When you say "pan," did you mean that you were literally panning the camera, like trying to track a moving subject? Could you elaborate a little? Thanks.
-
Welcome to the new forums! Please read this first. For known issues we are working to resolve, click here.
VR: Why you should turn it off when it's not needed
- Thread starter em_dee_aitch
- Start date
edwardaneal
Veteran Member
well - - I am not an engineer and Nikon does not, to my knowledge, publish the limitations of the VR system.
What I can say with certainty is that shooting BIF hand held with my 70-300VR at 300mm with shutter speeds between 1/500th and 1/1000th with VR turned I have absolutely no problems with soft images due to camera shake. I can also say that shooting BIF with my old 300mm f/4 at the same shutter speed range without VR the number of sharp images unaffected by camera movement was more like 70%.
to me this clearly indicates that having VR at shutter speeds between 1/500th and 1/1000th at 300mm is an asset
Also I have viewed all of these images at 100% and I see no degradation or difference in quality between these images and images shot at higher shutter speeds without VR with either lens. This is why I sold the 300mm f/4
Primary kit - D200, 10.5mm f/2.8D, 35mm f/1.8G, 50mm f/1.4G & 70-300VR
Backup kit – D80, 18-105VR
SB800, SB600 and other misc lighting equipment
Lenses worth mentioning owned and sold– 12-24 f/4, 17-55 f/2.8, 35-70 f/2.8, 80-200 f/2.8, 20mm f/2.8, 35mm f/2, 50mm f/1.8, 50mm f/1.4D, 60mm f/2.8, 85mm, f/1.8. 105mm f/2D-DC, 180mm f/2.8, 300mm f/4D-ED
What I can say with certainty is that shooting BIF hand held with my 70-300VR at 300mm with shutter speeds between 1/500th and 1/1000th with VR turned I have absolutely no problems with soft images due to camera shake. I can also say that shooting BIF with my old 300mm f/4 at the same shutter speed range without VR the number of sharp images unaffected by camera movement was more like 70%.
to me this clearly indicates that having VR at shutter speeds between 1/500th and 1/1000th at 300mm is an asset
Also I have viewed all of these images at 100% and I see no degradation or difference in quality between these images and images shot at higher shutter speeds without VR with either lens. This is why I sold the 300mm f/4
--The effect at a given spot, and the transition to unaffected spots, is better "frozen" by the higher speed/narrow slit. With a slower exposure and wider slit (or full frame exposure), the motion of the VR system combined with any residual vibration that it can't quell could lessen the harshness of the transition and also make the better part of the image softer.
I tend to disagree for the reason noted above, and also because with slower speeds expectations are lowered, and you're more likely to say "wow, I can't believe VR got me that shot"
--
David Hill
http://www.sfbayweddingphotographer.com
San Francisco, CA | Austin, TX
Certified Wedding Photography Junky™
Primary kit - D200, 10.5mm f/2.8D, 35mm f/1.8G, 50mm f/1.4G & 70-300VR
Backup kit – D80, 18-105VR
SB800, SB600 and other misc lighting equipment
Lenses worth mentioning owned and sold– 12-24 f/4, 17-55 f/2.8, 35-70 f/2.8, 80-200 f/2.8, 20mm f/2.8, 35mm f/2, 50mm f/1.8, 50mm f/1.4D, 60mm f/2.8, 85mm, f/1.8. 105mm f/2D-DC, 180mm f/2.8, 300mm f/4D-ED
edwardaneal
Veteran Member
I agree with this, but, if as the OP says, the movement of the elements in VR can degrade an image at high shutter speeds why wouldn't this element movement degrade the image even more at longer exposures?
Wouldn't the elements move even more during a longer exposure?
Personally I don't believe the movement of the elements can be seen by the sensor. Think about it - if the movement of the elements is in perfect sync with the vibration it is counteracting it should be completely invisible to the sensor regardless of shutter speed.
