Megapixels vs. Sensor size.

[maflynn]

So, I was wondering what people's take on this are.

I'm testing out a compact camera with a 1" sensor size but at 20 megapixels vs. My M43 camera at 16Mp

All things being equal, wouldn't the images from the 1" sensor produce higher image quality then the M43?

Wouldn't I be able to crop the image from the G5X II to a higher degree then the OMD?

I'm playing around with the images, but I'd like other people's input and thoughts.

 

[mollyc]

I would expect the 4/3 to be cleaner since it is a larger sensor. I would also expect larger prints/enlargements from the 4/3.

 

[dwig]

There are way too many other variables to merely use sensor size and megapixels to determine "quality".

As a rule, the larger the photosites are the lower the noise. A m43 sensor at 16mp has significantly larger photosites than a 1" sensor at 20mp. Also, higher mp usually means more detail, but if and only if the lens is sharp enough to deliver the detail to the sensor. Confusing everything is that other subtleties in sensor design and make one sensor deliver better quality than another of the same size and mp.

 

[kallisti]

There are way too many other variables to merely use sensor size and megapixels to determine "quality".

As a rule, the larger the photosites are the lower the noise. A m43 sensor at 16mp has significantly larger photosites than a 1" sensor at 20mp. Also, higher mp usually means more detail, but if and only if the lens is sharp enough to deliver the detail to the sensor. Confusing everything is that other subtleties in sensor design and make one sensor deliver better quality than another of the same size and mp.

Well stated and I believe this is correct. There are two main issues regarding the OP's post--differences in how the sensors handle noise and potential differences in whether the lens on the compact camera is able to take advantage of the increased MP of the smaller sensor.

Including a pic below (with a crop) which may serve as an example for some of the issues involved. This was taken with a Nikon Z7 and Nikon Z 24-70 at 70mm, f/2.8, 1/50th sec, ISO 25600. The camera has a 45.7 MP full frame sensor.

The noise is obvious and nasty. It's a large file taken with a large MP sensor, but the IQ of the image is crap despite the large file size. I intentionally shot this with a crazy-high ISO. Larger MP doesn't necessarily translate to better image quality if there is significant noise in the image. To get back to the OP, if the noise in the image with the smaller sensor is worse then you aren't going to see improvements in IQ (assuming you are shooting images where noise can be an issue--i.e. not perfect light, low light, high DR images, etc.). If the bigger (but lower MP) sensor can handle noise better, the smaller files may actually be more usable. Stated another way, garbage in equals garbage out.

As stated by @dwig, the theoretical advantages of the smaller-sized but higher-MP sensor regarding sharpness can only be realized if the lens on the smaller sensor camera is of sufficient quality to take advantage of the higher-MP sensor--something that may not be true with a compact camera.

If you have the inclination to get really technical, you can use a program like RawDigger to compare the quality of the RAW files taken with both cameras. This program doesn't address sharpness issues, but it does compare the tonal continuity in the files (which is important for things like banding and smooth tonal transitions).

 

[Darmok N Jalad]

I’ve owned both a 20MP 1” sensor (FZ1000) and a 16MP m43 (G85). You can get great results with both, but I think optics also come into play. Also, make sure both cameras shoot in the same aspect. My FZ1000 only did 20MP in 3:2, but dropped to 16MP in 4:3.

 

[Ledgem]

There are way too many other variables to merely use sensor size and megapixels to determine "quality".

As a rule, the larger the photosites are the lower the noise. A m43 sensor at 16mp has significantly larger photosites than a 1" sensor at 20mp. Also, higher mp usually means more detail, but if and only if the lens is sharp enough to deliver the detail to the sensor. Confusing everything is that other subtleties in sensor design and make one sensor deliver better quality than another of the same size and mp.

Everything written there is generally accepted as the general consensus for most photographers. Yet I've also read dissenting opinions from people who seem (key word) pretty well-versed in optics and electronics who claim that increased megapixel count increases performance of older lenses, can improve noise, and can improve dynamic range. Sure, sharper lenses will make a higher megapixel count really stand out, but the key thing is that a higher megapixel count will never make an image look worse.

The long and short of it is that sensors are complex enough that the number of variables go far beyond photo well size and physical size/megapixel count. dwig's first line was absolutely right. If mayflynn's µ4/3 camera is much older and the compact camera is much newer than it's quite possible that the performance could be similar or even better on the compact. Heck, modern µ4/3 cameras outperform older "full frame" cameras... so why not? But if both cameras are of the same generation then the µ4/3 camera would be expected to be superior.

 

[kallisti]

Sure, sharper lenses will make a higher megapixel count really stand out, but the key thing is that a higher megapixel count will never make an image look worse.

Another excellent reply and food for thought.

