Cannon Fire - burnt highlights

[tandeejay]

Hi

This post is about the "artillery" type of cannon, not the photographic type :)...

At regular intervals throughout the year, the Fort Lytton National park fires their 64lb cannon. I've been to the "History Alive" event there for the past 2 years and tried to photograph the cannon firing, with this result from last years event (previous year I didn't take the tripod and had lots of camera motion):

https://farm4.staticflickr.com/3901/14380121054_ab577ddcb9_c.jpg (https://flic.kr/p/nUHSZC)
Cannon shot 2014 (https://flic.kr/p/nUHSZC) by John & Tracy Blackburn (https://www.flickr.com/people/99216317@N08/), on Flickr

Apart from the framing, which I aim to improve on this year, I've noticed that often with cannon shots, parts of the blast are often blown out.

Does anyone have any good suggestions on how to reduce the blown highlights on the blast without under exposing the rest of the image?

Regards,
John

[tduell]

Hi

Does anyone have any good suggestions on how to reduce the blown highlights on the blast without under exposing the rest of the image?

Regards,
John

Well I would be guessing a bit, not having shot a subject quite like this before.
I would think bracketed exposures are probably out of the question due to dynamics of the subject.
I would try shooting raw, and setting (say) 1/2 stop underexposure, then bring out the detail in the underexposed parts of the image in post processing.
Just my two bob's worth.

Cheers,
Terry

[tandeejay]

Thanks for the suggestion Terry. Unfortunately I probably won't have a camera capable of shooting in raw until after the History Alive event in June this year :(.

I might experiment with exposure compensation anyway, plus I'll try and do better with the framing :)

[ameerat42]

... Unfortunately I probably won't have a camera capable of shooting in raw until after the History Alive event in June this year...

I might experiment with exposure compensation anyway, plus I'll try and do better with the framing :)

Yep. Nothing much until you got a wide DR raw file. If trying this again, a bit of underexposure might help, but I think there is a fair
range of intensity in that highlight. Anyway, that amount of blown high for this subject doesn't (IMO) matter much. It looks
natural enough. Mainly here, do better with the framing next time, as you said.

The camera did pretty well else.
Am.

[tandeejay]

The camera did pretty well else.

I had the camera on a tripod, set on the 60 frame burst which takes 60 frames at 1megpix in 1 second. That was the 1st frame of that burst. The cannon at the front was fired 15 seconds later. I'd set up the framing for the front cannon, and decided to photograph the rear cannon as a test, and then found that the camera takes 15 seconds before it is ready, so I missed the front cannon... just got 60 frames of a cloud of smoke...

In hindsight, my framing for the front cannon would have been to tight too... I greatly underestimated the size of the blast... I worked out I need to allow at least 3 times the cannon barrel lenght in front of the cannon.

John

[swifty]

Just a few things to consider, John.
Was it necessary to tripod mount the camera? It seems to me you could have more flexibility just hand holding and shooting without loosing any support but gaining more freedom to recompose quicker.
Is the 60 frames a bit overkill. I would probably have used a few short bursts anticipating and following each cannon. That way you could have used a higher resolution and wouldn't have clogged up the camera buffer so quickly. Unless the cannon sequence were unpredictable?

In regards to the original question of blown highlights, I agree with the other posters in that it's just the lack of DR to cover the entire spectrum in the scene. Your only alternative would be to underexpose and post process the shot. Bear in mind there are usually more detail in the shadows than you think and a little noise won't ruin the shot.
Since your camera doesn't shoot RAW, use the flattest in camera JPEG setting. Increasing contrast pushes both highlights and shadows to the extremities and you're more likely to get blocked shadows and blown highlights. The flat JPEG can then be post processed to add 'pop' and contrast without blowing highlights provided you captured the DR to begin with. Use something like a curve adjustment in post processing.
Use either exposure compensation of preset your exposure manually to underexposed a little bit. You'll have to estimate how much to underexpose but since its brights and sunny and you're at base ISO you can probably get away with up to 2 stops.

- - - Updated - - -

In regards to the comment on camera motion without a tripod, your Flickr exif indicated 1/500 shutter speed. So in the lighting conditions this year, hand holding shouldn't be an issue unless you were using an extremely zoomed in focal length. Again your exif indicates equiv of 144mm (25.8mm actual) focal length so that combined with 1/500 shutter speed is perfectly hand hold able.
Perhaps in previous years the conditions were darker and you were more zoomed in?