Are there things that can throw it out of sync for a split second? - - - Sure there are, But if used properly and given the time it needs to settle down and sync I just dont see how it could cause degradation to the image
--
Primary kit - D200, 10.5mm f/2.8D, 35mm f/1.8G, 50mm f/1.4G & 70-300VR
Backup kit – D80, 18-105VR
SB800, SB600 and other misc lighting equipment
Lenses worth mentioning owned and sold– 12-24 f/4, 17-55 f/2.8, 35-70 f/2.8, 80-200 f/2.8, 20mm f/2.8, 35mm f/2, 50mm f/1.8, 50mm f/1.4D, 60mm f/2.8, 85mm, f/1.8. 105mm f/2D-DC, 180mm f/2.8, 300mm f/4D-ED
Ronnie Gaubert
Senior Member
Hi Jarno,
Those lines only show up under certain lighting conditions. It mostly has something to do with the type of background elements and the distance that those elements are from the subject.
In this shot you will see the bush directly behind the bird's head is not affected. That bush is only a few feet away. The background area about 5-10 feet in back of the bird is greatly affected by the double lines. Especially the area below the bird. In many of my images that lined affect is not as obvious. That's why I think it has something to do with lighting. Darker areas don't seem to be affected. It seems to be more related to areas with bright highlights. But I'm no expert and don't positively know what causes them. I can say it's not related to VR.
This one with the 400mm Non-VR and a Nikon TC17EII teleconverter.
Date/Time 27-Apr-2008 09:54:41
Make Nikon
Model NIKON D200
Flash Used No
Focal Length 650 mm
Exposure Time 1/500 sec
Aperture f/6.3
ISO Equivalent 320
Exposure Bias -2/3
White Balance
Metering Mode matrix (5)
Raw
Exposure Program aperture priority (3)
Focus Distance
--
http://www.pbase.com/ronnie_14187
It's less about the subject and more about the composition & light- Ronnie Gaubert.
jfriend00
Forum Pro
My previous post (5 posts earlier in this thread) addresses this question. It has to do with the differences/degradations in the image quality as the VR lens moves through it's range. If you believe as I do that the decentered VR lens causes some image degradations in different positions (no moving optical design like this is perfect), then the moving slit captures a potentially different IQ at different locations in the frame (as the slit and VR lens move). At slower speeds when the shutter is open entirely (no moving slit), it at least captures the same IQ everywhere.
--
John
Gallery: http://jfriend.smugmug.com
Popular: http://jfriend.smugmug.com/popular
Portfolio: http://jfriend.smugmug.com/portfolio
RBFresno
Forum Pro
I'm partial to these shots of critters eating other critters!
Regarding the VR.
For me, the biggest disadvantage of turning VR off, is that I forget to turn it back on when I really need it. I've ruined a number of shots by having the VR off when I really needed it.
Best Regards,
Bob
http://www.pbase.com/rbfresno/profile
JC1306
Senior Member
That's not what I said and it's not, that's precisely the problem. The sampling speed of the VR system is fixed at 1/1000s according to Nikon. So you press the shutter and the VR system takes 1/1000s to analyze movement, plus a little bit more to actually compensate for it. So if the exposure is 1/1000s or faster the VR system will only begin to respond after the picture has been taken. And since at those shutterspeeds you're not likely to get blurred pictures anyway you might as well turn VR off.
Jarno
--
˙ǝɹnʇɐuƃıs sıɥʇ pɐǝɹ oʇ ƃuıʎɹʇ pɐǝɥ ɹnoʎ ʇןıʇ noʎ uǝɥʍ ʎuunɟ sʞooן ʇı
hoof
Senior Member
For me, I don't have the discipline to remember to turn on VR when I need it every time I need it. Thus, the trade-off of leaving it on is a small chance of the image being degraded due to decentering when I don't need the VR in exchange for the VR being on when i need it and the shot would otherwise be ruined due to shake.
Karl Grobl, who shoots photos professionally all over the world, literally tapes his IS on his canon lenses to on to prevent this very problem:
http://karlgrobl.com/EquipmentReviews/ThrashedCanons.htm
In addition, even at high shutter speeds, if the camera is in the apex of it's vibration (when it's moving the image the most), the VR can still help. This is because the VR is also moving the image in sync with this motion. Thus, for lower-frequency vibrations (under 1/500th of a second if the sampling is 1/1000th), the VR can still help, usually on the 1-2 pixel blur level (or however far the image would move in the most violent part of the handshake in that timeframe).
