Choosing a Camera format for Underwater Photography

My last post on MFT vs APSC stirred a bit of noise so I though of doing a full write up

I compare pretty much anything from MFT to Full frame starting at 20 megapixels and looking at sensor data ad equivalent DOF

I am conscious this is not what everyone will want to hear however I have done some real research on the topic so I invite you to read the data before coming out with blanket statements and more misconceptions!

https://interceptor121.com/2020/03/07/choosing-a-camera-format-for-underwater-photography/

Great article and much needed. To support your article/post, I shoot with an MFT and when comparing shots am always pleased the my shots look as good as high end FF set ups. I do have to invest in good glass, including ports. 

I'm retiring from SCUBA for age (74) and physical issues. I will sell my rig and buy a another camera. As I read your article, as a walk about camera, it appears that I would be happy with a quality APSC. Correct? I have my eye on the Sony a6400. Would greatly appreciate your opinion. Even so, I take responsibility for my decisions.

Hi Bill

My article is focussed on underwater photography that has issue of depth of field domes etc etc.

On land I see more benefits from a larger sensor but frankly for walkabout I do not see how you could be disappointed with the Sony A6X series. I personally do not like Sony ergonomics, menu system and quirks about lenses but if you are happy with that I see no issue of any sort and is very small

121,  Great article.  You have really explained things in easy to understand terms.

I imagine you will have opened another can of worms with this article.

Thanks

Interesting article, as always.

I think that there is a think that we normally tent to miss when we talk about new gear. Who is going to use it? A beginner, an experienced hobbyist or a Pro?

I’m not surprise to hear that an experience photographer on a point and shoot pulls out better images than a hobbyist with an expensive rig. This is the same for any other field.

2.5 years ago I upgraded from a RX-100mii, after spending months on spreadsheets trying to balance my budget with the options I settled on the Oly E-M1m2. At that time, I didn’t wat to go into FF.

At the end, just before hitting the buy button I switched to the D500. I’m very happy with the D500, but I’m sure that I would be as happy with the Oly. With the time, I have seen improvement in my images that keep me from regretting expending all that money.

UW photography is “just” a hobby for me, and other from work (with phone) I normally don’t take pictures on land. I know that I don’t need / use all the options of the D500. But I love how fast the autofocus work with the Nikkor micro lenses.

Regarding your conclusion I agree with you, we hobbyist tent to expend way much money on equipment and gadgets. However, when I look at the equipment used in award winning competitions it is always FF & APS-C formats. I’m not saying DSRL is better, I have seen great videos and shots with the Panasonic and Sony cameras.

As a hobbyist, I don’t think that I can tell the difference from shots taken with DX and MFT system. Is it the same for everyone? Maybe…

I think one should consider the size and weight of the final rig (housing, arms, strobes, etc) and make sure you are happy with it.

About the extra cost from MFT and DX, for me it was around 15% and it wasn’t the most important thing. And there is nothing worse than looking at a shoot thinking that you should have bought a “better” system.

I think that this discussion involves more than technical facts and comparison, making it very hard to get a conclusion where everybody agrees.

Thanks Joss

The point was not agreeing on anything quite the opposite to give some options. Take also into account that the choice of a system cannot be based purely on sensor but involves ergonomics availability of optics and ports and other factors.

Most competitions as far as I understand are judged on screens and there is not guidance of what screen is used so unless you go 1:1 pixel and have same screens I would not think there is a difference between formats.

At risk of stirring another can of worms today the system that has the best compromise amongst all for underwater photography in my opinion is APSC DSLR. Despite the attempts of Nikon to self castrate it to avoid competition with larger formats (the D7200 was too good for the price point) you can see many stunning images with D series and also Canon 80D 90D series. The key asset of APSC are the lenses and port combination you can't go wrong with Tokina 10-17 and a 100mm macro canon or nikon. A native zoom fisheye option is lacking on other formats and creates the need for expensive water contact optics.

APSC is not that far from a 20 Mp MFT system however the choice depends on your path. I wanted a system that could take decent video and stills so I settled for mirrorless, someone with an estate of DX lenses will take APSC, someone with nothing may take any trajectory I guess.

