Now, this seriously has me beaten, suggestions as to what is happening here please?

[Ed Bray]

Okay, this is weird. I was processing another of my recent South Devon Railway images and I realised that the front of the locomotive is not as sharp as I would have expected.

The image was shot on 5x7 film using a 210mm Apo-Symmar lens at 1/250th sec at f11 on HP5+ film, now I would have expected a sharper image than that recorded, but when I started working on the image in Photoshop CS5 I noticed that the number on the front of the engine was a distinct double image, but nowhere else showed a similar double image, this is not a blurred image such as would normally be expected from too slow a shutter speed. The film was processed in Pyrocat HD for 8 ins at 24 degrees C with a TF-3 alkaline fix. The image was scanned on my Epson V750 scanner with the better scanning film holder. I have checked the negative on a lightbox with an 8x lupe and it is definitely a double image on the negative!

Full Image: focus was on the sleeper 8 up from the trackside marker as can be seen in the third image:

http://farm8.staticflickr.com/7110/7871269590_12719b1444_b.jpg (http://www.flickr.com/photos/edbray/7871269590/)
SDR: 3205 GWR Collett 0-6-0 5x7 (http://www.flickr.com/photos/edbray/7871269590/) by Ed Bray (http://www.flickr.com/people/edbray/), on Flickr

Second image, the locomotive's number at 100% crop

http://i332.photobucket.com/albums/m338/EdBray/Doubleimage.jpg

Third image: Focus point on 8th sleeper from tackside marker, sleeper marked 631.

http://i332.photobucket.com/albums/m338/EdBray/focuspoint.jpg

[Erik Larsen]

Maybe the train hit an in opportune bump as you tripped the shutter? Just a guess.

[Ed Bray]

Maybe the train hit an in opportune bump as you tripped the shutter? Just a guess.

Thanks Erik, but wouldn't that have given a blurred transition between the two sets of numbers?

[cowanw]

Wouldn't such a movement pause at each extreme causing a increased exposure at the ends? Such a bump would show the worst at the part moving at the closest and greatest angle of movement to you. On the other hand, other highlights on the front don't seem to have moved. perhaps the plaque was loose?

[Kirk Gittings]

What was the temperature/humidity conditions? It almost looks like the negative popped from like going from a cold cooler to a hot exterior a. However I would have expected a larger (not so small and defined) area to show the shift.

[Ed Bray]

Thanks both.


Wouldn't such a movement pause at each extreme causing a increased exposure at the ends? Such a bump would show the worst at the part moving at the closest and greatest angle of movement to you. On the other hand, other highlights on the front don't seem to have moved. perhaps the plaque was loose?

That's a possibility


What was the temperature/humidity conditions? It almost looks like the negative popped from like going from a cold cooler to a hot exterior a. However I would have expected a larger (not so small and defined) area to show the shift.

The temperature was about 20 degrees C, not too hot or too cold, the film holder was in position in the camera for about 30 minutes whilst I waited for the train to arrive, my blackjacket was draped over the camera during this time.

[E. von Hoegh]

Maybe the train hit an in opportune bump as you tripped the shutter? Just a guess.

The locomotive is rocking on an uneven section of track. The contrast of the black placque and white numbers makes it most apparent on the # placque.

[Bill_1856]

Repainted sign.

[Brian C. Miller]

You have a problem with your shutter. Above the sign, look at the lamp. It also exhibits the same problem. You may have a shutter leaf that is sticky. Get a CLA (clean, lube, adjust) done for your lens.

[IanG]

Looks normal for a shot of a moving engine, it's moviment and 1/250th isn't always fast enough to stop it.

Ian

[Ed Bray]

Thanks all, Brian, if it was the shutter would not the whole image suffer a problem?

[Ed Bray]

Your lens isn't good at math?

Sorry, that's gone right over my head?

[E. von Hoegh]

You have a problem with your shutter. Above the sign, look at the lamp. It also exhibits the same problem. You may have a shutter leaf that is sticky. Get a CLA (clean, lube, adjust) done for your lens.

Shutters don't work that way. The leaves are actuated in unison by a ring; if one leaf sticks, the rest of the leaves are retarded also.

[E. von Hoegh]

Your lens isn't good at math?

It's a dysnomic lens.

[DennisD]

Something's going on with the number plaque.