Oh, and you can guarantee all vibrations are under this threshold (1/500th of a second), otherwise high-speed shots would be blurry. Think about that for a moment. If the frequency is low, the VR can easily keep up. If the frequency is high, then VR can't help, but neither will fast shutter speeds. Since fast shutter speeds work, we know the frequencies of the vibrations are easily handled and tracked by the VR.
If you've seen Mythbusters with their slow-mo camera, notice how a bullet still is moving at high speeds. If the camera was panning in sync with the bullet (which is what VR does), then the bullet would remain stationary, even with the high-speed camera. Thus, even at high shutter speeds, VR can still help keep the image just a bit sharper.
For more evidence, take a fast shutter speed while panning very fast. Notice how blur is still visible. VR can help reduce this, even at high shutter speeds.
That said, decentering issues with the VR (as that's how it works) can degrade the image, just as the OP shows. But for me, the risk of having VR off when I need it, combined with the slight IQ improvement if my hands are really shakey and the shutter catches the shake at it's maximum, mean I leave VR on all the time, even at the cost of one edge being degraded due to VR decenteriing.
Karl Grobl, who shoots photos professionally all over the world, literally tapes his IS on his canon lenses to on to prevent this very problem:
http://karlgrobl.com/EquipmentReviews/ThrashedCanons.htm
In addition, even at high shutter speeds, if the camera is in the apex of it's vibration (when it's moving the image the most), the VR can still help. This is because the VR is also moving the image in sync with this motion. Thus, for lower-frequency vibrations (under 1/500th of a second if the sampling is 1/1000th), the VR can still help, usually on the 1-2 pixel blur level (or however far the image would move in the most violent part of the handshake in that timeframe).
Oh, and you can guarantee all vibrations are under this threshold (1/500th of a second), otherwise high-speed shots would be blurry. Think about that for a moment. If the frequency is low, the VR can easily keep up. If the frequency is high, then VR can't help, but neither will fast shutter speeds. Since fast shutter speeds work, we know the frequencies of the vibrations are easily handled and tracked by the VR.
If you've seen Mythbusters with their slow-mo camera, notice how a bullet still is moving at high speeds. If the camera was panning in sync with the bullet (which is what VR does), then the bullet would remain stationary, even with the high-speed camera. Thus, even at high shutter speeds, VR can still help keep the image just a bit sharper.
For more evidence, take a fast shutter speed while panning very fast. Notice how blur is still visible. VR can help reduce this, even at high shutter speeds.
That said, decentering issues with the VR (as that's how it works) can degrade the image, just as the OP shows. But for me, the risk of having VR off when I need it, combined with the slight IQ improvement if my hands are really shakey and the shutter catches the shake at it's maximum, mean I leave VR on all the time, even at the cost of one edge being degraded due to VR decenteriing.
Windancer
Veteran Member
..... for the reasons you stated.
--
Graham Fine Art Photography
http://grahter.sasktelwebsite.net
http://gallery.reginaphotoclub.com/TGraham
Photography on the North American prairies & plains:
http://groups.yahoo.com/group/PrairiePhoto/
Disclaimer: This e-mail is intended to impart a sense of humor. Given e-mail's inability to carry inflections, tone and facial expressions it may fail miserably in its intent. The sender acknowledges the limitations of the technology and assigns to the software in which this message was composed any ill feelings that may arise.
edwardaneal
Veteran Member
so if I understand you correctly at faster shutter speeds the VR only degrades some areas of the picture because of the shutter being a small slit
but at slower shutter speeds it degrades the entire image equally?
See, my point is that the element is supposed to move exactly the correct direction and amount to be the exact opposite of the camera shake.
if it it is doing this properly it should appear at the sensor as if all elements are perfectly aligned and nothing is moving - this is what VR does. If as you say the movement of the elements isnt perfectly in sync because nothing can be perfect then all images taken with VR would suffer from some degredation
I find it hard to accept that in lenses costing as much as nikon's long VR lenses cost that Nikon would find this acceptable especially when as you point out these lenses are normally used on a tripod.
Sorry - I just think you are mistaken and I also think there are too many factors that could have caused issues in your sample photos to arive at any kind of reliable conclusion.