At the end everything pretty much flattens out however it is important to correct wrong assumptions and myths and sensor choice being a very technical topic is number 1 on the culprit list. I just tried to move the focus elsewhere as there are other factors to be considered.

We are coming very near to the end of DSLR technology the path is clear and mirrorless will prevail. The last attempt of Nikon positions APSC mirrorless as a hobbyst camera the autofocus and other features are sub par and it seems this system will succumb to Sony, Panasonic and Olympus that have invested the last few years on EVIL camera while the big guys will go for full frame mirrorless. This may deprive underwater photography of choices and I think it is just a couple of years away. Both Canon and Nikon have launched new lens formats and the development on classic format has come to a stop. In perspective you will find my article very interesting but not too soon!

Thanks,

I am considering upgrading my Olympus EM5 Mk1 setup and main reason is low light performance.  The main issue putting off upgrade for the moment is that current usage frequency is not supporting the investment yet. I hope frequency and thus  incentive will increase with inverse proportionality to offspring age:-)

My secret whish is to be able to stay MFT, hoping for a disruptive increase in sensor sensitivity.    

 Cheers

/O

The post explains that at constrained depth of field low light performance does not depend on sensor size but on the camera having specific circuitry for high gain. This works better in conjunction with larger pixels compared to smaller.

The cameras with dual circuit tend to be video orientated not stills as having more megapixels even if smaller is better

In the case of the EM5 MKI the only MFT upgrade that makes sense is the OMD EM1 MKII as this has some unique features that stand out from the rest of the range. APSC will not make a difference and the next stop is full frame...

The article is interesting, and the theory described interesting. Overall, it is a good idea to debunk "myths". However, I think that some of the conclusions do not reflect actual photography usage, and should therefore be nuanced.

If we take the first "myth", you draw conclusions from the theory based on the assumption that the photographer works at constant depth of field. In that situation, indeed there is no difference in signal to noise ratio between large and small sensors.

When working at constant aperture, there is a significant difference, as illustrated in the Clarkvision article that you quoted:

These two pictures are taken with the same parameters: ISO 400, f/4.5, 1 second exposure. In this situation, the larger sensor gives significantly better results.

Depending on the situation, the photographer may have different objectives related to the depth of field (DoP). In some situations, the key will be whether one can shoot at the hyperfocal distance. In others, the objectives will be to minimize or maximize the DoP. Given these other parameters, the smaller sensors may be at disadvantage depending on the objectives of the photographer.

So I believe that beyond the theory, your guide should take more into account the type of usage and photography.

Your point related to the final output medium is also very relevant. To push the comparison a bit further, for most online media, the small camera sensors found on smartphones are totally sufficient.

My own conclusion was that APS-C provides an excellent compromise between the various parameters. But I understand that others may have different needs and objectives and would weight the strengths and weaknesses differently.

Not at all. The comparison between two images needs to be done with

Aperture is a relative concept to the focal length so same f number on two different lenses means something different

1. same field of view

2. Same physical aperture

3. Same exposure time

This creates a constrain on DOF that isolates how a sensor performs compared to another

If you change those settings that no comparisons is possible. Typical misconception you take

1. Same field of view

2. Same shutter

And then change the apertures so that the physical aperture is different the amount of light is not the same, the image looks different and there is no comparison. Which is exactly what is shown in the image you posted to show you don't draw comparisons if the two exposures are not equalised.

I am not selling camera or housing and I think that the articles substantially shows that there are a lot of misconceptions and errors and therefore make incorrect judgement. You cant compare two cameras just with the fnumber iso and shutter because other than shutter the other parameters are relative not absolute

Most people do not even know that ISO on format A is different from ISO on format B

Looking at usage of the gear underwater photography is a nice middle of the road, is not very bright so that you have DR issues, it is not too dim that you are talking about night shots so the consideration on constrained DOF applies in full and you need to make additional considerations to choose a camera. This is in fact the reason why you don't really see a gap between formats and cropped formats are the majority otherwise everybody would be shooting full frame. In fact you shoot APSC, if you had to shoot same aperture of full frame you would probably end with worse images...

1 hour ago, Algwyn said:

The article is interesting, and the theory described interesting. Overall, it is a good idea to debunk "myths". However, I think that some of the conclusions do not reflect actual photography usage, and should therefore be nuanced.