Looks like it might have been moving at the moment of exposure, despite the seemingly high shutter speed. There's no indication of other features shifting so noticeably in relation.

The difference in the brightness of the duplicate numbers on the sign is bothersome. The white letters seem to stand by themselves and within them alone there appears to be some vibration or movement. This suggests that the darker letters might be remnant of former lettering. Seems odd, however, that a number plaque would be treated this way.

[Ed Bray]

I am so perplexed by this I am seriously considering visiting the main station area in the next couple of days to check it out.

[sully75]

I see a double image on the smokestack too.

Not sure exactly why there's a double image, perhaps the train bounced. But there is a lot of blur in general. There is more blur in the front than in the rear, which would make sense because that it closer to the lens and going to have more apparent motion.

[Sevo]

As others noted, it seems to be a motion issue, in that nothing similar is visible in the background. But we'd need to see other bits of the image at similar enlargement to make sure.

[Ed Bray]

It's not the blur that I am worried about, I expected there to be some blurring. It is the double registration of the numbers that has me beat, again, had there been some blurring of the numbers that would have been okay but it is almost like I have taken the shot and then fired a flash (parts of a second later) which has then registered a second sharp rendering of the numbers (I haven't by the way, I didn't have a flash with me).

[Brian C. Miller]

Shutters don't work that way. The leaves are actuated in unison by a ring; if one leaf sticks, the rest of the leaves are retarded also.

I still think the shutter is at fault. The ground and ties are sharp, and the train exhibits a double-image problem. The lamp and the sign are the most obvious examples of that. The train did not bounce, as the photograph shows that the tracks are smooth and normal. This is a steam locomotive, and would turn something like a bolt on the tracks into foil. If the train jumped far enough in "1/250th" of a second to produce that kind of a blur, it would have left the tracks completely.

Therefore, the shutter didn't actually fire at 1/250th of a second, and there's a problem with it. It just needs a CLA, and all's well.

[Ed Bray]

As others noted, it seems to be a motion issue, in that nothing similar is visible in the background. But we'd need to see other bits of the image at similar enlargement to make sure.

Click on the flickr link and then select full size, it's there in the size scanned 1200dpi

[E. von Hoegh]

I still think the shutter is at fault. The ground and ties are sharp, and the train exhibits a double-image problem. The lamp and the sign are the most obvious examples of that. The train did not bounce, as the photograph shows that the tracks are smooth and normal. This is a steam locomotive, and would turn something like a bolt on the tracks into foil. If the train jumped far enough in "1/250th" of a second to produce that kind of a blur, it would have left the tracks completely.

Therefore, the shutter didn't actually fire at 1/250th of a second, and there's a problem with it. It just needs a CLA, and all's well.


How can you see the tracks that the train is actually on? It's entirely possible that the engine has rocked.

[Peter Mounier]

It's clearly (or not) motion blur as others have suggested.

http://www.morrobaygiclee.com/TrainBlur.jpg

Peter

[Sevo]

Therefore, the shutter didn't actually fire at 1/250th of a second, and there's a problem with it.

If any, it might be a matter of shutter bounce. 1/250s is not fast enough to fully freeze a moving train at such a close angle - and a shutter speed error minor enough not to affect the exposure to a critical degree won't change that one way or another. The real question is why the motion shows up as a double image rather than a continuous wipe - both a rocking engine and shutter bounce could cause that (plus other odd causes like another photographer's flash triggering right at the end of the exposure).

And after seeing the last enlargements with a triplicate edge, it seems very much like a rocking engine - that pattern is a bit too complex with too much transit between the stages to be shutter bounce.

[Steven Tribe]

A double image showing up in high contrast suggests to me a filter orientation/lens element problem.

[Ed Bray]

Thanks all for your thoughts so far.

I can tell you that there was no one else there, so no additional flash exposure although that would have been my first thought too if I did not know better.

[Peter Mounier]

The train rocks but the motion on the highlights goes by to fast to show a blur during the rocking motion. It shows, and is recorded on the film when the highlights stop rocking for a split second to reverse direction.

Peter

[William Whitaker]

Perhaps you have a tiny pinhole in the bellows that's casting a secondary image slightly out of registration.

[Ed Bray]

A double image showing up in high contrast suggests to me a filter orientation/lens element problem.

