7
\$\begingroup\$

In this photo lines are clearly visible.

  • Shutter speed - 32 sec.
  • Aperture - f/9
  • ISO - 100
  • Lens - EF-S18-55mm f/3.5-5.6 IS II
  • Focal Length - 39 mm

I have took the photo through the window, with the lights off

\$\endgroup\$
9
  • 14
    \$\begingroup\$ It's not clear which lines you're talking about, can you provide more information? \$\endgroup\$ Commented Aug 3, 2016 at 9:07
  • 16
    \$\begingroup\$ Specifically, which lines are you talking about - there are a lot of lines in that image, which ones are you confused about? Edit the image, and circle in red the bits you're asking about. \$\endgroup\$ Commented Aug 3, 2016 at 9:27
  • 1
    \$\begingroup\$ Lines of street lamps. \$\endgroup\$
    – iEPCBM
    Commented Aug 3, 2016 at 9:34
  • 30
    \$\begingroup\$ When someone asks you to clarify, or to give more information, just repeating the same phrase you used before is not helpful. Give us more words. What do you mean by "lines of street lamps"? Mind you, just looking at the image it's pretty clear that most of the noise is that you didn't hold the camera still. \$\endgroup\$ Commented Aug 3, 2016 at 15:31
  • 1
    \$\begingroup\$ Did you open the window or did you shoot through the glass? \$\endgroup\$
    – MirekE
    Commented Aug 4, 2016 at 1:10

6 Answers 6

34
\$\begingroup\$

It looks like there are parallel light trails below each streelamp -- going down, then right, then down some more (ASCII art):

 /
|
|
|
\_
  \
   |

And highlighted on the original: enter image description here I would guess that these are when the shutter button was pressed, tilting the camera, because only bright sources show this effect. This is in addition to the normal, more horizontal/wandering light trails. If you view full-res you can see the individual pulses of light caused by the AC power (fluorescent and similar lights used to run at twice mains frequency, but newer models like these are high frequency). These pulses are more widely spaced in the long straight of the trail, close together (they run into each other) where the camera is changing direction more slowly. I can tell these are high-frequency lights because counting the pulses at the usual 100-120Hz twice mains frequency would account for a significant proportion of your exposure time, and the windows on the opposite buildings demonstrate that this isn't the case.

Zoomed in on the pulses:

enter image description here

This also shows the tail light trails you'd expect, indicating that the disturbance didn't last for very much of the exposure.

The solution is to reshoot ideally with a tripod and a cable release. You can improvise for the tripod but then the cable release becomes essential (or wireless release, or self timer, just don't touch the camera). Even slamming doors can make the camera move. I'd turn off the IS - it can't help you on such long exposures, with the possible exception of vibrations through the floor.

\$\endgroup\$
5
  • 1
    \$\begingroup\$ Yes. You are probably right. Tonight I will try to do it again, following your instructions. \$\endgroup\$
    – iEPCBM
    Commented Aug 3, 2016 at 10:22
  • 10
    \$\begingroup\$ You can capture without a cable release by turning on the 2s (or 5-10s) delay, or remotely control the camera via wifi if supported \$\endgroup\$
    – phuclv
    Commented Aug 3, 2016 at 10:45
  • 4
    \$\begingroup\$ @LưuVĩnhPhúc you can, hence my "or self timer". I hadn't considered Wifi because remote triggering over Wifi isn't in any camera I've used; arguably it's covered by "wireless release" though what I had in mind was a dedicated IR or radio trigger. \$\endgroup\$
    – Chris H
    Commented Aug 3, 2016 at 10:51
  • 2
    \$\begingroup\$ Even with the cable release, you want the two-second delay: this lifts the mirror at the beginning of the timer period, letting any vibrations die out before opening the shutter. \$\endgroup\$
    – Mark
    Commented Aug 3, 2016 at 21:18
  • \$\begingroup\$ @Mark, good point though not all models do this quite the same way (that delay may be set separately). \$\endgroup\$
    – Chris H
    Commented Aug 3, 2016 at 21:23
9
\$\begingroup\$

How annoying :-) - I finished marking up a photo and now see that Chris H has done much the same but better. So I'll post this 'for completeness' but with less comment than I otherwise would have.

I saw a movement pattern similar to Chris's. This may have occurred in one motion, but the variable brightnesses along each path suggest it may have been a series of small steps with some delays between as the camera shook and settled.

Clicking on this marked up version will display the full resolution upload that allows the paths to be better seen. I have numbered the separate directions of movement from 1 to 6 but these probably occurred in the opposite order.