If we take the first "myth", you draw conclusions from the theory based on the assumption that the photographer works at constant depth of field. In that situation, indeed there is no difference in signal to noise ratio between large and small sensors.

When working at constant aperture, there is a significant difference, as illustrated in the Clarkvision article that you quoted:

These two pictures are taken with the same parameters: ISO 400, f/4.5, 1 second exposure. In this situation, the larger sensor gives significantly better results.

Depending on the situation, the photographer may have different objectives related to the depth of field (DoP). In some situations, the key will be whether one can shoot at the hyperfocal distance. In others, the objectives will be to minimize or maximize the DoP. Given these other parameters, the smaller sensors may be at disadvantage depending on the objectives of the photographer.

So I believe that beyond the theory, your guide should take more into account the type of usage and photography.

Your point related to the final output medium is also very relevant. To push the comparison a bit further, for most online media, the small camera sensors found on smartphones are totally sufficient.

My own conclusion was that APS-C provides an excellent compromise between the various parameters. But I understand that others may have different needs and objectives and would weight the strengths and weaknesses differently.

 

 

 

This is what is written under the image you just posted. If you have not understood is a case of reading it again.

Figure 5. Full image crops from Figures 4a and 4b. The crops are from raw converted 16-bit TIF images converted by Canon Zoombrowser software with the same (default) settings for both cameras. Both cameras recorded the same exposure at the same ISO, and same f/ratio: ISO 400, f/4.5, 1 second exposure. The larger lens diameter collected more light and fed that light to the pixels of the 1D Mark II to produce higher signal-to-noise images than the small lens diameter feeding the small sensor, small pixels of the S70. Both cameras are photon noise limited, as are these images. The signal-to-noise ratio is abut 3.5 times higher for the 1D Mark II image, close to the ratio of the lens aperture diameter, the pixel pitch, and the sensor sizes from each camera. The CAUSE is the lens which delivered more light to the 1D Mark II image, making the better image. The sensor and pixels are just buckets to collect the light delivered by the lens.

In the most simplistic terms... mid to high end DSLR's (current generation) offer better AF performance than comparable current generation MFT cameras and mirrorless cameras. This is from first-hand experience (our family shoots D850, D500, Sony A7iii, and Oly). Add in to this low light performance and the gap grows.

This is changing, and it's pretty clear that mirrorless will eventually pass DSLRs in performance and output capability, however there is still (also) the lens availability challenge. In many of these discussions we read about interesting (and in some cases, fairly exotic) camera lens & port combinations - usually optimized for some specific scene or to capture a specific subject, however, until either a MFT & lens or a mirrorless camera & lens can match the all-round performance of a D500 matched with Nikon AF-S Micro Nikkor 60 f/2,8G ED lens there will be a gap (I know the following are not necessarily representative of everyone's way of shooting - just using for illustration).

Two straight-forward test scenarios:
D500 + 60mm using a snoot (like the new Backscatter strobe/snoot) 
D500 + 60mm on a blackwater dive (shooting in low light/variable light)

(personally I would then add in the D850 + 28-70 shooting with the WACP for W/A)

I don't shoot Canon, but I'm assuming they have a combination that performs similarly.

Until MFT and mirrorless can match this type of performance there will remain a noticeable difference. 

As I said, mirrorless is getting there - however even with recent improvements the blackwater scenario still shows the gap (I've seen it first hand).



 

41 minutes ago, oneyellowtang said:

In the most simplistic terms... mid to high end DSLR's (current generation) offer better AF performance than comparable current generation MFT cameras and mirrorless cameras. This is from first-hand experience (our family shoots D850, D500, Sony A7iii, and Oly). Add in to this low light performance and the gap grows.

This is changing, and it's pretty clear that mirrorless will eventually pass DSLRs in performance and output capability, however there is still (also) the lens availability challenge. In many of these discussions we read about interesting (and in some cases, fairly exotic) camera lens & port combinations - usually optimized for some specific scene or to capture a specific subject, however, until either a MFT & lens or a mirrorless camera & lens can match the all-round performance of a D500 matched with Nikon AF-S Micro Nikkor 60 f/2,8G ED lens there will be a gap (I know the following are not necessarily representative of everyone's way of shooting - just using for illustration).

