Jump to content
Sign in to follow this  
mathis94

How to have a clear photo of a fish moving in front of a camera ?

Recommended Posts

Hello,

 

I'm currently working on a project with Salmon fish in Norway.

 

The goal i to have a high resolution and very sharp picture of a fish moving in front of an underwater camera in the Fjords.

I tried with a GoPro 4 using high FPS video , "normal photo ", timelapse etc. and I always end up with a blurry picture (see attached for example).

 

I thought about different ways to do as :

 

- Use lights with fast shutter speed : Problem > I must not have any light reflection on the fish skin

- Move the camera while the fish is moving : I tried at my office with a simple object, I got sharper image but it still blurry

- Use a very high speed camera (1000fps for example) : Didn't try that

- Use a very high ISO camera like the Sony a7S II : Didn't try that

- Use a special algorithm for motion blurry (http://cg.postech.ac.kr/research/fast_motion_deblurring/) : Try it, a bit sharper but not enough

 

What would be according to you the best way to do so ?

 

Thanks a lot !

 

 

 

post-53017-0-13299000-1446726703_thumb.png

  • Like 1

Share this post


Link to post
Share on other sites

Hey Mathis

 

It'd be worth contacting Tom Kline - who is a Wetpixler - in Alaska. Tom specialises in salmon: photographing them I think rather than just eating them. But he may do both.

 

I suspect the problem is the GoPro isn't really suitable.

Share this post


Link to post
Share on other sites

Thanks Tim! I do both!

 

I believe the problem is that a salmon sees us as a potential predator (a good reason!) and so will move away from us very quickly. However, if one is very still they can be approached. For example, last month I had one rest against my foot (but another salmonid, a Dolly Varden) while I was standing in a stream shooting them spawning by remote control - you can see examples of my work on salmonography.com. However, I avoid being in the water with them as much as is possible for a given shooting situation. I was in the water to shoot the DV because a treefall next to the spawners was blocking my view of them from out of the water. So I stood just behind it and ~1.5 m from the activity to see and know when to "push the button". Still, I could only see the front end of the fish (which was enough) as the fallen tree blocked the rest.

 

Your image suggests a possible solution. You will notice that some of the salmon in the background are sharp - also it looks like your visibility in Norway is much better than what is typical in Alaska fjords, at least in the summer. I do not have a Gopro but saw some University of Oregon students use them at one of my sites. They left the cameras on the bottom in some sort of self-shooting mode and got some decent shots. If your Gopro can do this you might consider clamping yours to a fixture holding the net (visible in your shot) and letting it shoot for you. The next level up in expenditure would be to get a pole cam set up for it. You may want to use this from the surface (from a boat?). The next level up would be a remote control to use with the pole cam so you could decide exactly when to shoot.

 

I use a pole cam type solution. There is a blurb on it on the about me page on my site. As well, I have made a few technique postings on Wetpixel over the past decade. Most recently I have added a Canon 1DX to my arsenal. When doing available light shots with the 1DX I have been using autoISO with a 12800 ceiling value. Many shots end up being done at 12800 as the light level is not too high here even during the long day time (i.e., July). I did one shoot at 25600 (in May) because it was after the sun went behind the ridge forming the valley I was in, which was forested, and it was getting close 7pm. Even with such high ISO the shutter speeds get a bit long for a moving fish. I do not go higher in ISO because I need to do quite a bit of color correction as well as some highlight recovery - recall that dynamic range goes down as ISO goes up and I need as much of this as well. I also do long soak times. I have had a camera in one spot without moving it for as long as 12 hours. More typically it is less, but a few hours nonetheless. The pro body DSLRs have a big enough battery power to last. IMHO this is one of the main shortcomings of any camera that uses live view as they use more power and seem to have undersized batteries. As well one needs very fast autofocus that works well in low contrast.

 

Any questions?

Good luck!

Tom

  • Like 1

Share this post


Link to post
Share on other sites

Thanks for answering Tom and Tim :).

Im curious to know what kind of lights where you using ? When I used lights on the fish the skin was reflecting it and I must not have any light reflexion on my pics :(

Also the idea is to put the camera/lights underwater, and let it do the job and get as much sharp pictures of fish swimming close to the lens (from 0 to 40cm close)

So I cant be behind the camera all the time to trigger a shot :/

Share this post


Link to post
Share on other sites

I am posting here a couple of "technique" shots that may better explain what I do. Both cameras in the shot showing two housings are set up with a single Seacam strobe positioned at 12 o'clock and above the water. These are my primary light sources. I also have some available light pix done with no strobe - most of the recent (last few years) shots on my site showing schools of salmon are done this way. High ISO is generally needed. Available light shots generally have less issues with reflective parts of salmon bodies. Even when it is sunny such as the shot with two housings, the light is very patchy in forest settings so strobes are generally needed on the spawning grounds. There is no sun at all hitting the ground late in the season, like right now (November), at spawning grounds located in valleys due to very low sun angle.

 

You can see the remote cord from the housing in top of the frame (two housing pic) - it ran over to where I was standing to shoot the technique shot. From this vantage point I was able to shoot both cameras. Note that there is a salmon pair in front of each housing. These salmon are about 0.5m long (for scale). I have to anticipate where a salmon is going to be in the next second when I push the button. Two housings (different models, one being older or "obsolete") is a convenient set up - I hold one release in each hand.