No filters and other images taken with this lens on the same day show no similar problems:

http://farm8.staticflickr.com/7268/7859394070_f673bc426e_b.jpg (http://www.flickr.com/photos/edbray/7859394070/)
6430 Pannier Tank & Auto Coach W225 on 5x7 (http://www.flickr.com/photos/edbray/7859394070/) by Ed Bray (http://www.flickr.com/people/edbray/), on Flickr

http://farm9.staticflickr.com/8447/7859408486_4b1d6227b4_b.jpg (http://www.flickr.com/photos/edbray/7859408486/)
Auto Coach W225 powered by 6430 on 5x7 (http://www.flickr.com/photos/edbray/7859408486/) by Ed Bray (http://www.flickr.com/people/edbray/), on Flickr

[Sevo]

A double image showing up in high contrast suggests to me a filter orientation/lens element problem.

It would - but it does only show up in moving parts of the subject, not in the equally high contrast of the branches against the sky or the black/white signal in the rear.

[Mike Anderson]

The locomotive is rocking on an uneven section of track. The contrast of the black placque and white numbers makes it most apparent on the # placque.


The train rocks but the motion on the highlights goes by to fast to show a blur during the rocking motion. It shows, and is recorded on the film when the highlights stop rocking for a split second to reverse direction.


I vote for this theory. In image #2 I can't see similar double registration of the high-light dots in the foliage in the upper right.

[sully75]

No filters and other images taken with this lens on the same day show no similar problems:

http://farm8.staticflickr.com/7268/7859394070_f673bc426e_b.jpg (http://www.flickr.com/photos/edbray/7859394070/)
6430 Pannier Tank & Auto Coach W225 on 5x7 (http://www.flickr.com/photos/edbray/7859394070/) by Ed Bray (http://www.flickr.com/people/edbray/), on Flickr

http://farm9.staticflickr.com/8447/7859408486_4b1d6227b4_b.jpg (http://www.flickr.com/photos/edbray/7859408486/)
Auto Coach W225 powered by 6430 on 5x7 (http://www.flickr.com/photos/edbray/7859408486/) by Ed Bray (http://www.flickr.com/people/edbray/), on Flickr

Nice pictures, for the record!

[Cletus]

+1 for repainted sign / numbers. When I look closely at the number plate in your crop image, the curves on the bottom, partial set of numbers don't seem to exactly match the curves on the top numbers. I can't think of any technical reason, or lens flaw that would only allow this one small section to be double imaged and nothing else directly adjacent to it.

That's my guess...

[Cletus]

That, or the more likely possibility that it IS a double exposure from a "high frequency" bump, or some other sudden movement of the train during the exposure and because of the higher EV value of the white lettering, the darker areas of the train did not register on the film.

Think cumulative exposures or multiple flash trips to get F128 or something. If there were any movement of the lit subject between exposures, where other areas of the scene of low value that weren't lit, you'd only show double exposure in the areas of highest brightness. In the dark areas where there wasn't sufficient illumination, the double exposure wouldn't appear as those areas would not register on the film. I guess I said the same thing twice there, but that's my story and I'm stickin' to it!

Kirk Giddings could probably comment with some direct experience with this - architectural photogs use this method all the time.

[Bill_1856]

Ed, how about the 85A sign below the 3205? Does it look okay or the same?

[E. von Hoegh]

+1 for repainted sign / numbers. When I look closely at the number plate in your crop image, the curves on the bottom, partial set of numbers don't seem to exactly match the curves on the top numbers. I can't think of any technical reason, or lens flaw that would only allow this one small section to be double imaged and nothing else directly adjacent to it.

That's my guess...

Those number plates are often cast with raised numerals. If you look closely, you will see that the strong image of the numerals, while blurred, is centered on the plate and the ghost numbers are not centered.

[Ed Bray]

Nice pictures, for the record!

Thanks.

Bit of luck having unlimited access to this spot which is off the beaten track and down a Private road, the only think the other side of the railway lines (unmanned level crossing) is a field and a building with a Borehole under it of which comes under my 'Remit' as the Asset Manager for the area. I did spend a nice morning down there on Saturday by myself for a few hours.

[Greg Miller]

It's clearly (or not) motion blur as others have suggested.

http://www.morrobaygiclee.com/TrainBlur.jpg

Peter

I agree. there is ghosting in several areas, and only on objects that move. The ghosting varies in amount, which would be consistent with a moving train with varying degrees of relative motion, as well as varying degrees of rocking of the train cars. Other areas of high contrast, such as along the non-moving train tracks, show no ghosting.

