Low light + 85mm 1.8 ---> 1/160 not enough?

KnightRT wrote in post #14917753
Printing is immaterial to how the image is recorded.

That much is true.

What is "blur at the pixel level"? Blur can only be observed in images, not in pixels, and that requires some physical manifestation, print or display, and that menas it will have a physical size. That physical size is reached by magnifying the image captured by the sensor (whatever its resolution or the number of pixels present) by a factor that depends on the sensor size. Thus, images from smaller sensors will be magnified more, so any blur present will be more apparent.

Part of that makes some sense, I guess, but it would only apply if you routinely discarded the parts of your FF image that wouldn't fit on an APS-C sensor.

That's not how the real world works though. On different formats, different focal lengths are used to achieve comparable framing. Do that, and you need to magnify by different amounts to get to the final print/display size, and that's where subject blur is magnified.

If you don't do that, and just crop the larger image to match an APS-C frame, then the magnification and blur would be the same, as you say. That's the part that makes sense (in a limited way, given that you've paid the FF price, but are only ever using APS-C capabilities). However, since you would necessarily have to throw away the content that would have made the scenes different, that part doesn't follow.

TeamSpeed wrote in post #14916793
For example, you would think 1/250th would be able to stop a slow dance, but here is a shot I screwed up on because just her hand moving through the air slowly wasn't captured very crisply at 1/250th. Even at 1/500th it was questionable, so 1/160th shooting at people who move could very well be a recipe for disaster if you want crisp shots.

If you are observant and have the opportunity to time your shot, then you can get away with slower shutter speeds then would normally be deemed sufficient. Unfortunately I really do not want to be limited by waiting for "peak" moments and, maybe, miss a better shot.

The following were shot at 1/250s. In the first, the front foot is blurred, in the second only her core is not blurred but I think it works. It doesn't always have to be blur/motion free.

I needed no blur for what I was shooting (per the parents) , and I also had no idea what to expect from this Indian dance. Having gone through this first one, I will know more of what to expect the next time. Most of my shots (95+%) don't have any blur, but were shot at higher speeds.

The point is that 1/160th isn't always enough to stop movement of people, no matter how "fast" that speed seems and how one would "think" it is enough to stop the movement.

Great shots, and it shows you can definitely use blur for artistic shots!

Motion blur is indeed very nice to keep a sens of motion. However you need to master the shutter speed, get some experience with your subject or simply be lucky.

For exemple, a pan shot of a horse trotting... 1/30 will keep the leg blurry while the body will be sharp.
1/60 to nicely blur an helicopter propellor.

Anyway, in spite of all that semantics, you need to include the 1.6x factor when calculating your shutter speed. Consider a 85mm lens a 136mm one and raise your shutter speed to 250 or 320.

Earwax69 wrote in post #14919559
Anyway, in spite of all that semantics, you need to include the 1.6x factor when calculating your shutter speed. Consider a 85mm lens a 136mm one and raise your shutter speed to 250 or 320.

This is only a guideline for preventing handshake where you don't have IS. This has nothing to do with freezing motion.

For my examples, I had to go up to around 1/800th or faster, but that required ISO 12800, which is nothing I am afraid of.

The lighting was a bit rough, here was 1/400th at ISO 12800, and it happened to be enough simply because I must have hit the young lady mid-move, but her hands were still moving, creating a bit of blur.

HTTP response: 404 | MIME changed to 'text/html' | Byte size: ZERO

xarqi wrote in post #14919055
On different formats, different focal lengths are used to achieve comparable framing. Do that, and you need to magnify by different amounts to get to the final print/display size, and that's where subject blur is magnified.

With equivalent framing, bodies of equal resolution will have the same tolerance to handshake regardless of crop factor or pixel size. You're compensating for a change in pixel size with a change in focal length. With the same focal length, bodies with the same pixel size will have the same tolerance to handshake, albeit with different framing.

In more concrete terms, you can shake 60% more with a 20D shot at 200mm than a 5D II at 320mm. The framing will be the same, but the 5D II has 60% more linear resolution to record your movements. Same with the 7D; if you're using a shorter focal relative to the 5D II to keep the framing, the 5D II will be 8% less tolerant of handshake.

OK - I'll go one more round, but that's it.

KnightRT wrote in post #14920081
With equivalent framing, bodies of equal resolution will have the same tolerance to handshake regardless of crop factor or pixel size.

Let's see if it's possible to dissect out the red herrings here.
Equivalent framing on different formats implies different focal lengths. That will imply a difference in susceptibility to subject blur as the magnifications at the sensor will be different, with the larger sensor needing the greater focal length, thus emphasising blur.
Now add in the post-processing greater magnification needed by the smaller sensor, and hey presto, the two effects exactly cancel. Resolution and pixel size have no bearing on this and are extraneous elements of your statement.

Not so. Pixel size is not a factor.

Oi! Ignoring the spurious mention of pixel size what's left. Same focal length, different formats implies different compositions. This is the situation where you are driving your FF body as if it were an APS-C and rountinely throwing away much of the data. I've covered that edge case in an earlier post. That's just not how the real world works.

No - you've almost completely lost me there, but I can assure you that linear resolution has no bearing on the matter at all. At best, it may dictate how well the blur that is present is rendered.

Over and out.

Yes indeed, nothing to do with freezing subject motion. However this thread was originally about my poor shaky hands...

Let's just work on this. It's the clearest application, if you're not getting it, the rest of it won't make sense. I'll break it apart:

> In more concrete terms, you can shake 60% more with a 20D shot at 200mm than a 5D II at 320mm.

If the 20D has a 200mm lens and the 5D2 has a 320mm lens, you've got exactly the same angle of view. But the 20D only has 8MP to represent that image relative to the 5D2's 22MP. Movement affects more pixels on the 5D2 because the same physical movement is recorded with more pixels.

Your argument was that, because we're printing the scene at the same physical size, the blur will look the same in the end result. I agree. But the point of having more megapixels is to print larger. If we print with the same pixel density, or merely view both images on a monitor at 100%, it will be obvious that the 5D2 has more blur and requires a higher shutter speed.

How much higher? The horizontal dimension of the 5D2 frame is about 5600 pixels. The 20D, about 3500. Divide the first by the second to get 1.6; for every 1 pixel affected on the 20D, 1.6 will be affected on the 5D2. For a movement at a given speed, the solution to match the 20D's pixel-level blur is to boost the shutter by 60%.

The point of all this is that pixel size (or pixel density if you're prefer that phrase) is the ultimate arbiter of handshake. There's no single rule that works for all bodies. I brought up the 20D specifically because it was a case where the pixel density was so low that it was equally sensitive or less sensitive to movement than current full-frame DSLRs, regardless of the 'crop factor' rule.

Earwax69 wrote in post #14920339
Yes indeed, nothing to do with freezing subject motion. However this thread was originally about my poor shaky hands...

Okay, but that was hard to determine because interspersed through this thread, and in the original post, terms are thrown around like motion blur, then elsewhere hand shake, and the guidelines of shutter speeds of 1 / (fl * crop factor) only helps hand shake.

As to how to do that shot later and fear of higher ISOs, actually go up to 6400 or even 12800, get your shutter speed where you want it, then if the image is too bright, actually take the exposure slider in DPP down to where you want the photo to be in the darker area. This will give you a crisp shot, and low ISO noise, instead of a blurry shot and slightly lower ISO.

I always shoot to the right (ETTR or even HAMSTTR), so I would rather shoot at a higher ISO and push the exposure down if needed, than to shoot at lower ISOs to the left, and bring the exposure up later. Less noise when ETTRing than when ETTLing.