Pages: [1]   Go Down

Author Topic: 5DS R Motion blur impact on MTF for 3 shooting modes  (Read 6429 times)

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1893
5DS R Motion blur impact on MTF for 3 shooting modes
« on: February 07, 2016, 01:49:06 am »

Attached is a chart showing the effect of motion blur on MTF in the three different shooting modes, regular, mirror lockup, and live view.

Manfrotto tripod, 1 ft center pole extension, 70-200mm at 200mm. F5.6, 1/45s.

The attached chart shows the three modes in red (regular), blue (mirror lockup), green (live view).

Motion blur is the most extreme along one axis and least on the axis at 90 degrees. Maximum motion blur occurs along the 11 o'clock position, just to the left of vertical.  The dashed lines are the response along the maximum blur axis, the solid lines are along the least blur axis.

The regular shooting mode shows the most blur and it is rather extreme. MTF of .5 at only 21 lpmm compared to 54 lpmm for the live view. From other experiments the live view MTF numbers are almost entirely mostly from lens and diffraction effects. What's interesting is that the mirror lockup mode is only reducing motion blur by about half.

The normal use MTF values (dashed red line) bobbles about after lpmm > 30 and those numbers should be considered artifacts that are mostly from spacial frequency aliasing. The lined out text is incorrect. The results are from the nature of the low pass spacial filter that motion blur induces, not spacial aliasing.

Note: autocorrect turned "spacial into special." I have fixed those instances.
« Last Edit: February 07, 2016, 11:52:05 am by Doug Gray »
Logged

Jack Hogan

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 795
    • Hikes -more than strolls- with my dog
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #1 on: February 07, 2016, 05:54:13 am »

Good work, Doug, your investigation seems to point to the mirror as the major culprit of the vibration.  On a tripod with good technique and MLU up shutter shock is also an issue for some cameras.  Does live view on the 5DSR switch to electronic first curtain?  If so that could explain the difference between MLU and LV.

Jack
Logged

ErikKaffehr

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 11251
    • Echophoto
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #2 on: February 07, 2016, 07:36:38 am »

Doug,

Thanks for good info. I was also thinking about the EFSC.

Best regards
Erik


Good work, Doug, your investigation seems to point to the mirror as the major culprit of the vibration.  On a tripod with good technique and MLU up shutter shock is also an issue for some cameras.  Does live view on the 5DSR switch to electronic first curtain?  If so that could explain the difference between MLU and LV.

Jack
Logged
Erik Kaffehr
 

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3354
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #3 on: February 07, 2016, 08:51:50 am »

Good work, Doug, your investigation seems to point to the mirror as the major culprit of the vibration.  On a tripod with good technique and MLU up shutter shock is also an issue for some cameras.  Does live view on the 5DSR switch to electronic first curtain?  If so that could explain the difference between MLU and LV.

Good point regarding EFCS, Jack. When shooting in live view, it makes sense to use ECFS if it is available so as to reduce shutter imposed vibration. If the tripod support is less than optimal, it would also be interesting to investigate the effect of image stabilization at relatively low shutter speeds.

Bill
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1893
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #4 on: February 07, 2016, 11:34:30 am »

Good work, Doug, your investigation seems to point to the mirror as the major culprit of the vibration.  On a tripod with good technique and MLU up shutter shock is also an issue for some cameras.  Does live view on the 5DSR switch to electronic first curtain?  If so that could explain the difference between MLU and LV.

Jack

Jack,

Indeed EFCS is in effect in live view. The last curtain is quite fast. While it imparts some mechanical momentum, it doesn't have much time to translate into significant physical motion. There is, however, another mechanical effect. The lens aperture reduction which must be done before the EFCS activates. It appears in live view that the EFCS is activated about 1 second after the lens aperture is closed. This allows a little time for any mechanical resonances to damp down. It would be nice to be able to control this - perhaps extend it to several seconds. The lens aperture mechanics will vary from lens to lens.

Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1893
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #5 on: February 07, 2016, 11:43:28 am »

Good point regarding EFCS, Jack. When shooting in live view, it makes sense to use ECFS if it is available so as to reduce shutter imposed vibration. If the tripod support is less than optimal, it would also be interesting to investigate the effect of image stabilization at relatively low shutter speeds.

Bill

IS is another variable and is typically quite effective. In earlier tests with long exposures (> 1/5 th sec) it appeared to introduce some motion blur due to slow drift. It is, however, quite good at reducing mirror and shutter induced blur at faster shutter times and is useful even at relatively high speeds.

Logged

bjanes

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 3354
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #6 on: February 07, 2016, 12:14:35 pm »

IS is another variable and is typically quite effective. In earlier tests with long exposures (> 1/5 th sec) it appeared to introduce some motion blur due to slow drift. It is, however, quite good at reducing mirror and shutter induced blur at faster shutter times and is useful even at relatively high speeds.

Do I interpret correctly that your response applies to tripod shooting at higher shutter speeds? There is some confusion about IS with tripods as discussed here.

Bill
Logged

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1893
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #7 on: February 07, 2016, 01:06:34 pm »

Do I interpret correctly that your response applies to tripod shooting at higher shutter speeds? There is some confusion about IS with tripods as discussed here.

Bill

Yes, On tripods I've noticed that IS introduces motion blur at very slow speeds but at higher ones ( faster than 1/5 sec or so ) IS significantly reduces motion blur on a tripod. And, of course, IS does produce major improvements handheld at most speeds.  Theoretically, there may be some sort of crossover point at very high speeds (I'm guessing at or above 1/500th) where IS slightly increases motion blur from mirror slap by overcorrecting. At most it's a quite small effect

With these high MP sensors these subtle issues have become more important. At least where the highest possible resolution is desired.
Logged

dwswager

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1375
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #8 on: February 11, 2016, 08:52:01 am »

Jack,

Indeed EFCS is in effect in live view. The last curtain is quite fast. While it imparts some mechanical momentum, it doesn't have much time to translate into significant physical motion. There is, however, another mechanical effect. The lens aperture reduction which must be done before the EFCS activates. It appears in live view that the EFCS is activated about 1 second after the lens aperture is closed. This allows a little time for any mechanical resonances to damp down. It would be nice to be able to control this - perhaps extend it to several seconds. The lens aperture mechanics will vary from lens to lens.

Good to know that I was right to enable EFCS even when in Mirror Lockup Mode.  Thanks you for the hard  work that we all benefit from!
Logged

FelixBelloin

  • Newbie
  • *
  • Offline Offline
  • Posts: 34
    • My gallery
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #9 on: May 27, 2016, 04:47:24 am »

Hi Doug,
Have you tried the 2 seconds delay mirror lock up mode? It's the one I use almost all the time, would like to know your view on it.
Cheers,

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1893
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #10 on: May 28, 2016, 07:33:48 pm »

Hi Doug,
Have you tried the 2 seconds delay mirror lock up mode? It's the one I use almost all the time, would like to know your view on it.
Cheers,

I sometimes use the 2 second MLU delay for normal work but haven't tested it. I was trying to separate out the Mirror slap from the shutter induced vibration. Just offhand I would say the 2 second delay gets rid of the large majority of the mirror induced motion except on very long, heavy lenses that can take longer for vibration to die down.
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8232
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #11 on: May 29, 2016, 10:21:48 am »

I sometimes use the 2 second MLU delay for normal work but haven't tested it.

Hi Doug,

I think you mean the self timer exposure delay of 2 seconds? In my testing of other cameras, a 4 second delay after the mirror slap would more or less allow all vibrations to decay to insignificant levels, but then the shutter release adds its own vibration which will impact image resolution.

Quote
I was trying to separate out the Mirror slap from the shutter induced vibration.

Yes, very much appreciated since this is usually poorly documented, but of high importance for high resolution cameras. In fact, adding the shutter release vibration to the not yet decayed mirror vibration might even cause an amplification if the amplitudes are in phase. So to find out the best combination of timings, one would need to add a timing component to the test.