In a rocking motion the car could stable during the start of the exposure, then rock sharply, towards the end of the exposure, causing the type of ghosting exhibited here.

[DennisD]

There is definite blur or double imaging especially higher up on the locomotive (Peter's post #24) which would support the "rocking" theory.

However, looking closely at the number plaque, there is clearly motion just within the brighter white numerals alone --- while the darker numerals, appear like a "shadow image" and look relatively sharp. It looks like a double exposure, one with noticeable movement and another without, taken a millisecond apart. That certainly supports some kind of (almost) simultaneous "light flash" theory.

Looking at the photos in post 24, I take back my previous thought that the sign was repainted. The "shadow image" is very clear at the top of the engine around the headlight, giving the same effect as on the number plaque below.

[Kirk Gittings]

That, or the more likely possibility that it IS a double exposure from a "high frequency" bump, or some other sudden movement of the train during the exposure and because of the higher EV value of the white lettering, the darker areas of the train did not register on the film.

Think cumulative exposures or multiple flash trips to get F128 or something. If there were any movement of the lit subject between exposures, where other areas of the scene of low value that weren't lit, you'd only show double exposure in the areas of highest brightness. In the dark areas where there wasn't sufficient illumination, the double exposure wouldn't appear as those areas would not register on the film. I guess I said the same thing twice there, but that's my story and I'm stickin' to it!

Kirk Giddings could probably comment with some direct experience with this - architectural photogs use this method all the time.

While that's true (though when doing multiple pops to build flash power each exposure is identical) I can't imagine how it would of occurred in this circumstance. I'm thinking instead maybe a brief glimpse of an alternate universe? :)

[E. von Hoegh]

While that's true (though when doing multiple pops to build flash power each exposure is identical) I can't imagine how it would of occurred in this circumstance. I'm thinking instead maybe a brief glimpse of an alternate universe? :)

Aliens. The little guys love to mess with photographers. I guess it's better than getting probed...

[Struan Gray]

Look at the track. There seems to be a join between two rail segments under the tender, with a distinct distortion to the rail around it. The front of the tender shows some doubling - but not as much as the engine - but the rear does not. Look, for example, at how the doubling of the rivet heads on the side of the tender dies away as you move back.

I suspect the engine was given a lurch as it passed over the rail joint, and it was still swaying a bit in response to that and/or any slight misalignment of the two rail segments. The front of the tender is just starting the same sort of response, while the rear is only feeling a bit of torque from the motion of the front.

I'm not an expert on the suspension of steam trains, but the idea seems consistent with the lurch-rattle-sway typical of the steamies I have travelled on where they have used the older discontinuous track - which had more pronounced discontinuities than modern continuously-welded track.

[DennisD]

Thinking about my previous post #40 and just now reading Greg Miller's post #39 ....

Can someone explain why a "ghost" or "shadow" image would be present under these conditions as compared to a continuous blur, given that this is a single exposure image.

[Gem Singer]

Shutter bounce.

At higher speeds (in this case, a 250th), when the shutter reaches the end of it's cycle, it rebounds and briefly opens again causing a second image.

I've read about this phenomenon but have never experienced it myself.

Could possibly be the cause of the ghost image.

[Greg Miller]

There is definite blur or double imaging especially higher up on the locomotive (Peter's post #24) which would support the "rocking" theory.

However, looking closely at the number plaque, there is clearly motion just within the brighter white numerals alone --- while the darker numerals, appear like a "shadow image" and look relatively sharp. It looks like a double exposure, one with noticeable movement and another without, taken a millisecond apart. That certainly supports some kind of (almost) simultaneous "light flash" theory.

Looking at the photos in post 24, I take back my previous thought that the sign was repainted. The "shadow image" is very clear at the top of the engine around the headlight, giving the same effect as on the number plaque below.

I don't think the double exposure theory hold up. Looking at Peter's samples, the ghosting varies in size and direction. With a double exposure, the ghosting should be consistent in direction, and pretty consistent in size.

[Ed Bray]

It certainly wasn't a double exposure, it was a common or garden 210mm Apo-Symmar in a Copal 1 shutter, this has to be recocked to allow a second exposure IE it was not in a press shutter.

