the number of analogue converters but It makes me wonder about tri or quad ISO approaches. There is a useful band of light gathering for *most* work but we all want whatever we shoot to be clean and full DR.
 
By the way, does anyone know how a sensor can have two native ISOs?

If you say it is by applying gain at different levels, then you're wrong, because "native" means no gain. If you say it is by having "two circuits" running through the sensor, then you're not specific enough. It's no better than saying, "The sensor has some stuff in it." It is almost a tautology, and you might as well say, "magic."

Does anyone have an electronic engineering explanation of dual native ISO?
 
Not entirely sure myself since I or anybody on here has never designed and built a sensor but I think it has something to do with the fact that a sensor is really just an input device. It sees light and converts that light to electronic values in a similar way that a solar panel turns light into energy. I think what the analogue circuits do is determine how to treat that electronic data. It can be mapped at one ISO level or another ISO level. The reason I believe you don't see every camera just set their base ISO to 3200 or something ridiculous is then a lower ISO level would suffer. That wouldn't make much sense since typically we try to shoot with a low ISO anyway. It would also force all of us to shoot with a lot of ND filtering outdoors. Thats why we have dual native ISO. Those basic sensor values can be mapped to a normal ISO level or a higher ISO level when needed.

Thats my take on it at least.
 
Actually, all sensors have an amplifier to get the analog voltage off of the photosite. After that, I'm not sure how they are doing dual ISO, but maybe with a low gain and a high gain internal amp (internal to the CMOS imager). https://www.youtube.com/watch?v=FKJFIzDfUNE

The advantage of CMOS is that you can stack the amplifier on the same silicon and right below the photosite making things smaller, less power, and less electrical noise since the amp is right behind the photosite.
 
My understanding is that it is all relative, where "native" really represents an analogue interpretation of the sensor output. Once it goes to analogue the characteristics are baked in. Since we only had one set of analogue converters in the past, the word native was used. Now that there is more than one analogue path, native is more of an epicenter or inflection point representing pure output from the digital sensor. As we move away from the native ISO the signal becomes less-pure. This makes me wonder if more than two are possible and if they would help.
 
Bassman2003 said:
My understanding is that it is all relative, where "native" really represents an analogue interpretation of the sensor output. Once it goes to analogue the characteristics are baked in. Since we only had one set of analogue converters in the past, the word native was used. Now that there is more than one analogue path, native is more of an epicenter or inflection point representing pure output from the digital sensor. As we move away from the native ISO the signal becomes less-pure. This makes me wonder if more than two are possible and if they would help.
Click to expand...

Not sure more than two would be worth it. The upper end is already likely going as high as it can without sacrificing quality. I think thats what the EVA1 can't push the dual native ISO as much as the Varicam. The upper ISO was as high as they could go due to the tiny pixels on the 5.7k sensor. Having a native ISO in the middle is likely a waste of effort since one either needs to shoot in low light or they don't. The gains between each level typically compensate enough making a middle ISO level almost pointless I would think.
 
ARRI ALEXA Sensor Tech

The Dual Gain Architecture simultaneously provides two separate read-out paths from each pixel with different amplification. The first path contains the regular, highly amplified signal. The second path contains a signal with lower amplification, to capture the information that is clipped in the first path. Both paths feed into the camera's A/D converters, delivering a 14 bit image for each path.
Click to expand...
 
But does the Alexa have dual ISO? The only cameras I have heard that have dual ISO are from Panasonic: the Varicam LT and the GH5-S.

And I have also been told that "native" means "no gain." Meanwhile, I would guess that "dual-gain architecture" means "gain."
 
Sony Venice, Red Gemini, and BMD PCC 4K have all introduced Dual ISO since Panasonic launched the Varicam LT.

Currently Panasonic has it in the Varicam LT, GH5S, and the EVA1.
 
combatentropy said:
But does the Alexa have dual ISO? The only cameras I have heard that have dual ISO are from Panasonic: the Varicam LT and the GH5-S.

And I have also been told that "native" means "no gain." Meanwhile, I would guess that "dual-gain architecture" means "gain."
Click to expand...

I don't know the details of how these sensors operate, but there is no such thing as "no gain". The electric signal coming from a pixel on the sensor is minuscule. It has to be amplified.
 
Dual native ISO generally means dual conversion gain. Conversion gain occurs at the electrons-to-voltage conversion stage which is before normal programmable analog gain stage which is a voltage-to-voltage stage just before the ADC. Many camera have used dual conversion gain for years. Panasonic was just the first to market it in camcorders. Most of the recent Sony mirrorless cameras have a low and high conversion gain mode that kicks in at certain ISOs.
 
Isn't the Alexa using this as a way to increase the range and not as a way to get better low light? They take two sets of outputs and combine then together to create 16bit. Thats a bit different than dual native ISO where the sensor can be amplified by different values for a specific output. I think.
 
combatentropy said:
But does the Alexa have dual ISO?
Click to expand...

ARRI is using the image signals differently. Instead of providing dual ISOs on the camera, they are using the two pixel gain paths to create a live HDR signal, where the low gain and high gain signals are combined to create a single HDR like image with extremely wide dynamic range. This is how ARRI was able to get very wide dynamic range from the sensor tech that was available in 2010.
 
nyvz said:
Dual native ISO generally means dual conversion gain. Conversion gain occurs at the electrons-to-voltage conversion stage which is before normal programmable analog gain stage which is a voltage-to-voltage stage just before the ADC. Many camera have used dual conversion gain for years. Panasonic was just the first to market it in camcorders. Most of the recent Sony mirrorless cameras have a low and high conversion gain mode that kicks in at certain ISOs.
Click to expand...

Thank you! This is the kind of answer I was looking for. So my confusion stemmed from not knowing that there were two different kinds of gain, that happen at two different stages in the pipeline.
 
I'm actually more interested in Arri's technique than being able to shoot a black cat in a cave. Would be lovely in an affordable, handheld camera. Seems like an elegant hack Panasonic could use to get 14+ stops.
 
joe12south said:
I'm actually more interested in Arri's technique than being able to shoot a black cat in a cave. Would be lovely in an affordable, handheld camera. Seems like an elegant hack Panasonic could use to get 14+ stops.
Click to expand...

I was thinking the exact same thing yesterday. Include the option to use the dual native ISO two ways. As a boost to ISO or combined as a way to push the range to 14+ stops. Would love to see this as a firmware update on the GH5S to give the industry reason to pay attention to it again. Include the option to boost the DR with the dual ISO levels and include the full V-log when doing so.
 
You must log in or register to reply here.
Back
Top