[tandeejay]

This was my shot from the previous year. I wasn't zoomed in as far, but had a much slower shutter speed at 1/60th

https://farm4.staticflickr.com/3805/11524793856_d31c1aab91_c.jpg (https://flic.kr/p/iypzGm)
Cannon Shot (https://flic.kr/p/iypzGm) by John & Tracy Blackburn (https://www.flickr.com/people/99216317@N08/), on Flickr

The timing of the firing of each cannon was announced by the guy in the black outfit with the white helmet, so I only had the visual cue of his arm dropping. So add to that my reaction time, plus the camera's shutter lag, I was trying to anticipate to have the cannon go off sometime in the middle of the 60 frame burst. The previous year, I managed to get the shutter open slightly earlier in the shot, but was not using a burst mode. Ideally, I would have liked to have the 1st frame of that burst just as the hot gases and smoke were just starting to emerge from the end of the barrel, then I would have been able to find the shot with the best part of the blast... however, that photo I first posted in this thread was the 1st frame of the burst, and the best one. Timing is an absolute pain. Would a faster SD card have enabled the camera to become ready sooner? My current card is a 30mb/s card.

I like the idea of going to the History Alive event this year without lugging around the tripod. I'll have to experiment with releasing the shutter earler to see if that stops the 60fps burst earlier, and see if that changes how long it takes before it is ready again. The higher res burst mode does about 7 fps for up to 7 frames. Would 7 frames in 1 second be able to get at least 1 good frame? Exactly how long does a cannon blast take? with the high res burst, the camera still takes its time before it is ready for the next shot...

[swifty]

On this shot, with a crop factor of 4.5X on a different camera (Kodak P880) and 25mm actual focal length, you're at 117mm equiv focal length so 1/60th shutter speed wouldn't be adequate for hand holding. The old 'rule' is shutter speed should be at least your 1/focal length but today with the ever increasing resolution of sensors, you might want to aim for double that. Note this is just a general rule and factors like technique and individual ability will influence this.
You can see that the day was more overcast (less light) and your position has changed.

The thing with 60fps at 1MP is that even 1080P video gives you over 2MP resolution. Of course there are other video limitations but there are probably better ways to capture that cannon burst.

In regards to your camera's lag. Is it the shutter lag or AF lag? Shutter lag is typically very low if you're already pre-focused.
Which leaves only the reaction time. So I guess it comes down to how the firing sequence is conducted and how easy it is to anticipate.

With your current Nikon P520, this is what I would try.
Set your exposure manually when it is close to firing. Hopefully shooting conditions don't change rapidly (eg. fast moving clouds where its full sun one second and full shade the next).
Set exposure to under expose maybe 1 or slightly more stops if you're at base ISO.
Check that the shutter speed is adequate. Check your tests shots to see if they are sharp at the focal length of your framing and that the rest of the scene is not too dark (you can boost the shadows a little in post processing).
Set to the max continuos shooting speed at the max resolution, 7fps in your case.
When the conductor starts his sequence, pre-focus with a half press and hold that position.
Then try to time the start of your shooting burst at the start (or just before) the cannon fire.
This way you have less files to sort through and the one you get is of higher quality.

I have to say I don't use continuous shooting much and only do it if I'm trying to capture a sequence. Depending on how many times the cannon fires, I'd be tempted to use just a single frame and time each cannon fire with a single frame each time. I just find I have better success doing it this way for capturing decisive moments, rather than bursts but that's just me.
Stick to continuous shooting if there aren't many cannon fires (ie. you don't get many goes at timing the shutter) or things are less predictable and difficult to time.

[fillum]

I don't think there is any way that you can eliminate the blown areas of the blast and still have the rest of the image with reasonable exposure - I think the blast is just too bright. Your first image here is 3 stops less exposure than the second one (post #7) and still shows a lot of blown highlights (although the coloured parts are a bit more defined). However like Am said, I'm not sure that it matters here, it's something we almost expect to see. I think the blown highlights actually help give a sense of the intensity of the blast.

The blown highlights do work much better in the second image where they are 'contained' within the yellow / red perimeter of the flash.

Having said that, if you do feel that the highlights are a problem it's a pretty easy job in post to paint in a bit of colour and texture...

115322




Good luck with it...