Regarding the line I bolded, I'm not certain this is entirely true. Higher MP sensors can make errors in focus more obvious. Assuming the lens is sharp enough, areas in perfect focus will be significantly sharper than areas not in perfect focus. On a lower MP sensor, there can be less of an obvious transition between areas that are in-focus and not in-focus (because even the in-focus area isn't as sharp as on a higher MP sensor), so the focus error might not be as obvious.

Consider a portrait as an example. With a higher MP sensor, it is much more obvious whether you got perfect focus on the subject's eyes or not compared to using a lower MP sensor. With the higher MP sensor, a focus error will result in a sharp cheek, forehead, or ear (or eyebrow or eyelashes in more subtle cases) with the eye obviously not as sharp in comparison. Not as obvious of a problem with a lower MP sensor since the focus error won't be as obvious (everything is a little softer in the image so the difference between in-focus and out-of-focus isn't as glaringly obvious). There are many other variables that influence how obvious focus errors are (lens sharpness, DOF, placement of the zone of sharp focus within the image, image output size/resolution, etc.).

If you *nail* focus (including placement of your DOF within the image), then a higher MP sensor will never make an image look worse. If you *don't* nail focus, then a lower MP sensor can help hide the focus error and perhaps result in a more pleasing image overall.

There are rumors in Leica forums that a future M might have a 40 MP sensor compared to the current 24 MP sensors. For a RF camera, I'm not sure this would be a good move as focus can already be problematic unless using an external EVF. A 40 MP sensor would make a future Leica M even more challenging to use, especially if one has aging eyes. But that's a topic for a different thread .

I'm getting into the weeds regarding the OP's question. But I thought it useful to comment on this particular aspect of your well-thought-out reply.

 

[dwig]

...Yet I've also read dissenting opinions from people who seem (key word) pretty well-versed in optics and electronics who claim that increased megapixel count increases performance of older lenses, can improve noise, and can improve dynamic range. Sure, sharper lenses will make a higher megapixel count really stand out, but the key thing is that a higher megapixel count will never make an image look worse.
...

Those claims need to be qualified:

• Increased performance of older lenses: False as stated, but higher MP sensors can lead to reduced image degradation caused by the sensor mechanics (micro lenses, AA filter, ...) if, and only if, the image is downsampled for final use. This is not increasing the performance but reducing the quality loss and only applied to older lenses designed for film. Modern high quality lenses designed for digital don't have the degradation problems so only display sharper images with high MP sensors when their native resolution is higher than a lower MP sensor can use.
  
• Improved noise: False as stated. Higher MP sensors often have lower noise because they are also using newer tech than the older lower MP sensors and thus less noise. It is not the fact that they are higher MP; it is because they are newer. Also, even when the tech is the same and a 16mp version has lower noise than a 24mp version, if you downsample both images to 8mp the downsampled 24mp image may well display less noise as more original pixels are averaged to produce the final pixels.
• Improved dynamic range: False as stated. The improved DR in higher MP sensors is because they use newer tech and not because of the higher MP.

 

[Ledgem]

Regarding the line I bolded, I'm not certain this is entirely true. Higher MP sensors can make errors in focus more obvious. Assuming the lens is sharp enough, areas in perfect focus will be significantly sharper than areas not in perfect focus. On a lower MP sensor, there can be less of an obvious transition between areas that are in-focus and not in-focus (because even the in-focus area isn't as sharp as on a higher MP sensor), so the focus error might not be as obvious.

You're right, but in this case I'd disagree with the notion that the image is any worse. If a lower resolution is hiding an error in focus, you still have the option of downsampling your higher-resolution image to a lower resolution to get the same effect. The flaw may be more apparent, but it was there to begin with - the higher resolution didn't introduce it.

Those claims need to be qualified:

• Increased performance of older lenses: False as stated, but higher MP sensors can lead to reduced image degradation caused by the sensor mechanics (micro lenses, AA filter, ...) if, and only if, the image is downsampled for final use. This is not increasing the performance but reducing the quality loss and only applied to older lenses designed for film. Modern high quality lenses designed for digital don't have the degradation problems so only display sharper images with high MP sensors when their native resolution is higher than a lower MP sensor can use.
  
• Improved noise: False as stated. Higher MP sensors often have lower noise because they are also using newer tech than the older lower MP sensors and thus less noise. It is not the fact that they are higher MP; it is because they are newer. Also, even when the tech is the same and a 16mp version has lower noise than a 24mp version, if you downsample both images to 8mp the downsampled 24mp image may well display less noise as more original pixels are averaged to produce the final pixels.
• Improved dynamic range: False as stated. The improved DR in higher MP sensors is because they use newer tech and not because of the higher MP.

I agree with you that it's not purely the megapixel count that leads to the improvements. However, the claims I cited were a response to some of the usual complaining of increasing megapixel count, as people remarked that low-light performance and dynamic range would be greatly improved if the underlying technologies advanced and the megapixel count remained the same.

Granted, I don't know the full explanations behind the claims so I can't really speak more about them than that. It's certainly intriguing ideas, though, and I wouldn't be surprised if there really was something behind them.