There was no one else around, the camera was about 60ft away from the train at the moment of exposure and levelled on a heavy tripod on a Tarmac Road.

There were no secondary light sources just the sun which was covered by a cloud otherwise the exposure would have been nearer 1/400 @ f16 (using the sunny 16 rule) rather than the 1/250 @ f11 (and I suspect the negative is a little underexposed too) giving at least a 2 stop reduction on sunny 16 if not slightly more.

Other images from that day do not exhibit the same attributes albeit they were probably moving a bit slower even though I tripped the shutter when they reached a similar point (AutoCoach in control above) and used the same lens and almost the same exposure f13 instead of f11.

[Struan Gray]

Ed. The AutoCoach shows the same sort of doubling. Look at the wire/rope to the bell, the rivets around the front roof line, the handle of the lamp, the two small triangular white doodahs on the front, and various other places around the front of the carriage. The logo on the side is fine, but then the AutoCoach has only just started to move across the rail join.

Doubling is typical of a oscillation, which you would expect to be small or non-existent for the parts of a train that have not yet passed the rail join - which acts as an impulse to set the joggling off. Motion blur gives a smooth blur (a bright highlight gets turned into a line) and if modulated by shutter bounce it should be present along the whole length of the train, even if to a diminished extent because of perspective and the reduced relative motion for objects which are farther away.

[Ed Bray]

Ed. The AutoCoach shows the same sort of doubling. Look at the wire/rope to the bell, the rivets around the front roof line, the handle of the lamp, the two small triangular white doodahs on the front, and various other places around the front of the carriage. The logo on the side is fine, but then the AutoCoach has only just started to move across the rail join.

Doubling is typical of a oscillation, which you would expect to be small or non-existent for the parts of a train that have not yet passed the rail join - which acts as an impulse to set the joggling off. Motion blur gives a smooth blur (a bright highlight gets turned into a line) and if modulated by shutter bounce it should be present along the whole length of the train, even if to a diminished extent because of perspective and the reduced relative motion for objects which are farther away.

I am a lot happier with the bluring on the Autocoach image as there is no distinct double image like on the Collett, the front of the cab shows motion blur which is to be expcted due to the proximity and the speed but it was the distinct 'double image' that has me curious. Thanks for your thoughts though.

[Greg Miller]

Can someone explain why a "ghost" or "shadow" image would be present under these conditions as compared to a continuous blur, given that this is a single exposure image.


In a rocking motion the car could stable during the start of the exposure, then rock sharply, towards the end of the exposure, causing the type of ghosting exhibited here.

[Struan Gray]

I am a lot happier with the bluring on the Autocoach image as there is no distinct double image like on the Collett, the front of the cab shows motion blur which is to be expcted due to the proximity and the speed but it was the distinct 'double image' that has me curious. Thanks for your thoughts though.

I'm not trying to be argumentative (honest). But I see doubling.

79529

[Greg Miller]

Blur can appear to be doubling if the blur happens for only a short period of the exposure. Such as if the train rocks suddenly.

[Ed Bray]

I agree Struan but there is a blurring in between which makes it less defined than the numbers on the Collett.

[Greg Miller]

I'm thinking that each part of the train will have its own motion. So it is very possible that some parts of the train will exhibit no problem,. some parts will exhibit doubling, and others will exhibit doubling with some blur in between. Just depends on how that parts of the train is moving/rocking. I sharp rocking will look more like doubling; and less sharp rocking will show more blur. The fact that the tracks don;t show any of this leads to explaining what appears to doubling & blurring by how each part of the train is being affected by the tracks. I konw from having ridden trains that there is plenty of motion; sometimes is is sharp and abrupt, and other times is is more of a slow roll.

[Ed Bray]

Thanks all for your comments, this pretty much does it for me:

It is quite common for leaf shutters to bounce at their highest speed, and this circuit will detect the bounce as a second pair of peaks shortly after the main exposure. This is caused by the shutter blades re-opening briefly as the last of the shutter spring's energy is released. [Note that this bounce may not be recorded if you use a slightly misaligned laser instead of a torch as a light source]. Strictly speaking, the bounce time should be added to the main opening time to get a total time for the shutter. In practice, if the bounce time is the equivalent of 5 stops less than the main exposure, which equates to about 1/30th of the main time, no ghost image will be recorded on the film. If the bounce time is excessive, say 1/8th or 1/4 of the main time, which is only 3 or 2 stops down from the normal exposure, this will make a significant difference and could leave a ghost image. Since most people avoid using the highest shutter speed, this may not be a problem for you, and in any case, a service may not help. What's most important is that you know about it!