Nothing personal - it's just my opinion
Primary kit - D200, 10.5mm f/2.8D, 35mm f/1.8G, 50mm f/1.4G & 70-300VR
Backup kit – D80, 18-105VR
SB800, SB600 and other misc lighting equipment
Lenses worth mentioning owned and sold– 12-24 f/4, 17-55 f/2.8, 35-70 f/2.8, 80-200 f/2.8, 20mm f/2.8, 35mm f/2, 50mm f/1.8, 50mm f/1.4D, 60mm f/2.8, 85mm, f/1.8. 105mm f/2D-DC, 180mm f/2.8, 300mm f/4D-ED
but at slower shutter speeds it degrades the entire image equally?
See, my point is that the element is supposed to move exactly the correct direction and amount to be the exact opposite of the camera shake.
if it it is doing this properly it should appear at the sensor as if all elements are perfectly aligned and nothing is moving - this is what VR does. If as you say the movement of the elements isnt perfectly in sync because nothing can be perfect then all images taken with VR would suffer from some degredation
I find it hard to accept that in lenses costing as much as nikon's long VR lenses cost that Nikon would find this acceptable especially when as you point out these lenses are normally used on a tripod.
Sorry - I just think you are mistaken and I also think there are too many factors that could have caused issues in your sample photos to arive at any kind of reliable conclusion.
Nothing personal - it's just my opinion
--
Primary kit - D200, 10.5mm f/2.8D, 35mm f/1.8G, 50mm f/1.4G & 70-300VR
Backup kit – D80, 18-105VR
SB800, SB600 and other misc lighting equipment
Lenses worth mentioning owned and sold– 12-24 f/4, 17-55 f/2.8, 35-70 f/2.8, 80-200 f/2.8, 20mm f/2.8, 35mm f/2, 50mm f/1.8, 50mm f/1.4D, 60mm f/2.8, 85mm, f/1.8. 105mm f/2D-DC, 180mm f/2.8, 300mm f/4D-ED
Sagittarius
Veteran Member
Hi,
VR is engaged when release button is half pressed. Shutter released at full press. I do not think that the lag between half press and full press is less than 1/1000s. Acoording to Nikon, VR has two algorithms: one for viewfinder and second one which is stronger for the sensor. Also the instant shuter is released, the VR lens is moved to the center. I think this movement starts after release button passed half way point but before shutter has been released. Now, if VR affects bokeh, why it affects background only and not the hole image?
--
Best regards
hoof
Senior Member
Not true. You do realize it takes a bit of time for the mirror to get out of the way, and settle (it bounces) before the shutter is opened? This can take 30-40ms depending on the design (the same time as a 1/30th of a second shutter speed). VR has loads of time (30-40 samples) to do its thing in this time, before the shutter can open. A visual example of this in action:
http://www.canonrumors.com/2009/07/5d-shutter-in-slow-motion/
In addition, if you go from undepressed to fully depressed, it's likely the camera will take a short time to autofocus before triggering the picture to be taken (otherwise the picture could be out of focus). Since it needs the mirror down to do this, this can make it take even longer (possibly 1/2 a second or longer if the lens is at minimum focus and the target is at infinity). That's why action shooters push halfway and hold, as it minimizes the time between fully depressing the shutter and when the picture actually gets taken.
If you're trying to take pictures on a hair-trigger, always always always halfway depress the shutter before the picture and hold it until the precise time. Not only does this minimize the time before the picture is taken, but both AF and VR have enough time to be fully 100% engaged and set before the picture fires. As a side bonus, you minimize shutter-depress camera shake, as you're not stabbing the shutter button, but gently pressing it a bit more.
