Exposure duration

[Exodies]

Advertisement (gone after registration)

How long does an exposure take for different shutter speeds on the M?

I'm guessing it won't be much quicker than the flash sync speed. This has an impact on choosing a minimum speed to avoid hand-held shake.

[AlanJW]

How long does an exposure take for different shutter speeds on the M?

I'm guessing it won't be much quicker than the flash sync speed. This has an impact on choosing a minimum speed to avoid hand-held shake.

Can you clarify?  A 1/100 second exposure will take 1/100 second, 1 second takes 1 second, etc.  Are you talking about shutter lag? -- the amount of time between pressing the shutter and it opening?  On my M it is imperceptible.  I am sure someone has measured, but I have felt no need.  I have been able to handhold on my M about the same as I was able on a film camera -- if I really pay attention and use good technique I can get usable images at 1/15.  This is in classic metering mode not the so-called "advanced" modes, so I will leave an answer on that to someone else.

[hx911]

I would (could) imagine the 'time' the op thinks about is from start of exposure till the end of the esposure - which is different (sometimes - always?) to the time of exposure that a pixes is exposed to.

It is the time it takes for the opening in the shutter to pass past the complete sensor, start is 1st curtain starts opening, end is 2nd curtain finished closing.

Sync speed, as I understand it, is the shortest exposure time where the 2nd curtain starts closing after the 1st curtain is fully open.

[Exodies]

Sorry, I always do this - I try my best to be succinct but merely succeed in being obscure.

I'm talking about the faster shutter speeds, not lag at all. At 1/4000 sec the gap between the two curtains must move across the frame relatively slowly.

Does each curtain always move at the same speed no matter what the shutter speed is? Is this the flash sync speed? If so then the fastest motion stopping or hand-held shake reducing duration is 1/180 sec.

[jaapv]

The speed of the curtain movement is always the same. The width of the slit between the curtains varies to obtain the exposure time.

 

At shutter speeds longer than 1/180th the second curtain must be delayed before starting out, (but the curtains still move at the same speed)  so there is a moment that the whole sensor is exposed, thus a flash, which is very short, say 1/10000th of a second (it varies with design and flash power), can obtain an even exposure if the flash is fired at the right moment.

 

This leads to the conclusion that the shutter curtains move at a speed of 24 mm/180th sec, which is 4320 mm/sec. As the distance between the curtains varies the total duration  of shutter movement varies linearly from just over 1/180th of a second @ 1/4000th to 8 seconds   @  8 sec.

 

 

Your conclusion about shake reduction is incorrect,  as a more narrow slit between the curtains will result in each pixel being exposed to light for a shorter time, thus the amount of subject  movement that can be recorded is less.

 

What does happen, though, is that with a moving subject (or with camera shake, which comes to the same) the relative  positions of the image elements on the sensor or film will vary from top to bottom, or left to right with a horizontal shutter, causing a distorted image.

This caused, for instance, the well known “oval wheel” symbolism for a fast-moving car early last century, as the shutters of the time were too slow to record a circle moving at speed correctly.

Welcome, dear visitor! As registered member you'd see an image here…

Simply register for free here – We are always happy to welcome new members!

[giordano]

With electronically-controlled Leica shutters it's a little misleading to say that the width of the slit between the curtains varies to obtain the exposure time. In fact there is no direct control of slit width at all. Instead, the two curtains (or sets of blades) move independently, each driven by its own springs. When the first curtain is released a timer is started which releases the second curtain after the appropriate delay. If the delay - the exposure time - is short enough the effect is a "slit".

 

On mechanical Leicas, the curtains still move independently, but for exposures shorter than the synch speed the delay is controlled by releasing the second curtain after the first has moved a certain distance.  

 

The above is true of almost all focal plane shutters ever on 35mm and medium format cameras. It was only on large format cameras (e.g. Speed Graphic) that exposure was set directly by the width of the slit.

[jaapv]

Advertisement (gone after registration)

Yes, that is indeed the way the width of the slit is set on a modern shutter. Slit width being defined as the distance between the shutter blades as they travel across the sensor, it is not really relevant which way it is controlled and released. It has indeed been a long time since this was preset mechanically  as  a distance and moving as one unit.

 

Thanks for the clarification

[Exodies]

So the shortest time you need to hold the camera steady is 1/180 sec. If you move during this time, whatever the exposure time, then the picture will be expanded or compressed depending on whether your twitch is in the same or opposite direction to the movement of the curtains.

If you are doing some setting or calculation based on 1/(x * focal length) then it is pointless if the result is less than 1/180 (for example, maximum exposure time for auto ISO).

[CheshireCat]

If you are doing some setting or calculation based on 1/(x * focal length) then it is pointless if the result is less than 1/180 (for example, maximum exposure time for auto ISO).

 

This is false. If you expose a picture for 1/4000s, every pixel (sensel) is exposed for 1/4000s (even if possibly at slightly different moments in time). Therefore your per-pixel motion blur will be limited to 1/4000s, as expected.

 

Nevertheless, the problem with the 1/FL rule of thumb is that it was invented in the film era based on the acceptable blur level at typical viewing distance.