The 5DS (R) offers a new adjustable shutter-release time lag for Mirror Lockup shooting, but the maximum delay is 1 second.
Therefore, I'd see it more as a handheld shooting feature (not for action shooting though) and to a lesser extend effective for tripod use (although it might help to avoid the in-phase amplification). If you can time and measure the vibration decay, it will probably show the requirement for a 2-4 second delay to be really effective.

I use a separate vibration logger which has modest weight/mass of itself, and it allows to record acceleration data at a high enough sampling frequency (user definable up to 512 Hz) to make sense for most of the exposure times we would be interested in. I have a previous version which is slightly slower (320 Hz), but it already shows a few vibration swings within a second, even when used on a heavy tripod without center column extension on a concrete floor. My 1Ds Mark III allows a 16 second shutter release delay, so I can isolate both impulses.

Quote
Just offhand I would say the 2 second delay gets rid of the large majority of the mirror induced motion except on very long, heavy lenses that can take longer for vibration to die down.

I don't have a 5DS (R) to test, but my 1DS Mark III (a heavier body but with older mirror technology) requires more than 2 seconds to get rid of measurable vibrations. The complexity then becomes, how the remaining vibration translates to image shake, which additionally varies with the lens used and the projected image magnification factor. But anyway, if there is no measurable vibration, there can be no visible vibration blur during exposure.

So your test is very useful for assessing the image effect, but it is an aggregate that is accumulated during the actual exposure time. It would require an acceleration sensor, or a flying spot on a scope or something like that, to gain insight as to which time delay is minimally required to avoid (most of the) vibration.

Cheers,
Bart
« Last Edit: May 29, 2016, 10:42:39 am by BartvanderWolf »
Logged
== If you do what you did, you'll get what you got. ==

Doug Gray

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 1893
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #12 on: May 29, 2016, 04:23:01 pm »

Hi Doug,

I think you mean the self timer exposure delay of 2 seconds? In my testing of other cameras, a 4 second delay after the mirror slap would more or less allow all vibrations to decay to insignificant levels, but then the shutter release adds its own vibration which will impact image resolution.

Yes, very much appreciated since this is usually poorly documented, but of high importance for high resolution cameras. In fact, adding the shutter release vibration to the not yet decayed mirror vibration might even cause an amplification if the amplitudes are in phase. So to find out the best combination of timings, one would need to add a timing component to the test.

The 5DS (R) offers a new adjustable shutter-release time lag for Mirror Lockup shooting, but the maximum delay is 1 second.
Very cool Bart.

I think you are right about the fractional time delay being most useful for handheld.  Vibrations damp down very quickly when handheld. I haven't done any measurements on that but it would be something worth exploring.

I love the idea of instrumenting with vibration sensors. There is also an opportunity for people to make something that rapidly absorbs or counters these vibrations. A given configuration should produce highly repeatable mechanical moments and a good system could use "feed forward" servo (control theory) techniques. Possibly, the camera/lens manufacturers could incorporate this into the lens stabilization mechanism. It would be especially effective for rapid pictures at 5+ fps.
Logged

Bart_van_der_Wolf

  • Sr. Member
  • ****
  • Offline Offline
  • Posts: 8232
Re: 5DS R Motion blur impact on MTF for 3 shooting modes
« Reply #13 on: May 30, 2016, 11:44:01 am »

I love the idea of instrumenting with vibration sensors. There is also an opportunity for people to make something that rapidly absorbs or counters these vibrations.

Yes, but it depends on the specific camera/lens mass and balance, and tripod + floor vibration dampening characteristics. Sometimes just adding a certain amount of mass in a strategic position can do wonders.

Attached a quick example of my 1Ds Mark III, first shutter curtain induced acceleration/vibration of the camera with an EF 100mm Macro lens on it, on a heavy tripod. Thee vertical vibration is largest, due to the travel direction of the shutter curtain. Such data allows to vary elements of the setup that would produce even less vibration. Less vibration during the shutter open time, will reduce visible blur.

Cheers,
Bart
Logged
== If you do what you did, you'll get what you got. ==
Pages: [1]   Go Up