I almost never go from undepressed to fully depressed when shooting. I don't trust the AF to work properly in this situation, so I want to make sure it gets proper focus before triggering the shot. People who stab the shutter from fully up to down simply aren't maximizing their camera's potential, and I have no sympathy if their images are lower quality, whether it's due to VR settle time or AF issues. One should learn how to use one's equipment if maximum IQ is important.
edwardaneal
Veteran Member
I just found this little writeup from Nikon here
http://www.nikon.com/about/technology/core/software/vr_e/index.htm
it says the system detects vibration every 1/1000th of a second
that is much higher sampling rate than the 1/500th you suggested earlier
now this part of what it says is important to what you are saying
"What's more, just before exposure, the VR lens will reset to central position (optical axis) from an off-centered position which is a result of VR operation during the shutter release button is lightly pressed . Since the shift amount of the VR lens is limited, this operation maximizes VR effects as well as optical performance. Only Nikon has this "Centering Before Exposure" feature. "
if correct this would seem much more reasonable to me as de-centering would not be a huge issue as you suggest and even if it were the sampling rate of 1000 times per second should keep things from degrading up to 1/1000sec shutter speeds and not as low as 1/500th as you suspect
--
Primary kit - D200, 10.5mm f/2.8D, 35mm f/1.8G, 50mm f/1.4G & 70-300VR
Backup kit – D80, 18-105VR
SB800, SB600 and other misc lighting equipment
Lenses worth mentioning owned and sold– 12-24 f/4, 17-55 f/2.8, 35-70 f/2.8, 80-200 f/2.8, 20mm f/2.8, 35mm f/2, 50mm f/1.8, 50mm f/1.4D, 60mm f/2.8, 85mm, f/1.8. 105mm f/2D-DC, 180mm f/2.8, 300mm f/4D-ED
http://www.nikon.com/about/technology/core/software/vr_e/index.htm
it says the system detects vibration every 1/1000th of a second
that is much higher sampling rate than the 1/500th you suggested earlier
now this part of what it says is important to what you are saying
"What's more, just before exposure, the VR lens will reset to central position (optical axis) from an off-centered position which is a result of VR operation during the shutter release button is lightly pressed . Since the shift amount of the VR lens is limited, this operation maximizes VR effects as well as optical performance. Only Nikon has this "Centering Before Exposure" feature. "
if correct this would seem much more reasonable to me as de-centering would not be a huge issue as you suggest and even if it were the sampling rate of 1000 times per second should keep things from degrading up to 1/1000sec shutter speeds and not as low as 1/500th as you suspect
--
Primary kit - D200, 10.5mm f/2.8D, 35mm f/1.8G, 50mm f/1.4G & 70-300VR
Backup kit – D80, 18-105VR
SB800, SB600 and other misc lighting equipment
Lenses worth mentioning owned and sold– 12-24 f/4, 17-55 f/2.8, 35-70 f/2.8, 80-200 f/2.8, 20mm f/2.8, 35mm f/2, 50mm f/1.8, 50mm f/1.4D, 60mm f/2.8, 85mm, f/1.8. 105mm f/2D-DC, 180mm f/2.8, 300mm f/4D-ED
JC1306
Senior Member
Sorry but that's not what I was talking about, it's about the effectiveness of VR during the exposure. The VR system checks for movement 1000 times per second and then still needs to act upon the information it receives. This means that even under the best of circumstances there is at least 1/1000s between two measurements of the VR system. So when you're using a shutter speed approaching 1/1000s or faster you get the following timeline: mirror is fully up - VR takes measurement - VR element shifts - exposure starts - exposure ends - VR takes next measurement - mirror starts moving down... As you can see the VR system did exactly nothing during the exposure so you might as well leave it off. VR is for long exposures, not short ones.
BTW, if camera shake is still an issue at 1/1000s or faster you might consider drinking less coffee...
Jarno
--
˙ǝɹnʇɐuƃıs sıɥʇ pɐǝɹ oʇ ƃuıʎɹʇ pɐǝɥ ɹnoʎ ʇןıʇ noʎ uǝɥʍ ʎuunɟ sʞooן ʇı
BTW, if camera shake is still an issue at 1/1000s or faster you might consider drinking less coffee...
Jarno
--
˙ǝɹnʇɐuƃıs sıɥʇ pɐǝɹ oʇ ƃuıʎɹʇ pɐǝɥ ɹnoʎ ʇןıʇ noʎ uǝɥʍ ʎuunɟ sʞooן ʇı
em_dee_aitch
Senior Member
Panning the subject matter over a series of shots, not panning the lens during the shot . Standard method for distant landscape panoramas.