[tandeejay]

Shutter lag I'm talking about is the one after the focus is pre-done. On reviews of the P520 I've read, the shutter lag is about 500ms and I'd have to agree with that based on my experience. If I have to wait for the camera to focus, depending on the conditions and amount of zoom, the autofocus can take several seconds as the lens hunts for something to focus on. I was well aware of this, so I had pre-focused when I took this.

The 2 cannons are fired once in the morning about 15 seconds apart, and once again in the afternoon, again about 15 seconds apart. So don't get many opportunities during the event. That gap of only 15 seconds was the killer for me for the second cannon, as the camera took most of the 15 seconds processing the shots from the first cannon, so I didn't have time to press the shutter button in time for the second cannon.

- - - Updated - - -

Thanks for all the great tips...

Phil, seeing your edit, I'm inclined to agree that the blown highlights isn't really a problem.

Comparing the exif data from 2013 with the one from 2014, it looks like the amount of cloud cover has potentially dropped the shutter speed significantly.

I'll have to keep that in mind this year, to ensure that I keep the shutter speed up high enough, and maybe ask the conductor to increase the time between the 2 cannons firing to compensate for the camera taking so long to get ready for the next shot...:lol:

[swifty]

Oh wow, 0.5s shutter lag. That is quite bad.
I guess you would have to start your burst sequence very early.
Try the 7fps at max resolution for about 3 seconds and see how long it takes to clear that buffer.
If its quite quick then you should have enough time for the second cannon fire.
Experiment with 4 seconds, or 5 seconds and see what gives you the most burst shooting time whilst retaining enough time to recompose, focus and shoot the second cannon fire.

[tandeejay]

Even the 7 fps mode cuts out after 7 frames, then it previews each frame for about 2 seconds. Most annoying especially since one of the 1st things I did when I got the camera was to turn off the preview mode.

But I like your suggestion of working out what gives me the most frames with the shortest delay. I'll definitely be having a go to see what I can do. Maybe even not pushing up to max resolution of 18mp. Even 10 would be a lot better than 1.

[swifty]

Sorry John, I may have given u unrealistic expectations as I haven't used a P&S for so long that I'm making a lot of assumptions about the camera ability that aren't too helpful.
At the end of the day Am and Fillum may be right in saying the DR of the blast may be just beyond the sensor's ability so you probably shouldn't worry too much about it.
Cheers.

[tandeejay]

Thanks for your advice. Much appreciated.

[tandeejay]

When thinking about my future plans to get a dslr, in a situation like this, which is more important: Shutter lag, or picture quality? According to the snapsort website, the 70d has a shorter shutter lag than the D7100 (75ms vs 251ms) but the D7100 has better picture quality having a greater dynamic range (11.5EV vs 13.1EV)

[ameerat42]

Tendeejay. Put like that, shutter lag vs IQ, I'd go for the latter.

Of course, if the IQ difference is minimal - that is, almost negligible - then OK, but what sounds better:

"Here's one good shot at least."
Or
"Here is a bunch of carppy shots. (You'll note there's hardly any shutter lag:rolleyes:)"

Am.

[swifty]

Hi John,
try these imaging resource links:
http://www.imaging-resource.com/PRODS/nikon-d7100/nikon-d7100A6.HTM[
http://www.imaging-resource.com/PRODS/canon-70d/canon-70dA6.HTM
Click the Performance tab and have a look at the AF timing.
I think Snapsort is mixing things up a little because basically shutter lag is imperceptible for pretty much all modern DSLRs.
According to Imaging Resource which is pretty trustworthy, pre-focused shutter lag is 54ms vs 59ms.
Put it this way, I think 251ms (1/4s) would be a completely unreasonable amount of lag and though I haven't used a D7100, I can't imagine it would be that far out of step with other DSLRs I've used and so I'm inclined to believe the imaging resource numbers.

[tandeejay]

Thanks for that Swifty. From the looks of that site, I see the D7100 has a lag of 248ms when pre-focused in live view mode, but with the optical view finder the lag is 54ms. The Canon is 59ms for both live view and optical view finder modes. I guess thats probably due to the canon introducing their dual pixel sensor.

I like that site you've linked. It provides far more detailed analysis than snapsort.

So, I was leaning towards the canon due to the shutter lag difference, but now that argument has just brought the Nikon's back onto my radar as possible future DSLR's for me...

Going to have to study the data on that site more....