The global blur in the overall image is the same on digital, however local per-pixel blur depends on pixel pitch, and is therefore proportional to the number of pixels in the blur linear direction. Hence, if you want per-pixel sharpness, you need to use a faster shutter speed on cameras with a higher number of pixels.

So, like you said, the new rule is 1/(x*FL), where "x" is proportional to the linear pixel density and to how shacky your hands are

[giordano]

In the presence of either subject movement or camera shake, it is the exposure time that counts as far as sharpness is concerned.

 

The time it takes the shutter to traverse the frame is only relevant when considering distortion of the kind illustrated in the photo in Jaap's post #5 in this thread. In that picture, Lartigue was panning the camera to follow the car, not fast enough to fully compensate for the car's motion (if he had, the car would not have been distorted but the spectators in the background would have had both more distortion and more motion blur).

 

In the picture, the racing car is a few metres from the camera and would probably have been travelling at say 100km/h or about 30 m/s. I don't know what sort of camera he was using, but the focal plane shutter would have taken somewhere around 100ms (1/10 second) to traverse the frame, 18 times longer than a digital Leica's, making the distortion 18 times worse.

 

What I'm saying is that motion-caused distortion was an issue with old, large-format focal plane shutters, but is seldom noticeable with small modern ones.

[jaapv]

 

So the shortest time you need to hold the camera steady is 1/180 sec. If you move during this time, whatever the exposure time, then the picture will be expanded or compressed depending on whether your twitch is in the same or opposite direction to the movement of the curtains.

If you are doing some setting or calculation based on 1/(x * focal length) then it is pointless if the result is less than 1/180 (for example, maximum exposure time for auto ISO).

No, as Cat says, the exposure time is not determined by the time that the shutter blades take to traverse the sensor but by the distance between the shutter blades. If that distance is small, the time that the pixel is exposed to light is short, if that distance is larger, the pixel will be exposed for a longer time.

[Jeff S]

 

In the picture, the racing car is a few metres from the camera and would probably have been travelling at say 100km/h or about 30 m/s. I don't know what sort of camera he was using, but the focal plane shutter would have taken somewhere around 100ms (1/10 second) to traverse the frame, 18 times longer than a digital Leica's, making the distortion 18 times worse.

 

 

http://www.largeformatphotography.info/forum/showthread.php?31903-Jacques-Henri-Lartigue-and-his-camera

 

Jeff

[giordano]

http://www.largeformatphotography.info/forum/showthread.php?31903-Jacques-Henri-Lartigue-and-his-camera

 

I may have underestimated the speed of the vehicle. It's identified in the fascinating thread linked to above as a Delage type F of about 1908, which would have had a maximum speed about 125 km/h. But that was a light car with a single-cylinder 1256cc engine; big racing cars in that era could do well over 160 km/h. http://www.classicandsportscar.com/news/classic-car-events/165-litre-edwardian-monster-to-take-on-beast-of-turin

[pico]

Speaking of the particular auto racing era - each car carried the driver and mechanic.

What was that about exactly? To push the car off the road when it failed? Or was it

to bind him to the responsibility of maintaining a good machine, as in "you build

it, you ride in it and I'll be there in your face with bloody hell if it breaks." ?

[giordano]

Almost all racing in those days was on public roads (hopefully cleared for the occasion). About the only purpose-built tracks were Brooklands (opened 1907) and Indianapolis (1909). Road circuits were much longer than we're used to: the 1912 French Grand Prix was run on a 79km circuit near Dieppe. Breakdowns were common and punctures and blow-outs very common, so the riding mechanic, toolbox and two spare wheels weren't just there for ballast.

 

In addition the mechanic was responsible for fuel supply (automatic fuel pumps hadn't been invented so he used a hand pump to pressurise the main fuel tank), lubrication (hand pumps again), watching the gauges, watching the competition (at least until rear-view mirrors were invented), cleaning the driver's goggles, and so on.

 

According to Wikipedia, riding mechanics in Grand Prix races were banned in 1924, seemingly to reduce the death toll - though with shorter circuits and better technology the need for them was much smaller.

[pico]

In addition the mechanic was responsible for fuel supply (automatic fuel pumps hadn't been invented so he used a hand pump to pressurise the main fuel tank), lubrication (hand pumps again), watching the gauges, watching the competition (at least until rear-view mirrors were invented), cleaning the driver's goggles, and so on.

 

 

I snipped good information. See Giordano's post.

 

I am familiar with Total Loss oil systems, for example in the early Harley-Davidson which had two tanks: on for gasoline, the other for oil which ran into the engine, then to the ground. Well, back then there was horse sh*t on the dirt roads. Simple times.

[giordano]

I am familiar with Total Loss oil systems, for example in the early Harley-Davidson which had two tanks: on for gasoline, the other for oil which ran into the engine, then to the ground. Well, back then there was horse sh*t on the dirt roads. Simple times.

 

Rapidly complexifying, however. By 1914 the Peugeot GP engine had hemi head, central sparking plugs, four valves per cylinder, dohc, and a proper dry sump.