From: p://www.mraggett.co.uk/shuttertes...ttertester.htm

[E. von Hoegh]

http://www.mraggett.co.uk/shuttertester/shuttertester.htm

Well that pretty much does it for me!

Finally, someone made a photocell circuit which works with Audacity and actually measures the time the light comes through the shutter, instead of the sounds the shutter makes! Bravo!

[Greg Miller]

Thanks all for your comments, this pretty much does it for me:

It is quite common for leaf shutters to bounce at their highest speed

1/250th is your shutter's highest speed?

[Charlie Strack]

At 60 mph, the train would be moving 88 ft. per second, and travel about 4-1/4 inches during a 1/250 sec expsoure. That would be along its axis of motion, which is oblique to the camera. This is probably enough to cause some blur in the image..

[Ed Bray]

1/250th is your shutter's highest speed?

Actually the highest one on that shutter is 1/400th but after doing a bit more research it can be all fast speeds but often the two fastest are the likeliest.

[Ed Bray]

At 60 mph, the train would be moving 88 ft. per second, and travel about 4-1/4 inches during a 1/250 sec expsoure. That would be along its axis of motion, which is oblique to the camera. This is probably enough to cause some blur in the image..

Blur wasn't the issue, it was the second set of numbers on the firebox plaque. There is no way that train was doing 60MPH, between 30 and 40 MPH tops I would have said.

[Cletus]

While that's true (though when doing multiple pops to build flash power each exposure is identical) I can't imagine how it would of occurred in this circumstance. I'm thinking instead maybe a brief glimpse of an alternate universe? :)

Kirk - I agree with what you said about the multiple pops and I guess I should've called the anomaly a 'ghost image' rather than a 'double exposure' - kinda misleading and probly not what's happening here. I think I'm back in the shutter bounce camp, although it's rather an interesting problem...oh and sorry for misspelling your name! :o

E., I also agree with what you said in that the numbers most probably are raised on a cast plate, eliminating the "repainted numbers" theory too.

[wombat2go]

I followed this thread all day and studied the full res image on Eizo here and I was stumped.
The trees and the rail are very sharp.
Then I saw the bounce issue mentioned.
I think that explains it, along with perspective on the train's motion, and the non linearity of the film which does not record the bounce in the darker areas of the loco.

Last month I tested my (new) Graflex Wollensak shutter using a biased photo transistor on film side to oscilloscope, and a lamp on lens side, to check the shutter speed calibration
Speeds were reasonably accurate. The illumination intensity on the film is trapezoidal in time as the shutter opens and closes.
There was no bounce. ( Not that I was aware at the time that bounce is an issue)

[EdWorkman]

What IanG said
It's speed blur.
Unless the train was stopped and posed, but you shoulda' said something.
There are plenty of sources for angle/speed/proximity tables for moving objects available- it's Airy.
That said, there seems to be no other clue that the train is in motion, so it's a good thing.

[Nathan Potter]

Interesting problem of motion during exposure. I think Struan is on track with his initial comments, then wombat2go nails the problem. The whole train is exhibiting differential motion across the film plane. The front moves faster than the distant rear as observed by the camera. The blurr pattern is generally horizontal, as is the motion of the train. Some other minor blurring may well be due to other directional motion at the front of the train and, as mentioned above, due to swaying from uneven tracks.

As mentioned by wombat2go the darker areas do not show the blurring as well as the light against dark areas - especially the letters at the front of the engine.

Well, that's my guess. Sort of a visual metaphor of the special theory of relativity and the Lorentz transform, if you want to stretch things a bit.

Nate Potter, Austin TX. / Corea ME.

[Jim Andrada]

Sort of a visual metaphor of the special theory of relativity and the Lorentz transform, if you want to stretch things a bit.

Umm - how fast did you say the train was going???? Probably safe to assume that Lorentz wasn't behind this. Nice idea though!!!!!

[Struan Gray]

Ed was right in his insight that the doubling was the biggest mystery. And it's the doubling, and the way the doubling varies through the scene, that rules out things like simple motion blur or shutter bounce.

