Does 5D have noise advantage because of full frame sensor?

[lucas_jarvis]

Looking at %100 cropps, one would conclude that the 5D has a minimal noise

advantage over a camera like the 30D. I think a third of a stop is the

official advantage.

 

BUT, recently I've been reading that this may be misleading - that in fact the

5D could have more that a stop advantage over the 30D.

 

The idea is that when the 30D files are 'upscaled' to match the pixel

dimensions of a 5D, the %100 crops of the 30D's noise have been magnified. A %

100 percent crop comparison will now show the 30D to have more noise.

 

Another example would be a typical 8x12 print lets say. If two identical

pictures were printed from each camera at that size of print, the 5D would

show less noise. The 5D's print would have a higher DPI making the

noise 'finer' I guess you could say.

 

Does all of this sound like reasonable proof that the 5D has over a stop of

noise advantage compared to the 30D? Or is it a concept that is flawed in it's

explanation?

 

I'm a little confused in the theory because when I originally read it, it was

explained that the size of the sensor was the reasoning behind all of this,

yet I'm wondering if it has to do with pixel count.

 

If anyone can add to this, maybe we can clear it up and get to the truth.

Hopefully I've explained it properly and have not messed up the concept.

[bennyboy]

The pixel density of the 5D's sensor is lower than that of the 30D, this means that there is less inteference/crosstalk electronically between them = less noise. Not a huge difference, but one nonetheless.

[lucas_jarvis]

That would explain the 1/3 stop advantage at %100 crop, but aren't we talking about a completly different thing here?

[bennyboy]

No, we're talking about exactly the same thing. But I've never heard of noise being measured in stops.

[robert lee]

Here's the deal. All things equal (not really the case, but close enough) all sensors have to contend with the same thermal noise level. Think of this as the raw background hiss that can't be gotten rid of unless the thing is cooled to 0 Kelvin.

 

Since there's only so much that can be done to lower the noise floor, the other end of the attack is to increase the amount of image energy hitting a sensor site.

 

The goal is maximizing the ratio of signal to noise. The most obvious way to do this is to use sensor sites with larger area, i.e., imaging chips with lower sensor density. To a large extent, this is why the 8MP image from a APS-C DSLR is much cleaner than that from an 8MP pocketable P&S.

[gluteal cleft]

From listening to the folks who know a lot more about the causes of noise than I do, the ultimate limit in noise depends upon the square root of the full-well capacity (in electrons) of a pixel, so yes, the larger pixels provide a theoretical advantage. In fact, it's not just theoretical, but also real-world.

 

However, that's not to say that cameras are actually at that limit. There are other things that can cause noise, and have to be controlled, such as noise in the amplifiers and atomic-level differences between pixels. As time goes on, manufacturers find more and more ways to keep those causes in check, so you see small, incremental improvements over time from that. As one example, Canon is reputed to use a different set of amplification transistors for each ISO stop, matching the characteristics of the transistors to the particular job.

 

Your idea about noise being less evident in a print from the 5D is interesting - your assumption that given equally-sized prints, a pixel of noise from the 5D will (on the print) be smaller than a pixel of noise from the 30D is something to think about.

 

In any event, whether the improvements are mostly from larger pixels size, mostly from other improvements, or from magic pixie dust, the 5D is impressively good. Some ISO 3200 shots have been circulated that are real jaw-droppers, even at a 100% viewing level, not just in a print. The overall advantage is probably a combination of all of the above factors.

 

steve

[harvey_edelstein1]

In a test of the 1dsII vs. the Nikon d2x, the main advantage imho is the greater latitude of the FF sensor 1dsII over the Dx2. Its like color slide film being comparred to color print film. Dpreview who ran the comparison gave the FF Canon about a 2 stop advantage. This means that metering if less exact will still process more shadow and highlight detail without burn out or blackend shadows. The Canon 5D operating with the latest sensor technologies is even better and the low noise is an extra added benefit. Now lets see improvements in the user interface and we will really have a winner.

[bennyboy]

The user interface works just fine for me. But then I don't want to shoot with just one hand ;)