--
David Hill
http://www.sfbayweddingphotographer.com
San Francisco, CA | Austin, TX
Certified Wedding Photography Junky™
jfriend00
Forum Pro
I think you haven't yet understood the point I'm making. Let's try this in smaller steps.
1) We agree that the movement of the VR lens element is trying to counteract lens movement.
2) We can probably agree that when the VR lens is in the center of it's range, it has it's best optical performance in all aspects of the image.
3) Would you agree that something in the IQ will degrade when the VR lens gets decentered? That it's not possible maintain all aspects of IQ and decenter the VR lens - that some aspect of IQ will not be as good as when the lens is decentered?
4) If you agree on the previous point, then the issue of a moving slit while there's a moving VR lens just means that one side of the image is captured with the VR lens in position 1 and the other side of the image is captured with the VR lens in position 2. Since position 1 and position 2 have different IQ, the two parts of the image are captured with different IQ. For purposes of this discussion, we can assume that the VR lens is perfect for removing vibration so there is no movement of the image on the sensor between position 1 and position 2. But, if position 1 has the VR lens in the center and position 2 has the lens off center, something in the IQ will be more degraded in position 2 than position 1 and thus one side of the image will have different IQ than the other. What exact IQ characteristic is degraded depends upon the lens design. I've seen the quality of bokeh degraded in some VR shots, but it could also be CA or edge sharpness or edge distortion, etc...
--
John
Gallery: http://jfriend.smugmug.com
Popular: http://jfriend.smugmug.com/popular
Portfolio: http://jfriend.smugmug.com/portfolio
jfriend00
Forum Pro
I wonder how Nikon handles shots at 8fps? Can they recenter the VR lens before every frame? The instances where I've seen VR IQ degraded where all while shooting sports at 8fps at shutter speeds of 1/1000 and up.
--
John
Gallery: http://jfriend.smugmug.com
Popular: http://jfriend.smugmug.com/popular
Portfolio: http://jfriend.smugmug.com/portfolio
em_dee_aitch
Senior Member
I think in the above arguments you are assuming that the correction BR provides is perfect, and also neglecting the window of opportunity that might exist between the sampling limit of VR (1/500sec) and the max stopping power of the shutter (1/8000sec).
An important distinction with regard to that statement: VR does not pan in sync with a moving subject, because it does not detect the subject. It only attempts to offset camera motion. I'm not sure of your exact point here, but using the wrong VR mode (i.e. Active) during a panning shot could actually hurt your shot as opposed to help it, as it would effectively be attempting to stabilize the background and not the moving subject.
Same point about panning applies, but also: I think you're assuming that there is not a margin of error within the VR correction itself which could be recorded as degradation, and that said margin of error could be larger than the stopping power provided by 1/8000, which could help explain the kinds of oddities seen in my shots.
--
David Hill
http://www.sfbayweddingphotographer.com
San Francisco, CA | Austin, TX
Certified Wedding Photography Junky™
em_dee_aitch
Senior Member
I don't think that BIF is good subject matter for revealing this, nor are shots of people, because the outer 50 percent (or more) of the frame may all be out of focus area.... Just a thought.... I think that is probably why I have never noticed this much in my regular work and only had an "ah ha" moment on a landscape.
--
David Hill
http://www.sfbayweddingphotographer.com
San Francisco, CA | Austin, TX
Certified Wedding Photography Junky™
JC1306
Senior Member
Yes they do, somewhat: http://www.nikon.com/about/technology/core/software/vr_e/index.htm
At 1/500s VR is certainly effective but at 1/1000 less so. It's simple mathematics: as stated on the Nikon website the VR system takes measurements 1/1000s apart so it cannot respond in less time than 1/1000s. When you're using an exposure time of 1/1000s or faster the shutter will open and close before the VR system can even determine which way to shift the correcting element, let alone actually move it.
Jarno
--
˙ǝɹnʇɐuƃıs sıɥʇ pɐǝɹ oʇ ƃuıʎɹʇ pɐǝɥ ɹnoʎ ʇןıʇ noʎ uǝɥʍ ʎuunɟ sʞooן ʇı
Keyboard shortcuts
- f
- Forum