The finest example of motion is perhaps from the small green light in the red rectangle at upper left. There are many more examples than the ones marked.

enter image description here

You were lucky to get as good a result as you did in this case - the movement is MUCH larger than what is caused by typical camera shake in long exposures. While camera shake usually causes blurring you have movement of about 1/3 of the frame. You should use either a tripod OR a rock solid mount. Pressing the camera onto a solid unmoving surface can be good enough but care is needed. If using a low stability tripod or similar them movement after releasing the camera can last for many seconds. Use of a 10 second timer helps but in some cases vibration may last longer than that. If a solid tripod or other mount is available a 2 second timer pr-release is usually enough to get your hands clear of the camera and for minor movement to die out. If your camera or lens has "antishake" turn it OFF for long exposures with tripod or similar mounting. Remember to turn it on again afterwards.

\$\endgroup\$
1
  • \$\begingroup\$ Your markup is nicer than mine though. I would guess something like an improvised beanbag was used so it was stable after poking the shutter button. \$\endgroup\$
    – Chris H
    Commented Aug 3, 2016 at 12:04
7
\$\begingroup\$

Although the poster hasn't specified it, this answer assumes this picture was shot on a tripod (it looks too sharp to be hand-held, even holding the camera steady against the window).

Many claimed that this is caused by a movement of the camera. However, I believe that it is actually due to the optical stabilization (IS in Canon parlance) trying to compensate for a movement that isn't there on a tripod. The fact that the movement is very irregular, slow, then turn, then fast (the fluorescent/led light showing irregular blinking patterns) indicates that something active is going on - the simple shake after pressing the shutter button cannot explain this.

I have seen very similar artifacts (although not as massive) once when I forgot to turn it off on my Tamron 70-200 f/2.8 (mounted on a Nikon D750). enter image description here

When shooting on a tripod, remember to turn off the optical stabilization (IS/VR/VC/OS or whatever it is called on your lens).

\$\endgroup\$
1
  • 1
    \$\begingroup\$ It's worth remembering some lens specs (generally higher-end) do state that they can detect when tripod mounted and so don't need the IS turned off, but there's no disadvantage that turning them off anyway. \$\endgroup\$ Commented Aug 4, 2016 at 8:09
6
\$\begingroup\$

If you consider the light trails of the headlights of a vehicle, in the foreground, that appear to pass through some parked cars - it seems likely that the camera was tilted vertically during the 32 seconds the shutter was open while exposing the image (moving those trails from the parked cars up to the same place in the image that all the other vehicle lights traverse the image). This tilt would also account for the trail of the street lights - which I assume are the ones that are causing confusion.

They can be removed using an image editor (like Photshop).

They can be avoided by not moving the camera during an exposure.

\$\endgroup\$
3
  • \$\begingroup\$ I suspect you'll find those are light trails caused by the cars pulling off the road and parking. The ones in the far left, no idea what caused those. But the street lamps don't look like the lens was moved. \$\endgroup\$ Commented Aug 3, 2016 at 9:28
  • 2
    \$\begingroup\$ @EightBitTony I agree with David, the camera definitely moved. The path of the white headlights is because the car was moving while the camera was moving. The direction of this suggests (I think) suggests that the camera moved up at the start of the exposure, rather than down at the end of the exposure. Assuming that the white lights of the cars are facing the camera, therefore those lights travelling right. In my mind that suggests that pressing the shutter button may not actually be the cause of the movement \$\endgroup\$ Commented Aug 3, 2016 at 10:27
  • \$\begingroup\$ Yep, i see it now, I thought the lamps were lamps, not simply the lamp light leaving a trail as the camera moved. \$\endgroup\$ Commented Aug 3, 2016 at 10:31
2
\$\begingroup\$

Those street light lines are due to the camera motion mostly at the time of shutter release (or shutter closing ) you may avoid them to some extent by having Camera Timer of say 3s or so

\$\endgroup\$
-2
\$\begingroup\$

I am not sure how to remove them in post but I think the reason you see those streaks of light is because your shutter speed is too low.

This looks like a case of the rolling shutter of a CMOS sensor. You might argue that there is no object in motion. But there is: light. Light in motion is being redrawn by your CMOS sensor because of your very low shutter speed.

Try taking the picture again by increasing the shutter speed and opening up the iris a bit further to compensate for the loss of light.

The non-linear fibers of light are due to the ISO. I know you've only got it to the lowest value but still, in my experience, that's what an ISO does to street lights.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ Exposure time was 32 seconds. He was trying for a head/tail light tracks type long exposure photo. The 'artefacts' are due to massive camera movement. \$\endgroup\$ Commented Aug 3, 2016 at 13:43
  • 1
    \$\begingroup\$ Rolling shutter shows up for very short exposures (where the time it takes the shutter to open/close is large compared to the time of exposure). This is certainly not the case with a 32 seconds exposure. \$\endgroup\$
    – Calimo
    Commented Aug 3, 2016 at 14:37
  • \$\begingroup\$ Oh yes, I said it backwards. This couldn't have been the rolling shutter. For the rolling shutter to have occurred, it would have had to have been a very high shutter speed. \$\endgroup\$ Commented Aug 3, 2016 at 15:02

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.