 

The shot showing one housing shows evidence of one of my "issues". It had rained a bit in the days before the shot. The water level had been much higher leaving a layer of sediment on the leaves of the trees - mostly willows (genus Salix). Rain brings in glacial sediments from the main stem of the creek into the side channel creek seen in the pic. There is still some fine sediment even though the water looks pretty clear. So almost all shots from here have very fine backscatter. Mature Sockeye Salmon do not have too much of a reflection issue except for their mouths. I have had blown highlights in the mouth and had to throw those pix out. Dolly Varden are partially reflective even when mature. This is most evident when the female turns on her side to make a dig as seen in the posted pic. They also have white in the mouth. I have had blown highlights from both of these reflective areas and have had to toss a few pix. This is not my only issue with shooting local Dolly Varden spawning. My local spawning area is just downstream from a small waterfall, which is barely visible in the background of the shot. It generates a lot of tiny air bubbles which you can see in the pic. They can also get stuck on the port due to surface tension. Bubbles are especially bad when there is a lot of flow such as after rain, which is when they spawn. This site does not have glacial sediments, instead the water is stained like tea. Staining is worse after a rain as well. Staining darkens the water a bit, about a stop less light transmittance. Lots of problems that need some adaptation in order to get the shots. I have worked them out (more or less) with trial and error. BTW, the fallen tree I described in my earlier post is what is seen in the middle of this pic. I was standing behind where the male is and the spawning was upstream of the tree in the dark part of this pic.

 

I hope this helps!

Tom

 

 

post-3540-0-71573700-1446841120_thumb.jpg

post-3540-0-49443700-1446841146_thumb.jpg

post-3540-0-28036700-1446841156_thumb.jpg

  • Like 1

Share this post


Link to post
Share on other sites

The issue with the GoPro image above is that the subject (salmon) is too close to the GoPro and the camera cannot focus on it.

 

This can be corrected by using a correction lens (http://wetpixel.com/articles/field-review-inon-gopro-accessory-system) or a camera/lens combination that provides close focusing ability (there are many!)

Share this post


Link to post
Share on other sites

HI Mathis

 

you have a couple of problems:

 

Firstly you are too close to the fish to get a sharp image as the minimum focus on the gopro's is around 12 inches or more. This can be partially addressed by fitting a close-up lens like a +10 diopter to the front of the camera and perhaps using the camera in medium view mode. This diopter moves the whole focal range a little closer whilst still maintaining great depth of field. It's not the same as a macro lens, which will get you closer but has a very limited depth of field and is less useful in this situation. You will need an adaptor plate to fit the front of the camera with a 58mm thread to take a slightly wider diopter.

 

Secondly you have no control of shutter speed with the gopro so if you don't have great light, the shutter speed won't be fast enough to freeze the action. The white streak in the centre of your image shows you the shutter was open too long and allowed the particle/bubble in the water to travel and blur in the image (also the fish). I don't know which Hero 4 you have but you might have an option to shoot video in 4K and then take a still from that, shooting video will force the camera to expose 25-30 frames per second (and hence shorter shutter times) rather than a longer exposure for a photograph, however you will still need good light, which I think is your underlying problem. From the iimage you posted it appears you are working at some depth so light is going to be your biggest issue.

Share this post


Link to post
Share on other sites

I am posting here a couple of "technique" shots that may better explain what I do. Both cameras in the shot showing two housings are set up with a single Seacam strobe positioned at 12 o'clock and above the water. These are my primary light sources. I also have some available light pix done with no strobe - most of the recent (last few years) shots on my site showing schools of salmon are done this way. High ISO is generally needed. Available light shots generally have less issues with reflective parts of salmon bodies. Even when it is sunny such as the shot with two housings, the light is very patchy in forest settings so strobes are generally needed on the spawning grounds. There is no sun at all hitting the ground late in the season, like right now (November), at spawning grounds located in valleys due to very low sun angle.

 

You can see the remote cord from the housing in top of the frame (two housing pic) - it ran over to where I was standing to shoot the technique shot. From this vantage point I was able to shoot both cameras. Note that there is a salmon pair in front of each housing. These salmon are about 0.5m long (for scale). I have to anticipate where a salmon is going to be in the next second when I push the button. Two housings (different models, one being older or "obsolete") is a convenient set up - I hold one release in each hand.

 

The shot showing one housing shows evidence of one of my "issues". It had rained a bit in the days before the shot. The water level had been much higher leaving a layer of sediment on the leaves of the trees - mostly willows (genus Salix). Rain brings in glacial sediments from the main stem of the creek into the side channel creek seen in the pic. There is still some fine sediment even though the water looks pretty clear. So almost all shots from here have very fine backscatter. Mature Sockeye Salmon do not have too much of a reflection issue except for their mouths. I have had blown highlights in the mouth and had to throw those pix out. Dolly Varden are partially reflective even when mature. This is most evident when the female turns on her side to make a dig as seen in the posted pic. They also have white in the mouth. I have had blown highlights from both of these reflective areas and have had to toss a few pix. This is not my only issue with shooting local Dolly Varden spawning. My local spawning area is just downstream from a small waterfall, which is barely visible in the background of the shot. It generates a lot of tiny air bubbles which you can see in the pic. They can also get stuck on the port due to surface tension. Bubbles are especially bad when there is a lot of flow such as after rain, which is when they spawn. This site does not have glacial sediments, instead the water is stained like tea. Staining is worse after a rain as well. Staining darkens the water a bit, about a stop less light transmittance. Lots of problems that need some adaptation in order to get the shots. I have worked them out (more or less) with trial and error. BTW, the fallen tree I described in my earlier post is what is seen in the middle of this pic. I was standing behind where the male is and the spawning was upstream of the tree in the dark part of this pic.

 

I hope this helps!

Tom

 

 

where did you take this pic at?

 

Sent from my SM-G900P using Tapatalk

Share this post


Link to post
Share on other sites

where did you take this pic at?

 

Sent from my SM-G900P using Tapatalk

 

In Alaska!

Share this post


Link to post
Share on other sites
Sign in to follow this  

Sponsors

Advertisements



×
×
  • Create New...