Motion blur would indeed be smaller for parts of the train further away from the viewer. However, it would decrease linearly along the train, and, crucially, would be the same for all parts of the train which are the same distance away (excepting the wheels, which are rotating and translating at the same time). Here, the blur dies off suddenly in the tender. More importantly, blur from linear motion turns highlights into straightforward streaks. Here there is doubling - tripling in fact, if you look closely at the plate with '85A' on it, and the smokebox handle.

A multiple image could be caused by shutter bounce. However, the multiple image would be of the same sort for all objects at an equal distance from the camera. There is a variation seen here which rules out shutter bounce. Look at the highlights on the ends of the axles/connecting rods at the front of the engine - they show the sort of straightforward blur that you expect from constant motion. If the doubling/tripling seen on the upper part of the engine were caused by shutter bounce you would see similar effects on the axel highlights, and you don't.

So there is nothing wrong with the camera, the lens, the shutter or the photographer. The train has merely passed over a jolt in the track and the top part is rattling sideways in the usual way. I am not old enough to have regularly travelled on steam trains, but the motion is entirely consistent with what I remember of rolling stock and non-continuous track from the 60s and 70s. Train suspensions have been improved immeasurably since then, and track is laid out much more smoothly, so the very sudden jolts and lurches that used to be commonplace have been largely eliminated.

[IanG]

So there is nothing wrong with the camera, the lens, the shutter or the photographer. The train has merely passed over a jolt in the track and the top part is rattling sideways in the usual way. I am not old enough to have regularly travelled on steam trains, but the motion is entirely consistent with what I remember of rolling stock and non-continuous track from the 60s and 70s. Train suspensions have been improved immeasurably since then, and track is laid out much more smoothly, so the very sudden jolts and lurches that used to be commonplace have been largely eliminated.

Add to that the front of a loco is some way from the wheels amplifying any vibration very significantly.

Ian

[RichardSperry]

If you had a metronome that made one complete transit in 1/250th of a second. And took a 1/250 of a second exposure.

Wouldn't you see two blades, where it slowed down at either side. And nothing in the middle, probably with no blur at all.

Outside of a stuttering shutter, which doesnt appear in any of the other photos, it seems plausible that it is reflecting the side to side vibration of the train at the same time that it is moving forward toward the camera(diagonally) and that the diplopia effect is reflective of where the vibration 'slows' down at either side to be captured.

[evan clarke]

Tracks are sharp, the train's not..I have had this happen with big lake boats in the Milwaukee harbor..They look static but are moving..Same effect with the numbers, they are high contrast against the dark background and it just shows more..

[mitomac]

Was there any swing on the image? I've seen localized ghosting like this when using front tilt with a heavy lens. A very small amount of drift will cause a double image or ghosting that may only be localized to specific regions depending on the plane of focus.

mitomac

[wombat2go]

I am still puzzled but I discount the rocking or rail jolting theories.
If the lantern etc rocked over 50 mm in 1/250 sec ( 4 millisec) and stopped that would be an average G force somewhere above 200 G - military shell territory and the lantern and probably the whole boiler would fail structurally. Correct me if i am wrong, and even if the shutter was 5 times slower, still too high G.

Furthermore Ed's image would be normal among all the 1/250th rocking train photos which it is obviously not.

However if the loco was travelling at smooth speed and a second image hit the film at about 4 millisec later (1/250th) and we saw the points closest to camera move by 50 mm that corresponds to a loco speed of 45 km/Hr towards the camera, and quite likely..
Shutter bounce.
If the shutter opened partially/briefly a second time after 4 millisec or so, all the stationary rails and leaves would stay sharp, which they do in this shot in fact those are very sharp, and the brightest moving points would be re exposed.

[Ed Bray]

No, no swing, in fact no movements at all.