Two straight-forward test scenarios:
D500 + 60mm using a snoot (like the new Backscatter strobe/snoot) 
D500 + 60mm on a blackwater dive (shooting in low light/variable light)

(personally I would then add in the D850 + 28-70 shooting with the WACP for W/A)

I don't shoot Canon, but I'm assuming they have a combination that performs similarly.

Until MFT and mirrorless can match this type of performance there will remain a noticeable difference. 

As I said, mirrorless is getting there - however even with recent improvements the blackwater scenario still shows the gap (I've seen it first hand).



 

Hi those are ergonomics and usability points that are indeed very important to make a camera choice however that was not the purpose of my post. My articles only want to show that sensor size is not necessarily the gateway to DR low noise etc etc

in terms of low light performance DSLR has no edge on anything as explained there are other dependencies typically linked to a high gain circuit or not in the camera.

In fact the posts show that the sensor sizes are almost equivalent at constrained depth of field so to choose a camera you should look at other criteria, port and lenses available, autofocus, color rendition, battery usage, controls and ergonomics, the list gets longer

I am not in the business of making recommendation of what camera is better than another I only like the intellectual challenge when I read some misconceptions

In terms of mirrorless vs DSLR I think is a matter of time and DSLR will die like it or not but that is a whole different story and debate

I believe that you are missing the point made by @oneyellowtang.

There are other sensitivity measures than the main sensor sensitivity. Low light sensitivity of the Autofocus system is one of them. He was referring to the ability of the camera to focus in low light situations, and more generally of the autofocus performance of the various systems.

Also, you cannot draw conclusions based on the performance of the camera sensor only. You cannot shoot a picture without a lens. What we need to take into account is the system: camera body + range of lenses available. 

20 hours ago, Interceptor121 said:

The signal-to-noise ratio is abut 3.5 times higher for the 1D Mark II image, close to the ratio of the lens aperture diameter, the pixel pitch, and the sensor sizes from each camera. The CAUSE is the lens which delivered more light to the 1D Mark II image, making the better image.

That's the whole point, the system (camera body + lens) deliver more light to sensor, which results in better image. For a picture where depth of field is not significant, this matters.

4 minutes ago, Algwyn said:

I believe that you are missing the point made by @oneyellowtang.

There are other sensitivity measures than the main sensor sensitivity. Low light sensitivity of the Autofocus system is one of them. He was referring to the ability of the camera to focus in low light situations, and more generally of the autofocus performance of the various systems.

Also, you cannot draw conclusions based on the performance of the camera sensor only. You cannot shoot a picture without a lens. What we need to take into account is the system: camera body + range of lenses available. 

That's the whole point, the system (camera body + lens) deliver more light to sensor, which results in better image. For a picture where depth of field is not significant, this matters.

Can you give me an example of picture you took underwater below f/1.4? That is physically where there would be a benefit

There are plenty of very fast lenses for all formats this is no longer a distinction unless you go below f/2.8 on full frame and there are no lenses like that used underwater

The copy paste you put there shows you don't actually understand how a camera sensor works my post had to objective to help people like you make better decision, if you decide not to is totally fine but please do not come up with alternatives to physics

Physics is always right as it is based on extrapolation of formulas out of facts

The autofocus discussion is a separate item and is not a prerogative of formats anymore, there are hybrid and phase detect on all formats but that was not the point of my post

Do the sensors (Or pixels actually)  in question sutúrate at the same point or can the larger pixel hold a higher charge?

23 minutes ago, Akoni said:

Do the sensors (Or pixels actually)  in question sutúrate at the same point or can the larger pixel hold a higher charge?

Excellent question!

The pixels  saturates. Think about buckets the small buckets fill first and this is the reason that at the lowest possible ISO larger pixels sensor (irrespective of sensor size) will have more dynamic range at matching technology

This is the reason why I wrote that unless you shoot at very low ISO you see no difference between sensors. In fact the difference and benefit appear at the lowest ISO

Larger pixels will also hold more dynamic range as ISO grows look at the image comparing same megapixels but different sensor (and pixel size). The point is that underwater you do not really shoot at ISO 12800 so this is actually not a practical benefit except on land

Essentially when you operate middle of the road gains or ISO that is the typical underwater scenario none of this makes a difference and other factors become predominant which was the point of the article really