[EdWorkman]

I posted after IanG yesterday and it disappeared [I think]
My login is now fixed [ I think]
I second IanG
It's simply motion blur.
You 2 sec @ f64 large format folks are over thinking this :>)
Motion blur is a function of proximity, speed and angle: geometry and Airy.
Thus a close broadside needs a fast shutter, a head-on , faraway can do it with a slow speed.
Subject movement at broadside is easily computed, and by geometry converted to movement at the film plane.
And so for an oblique view. If say an easy 45 degrees, the translation on the film is 70% that of broadside, less at 30 degrees, etc.
Draw some similar triangles, holding the lens to film distance constant and increasing the lens to subject distance and you will see how the translation distance on the film reduces as the distance from lens to subject increases.
For a head-on shot, movement is the increase in width on the film plane as a function of geometry and the amount of movement toward the lens in the exposure interval- at any reasonable distance that dimension is quite small
Although there are differences between FP and leaf shutters, the tables that Graflex published a century ? ago illustrate the parameters in tabular form.
Note that the second example is closer to head-on than the first and less speed blur.
I don't recommend head-on shots of speeding machines, but I have shot 8x10 action views.
I posted one in the "Show Us your Trains' forum a year or so ago. Near broadside but only 20-30 mph and sufficiently distant. Had it been faster I would have made it more oblique, but stuck with 1/250 f16

[Struan Gray]

wombat, I'll admit the needed accelerations seem large, certainly for simple harmonic motion over 5 cm at 250 Hz. Limited shutter efficiency (which lengthens the period), other plausible motion profiles, and assuming a smaller train can bring that down to less extreme values of the order of 10 g, but that still seems large, even it it won't shear any rivets.

Most modern passenger trains aim for a maximum lateral acceleration of 0.1 g, although recordings of trains and trucks passing over bumps in roads and tracks show that uncushioned cargo can be subjected to instantaneous accelerations of a few g in normal travel. (I've taken my numbers from figures given in Swedish public transport research).

As I said, I agree this seems a bit extreme. On the other hand, I cannot reconcile a straightforward case of shutter bounce with the way the doubling/blur gets worse the higher up the train you go. Bright features down at track level are barely blurred at all, while those up at roof level are clearly doubled or tripled. This pattern is consistent all along the engine and at the front of the tender.

Ed is of course welcome to test his shutter, but I suspect a better answer would be to measure the train carefully and/or take a ride on it with a recording accelerometer (aka iPhone) clamped to the body of the train.

[evan clarke]

The train is moving..

[Ed Bray]

The train is moving..

Thanks!

[EdWorkman]

Ed Ed Ed Ed EEEEEedddd
Your Lens is fine
Your shutter is just fine
You took fine pictures
You don't need nanometers, accelerometers, thermometers, rocket science or angst.
A little speed blur shows the train is in action, and a little MORE wouldn't hurt a bit.
But that's all it is
Be happy, listen to evan
regards
Ed

[Ed Bray]

Thank you all.

I got my Toyo Field to Linhof lensboard adapter today so am now able to use all my lenses on the Canham, so I suspect I shall have even more blurry images going forward as for the first time I shall be able to use my 3x Compur3 lenses (240mm, 300mm & 360mm) on the Canham.

Longer focal length lens with slower Shutter Speed = more blur

[Mike Anderson]

Maybe next time inspect those tracks and file down any bumps. :)

[al olson]

Ed, inspect the lens board for a small hole off axis. It could be that you are making a pinhole exposure during the time that you pull the dark slide and when you put it in. With other photos, the pinhole exposure may be shorter and, hence, less noticeable.

[evan clarke]

And the whole picture really looks good, enjoy the subject and forget the tech..

[Ed Bray]

Ed, inspect the lens board for a small hole off axis. It could be that you are making a pinhole exposure during the time that you pull the dark slide and when you put it in. With other photos, the pinhole exposure may be shorter and, hence, less noticeable.

Thanks for the suggestion Al, the lensboard is 2.5mm thick aluminium. There are no other holes in it apart from the Copal 1 sized hole for the lens. That would be moot now as the 210mm and 72mm lenses are now in Technika boards as I have converted the Canham to use them rather than the Toyo field boards it usually uses.

The Toyo Field to Linhof adapter is a very nicely machined bit of kit and seems to be milled from a block of Aluminium which has then been anodised. Not too expensive either compared to comparable type boards from elsewhere, ebay #140559022957.

[Struan Gray]

Longer focal length lens with slower Shutter Speed = more blur

Where there's a will there's ...... supplemental lighting (http://en.wikipedia.org/wiki/File:Mosquito_photo-reconnaissance_photoflash_bomb_loading_WWII_IWM_C_4998.jpg).

[Ed Bray]

Where there's a will there's ...... supplemental lighting (http://en.wikipedia.org/wiki/File:Mosquito_photo-reconnaissance_photoflash_bomb_loading_WWII_IWM_C_4998.jpg).

Very good ;)