So so disheartened with my studio shots. Cannot seem to eliminate blur. Please help.

[HoppyUK]

The Flash duration is shown in T5 times. This is a higher number than the actual usable flash duration, known as T1 time. To get the proper flash speed in use take the T5 time and as a rough calculation, divide by 3. This means that the Bxri has a full flash duration (shutter speed) of around 1/500th of a second. This is where things take another turn.

With studio flash heads, the flash duration is generally fastest at full power. As you decrease power the flash duration gets slower. So by 1/8 power the duration has become substantially longer, hence why you say the Brxi is not freezing action.

Speedlites on the other hand are the opposite. The flash duration gets faster as the power goes down, so it adds another layer of confusion. When using speedlites they tend to have fast durations anyway, but the less power, the faster the duration.

Estimating the true action-stopping potential of flash units is difficult. Multiplying t.5 times by 3x is only approximate, hardly any manufacturers quote t.1 times and they tend to under-estimate in my experience. Frankly, the only way of doing it is to compare to actual shutter speeds and for the Elinchrom BX-500-Ri I put that at 1/900sec at full power, to 1/700sec at minimum (1/16th). The BX-250-Ri has smaller capacitors that dump faster and that gave 1/1300sec at max to 1/900sec at 1/16th. Those numbers are fast for studio heads that don't claim to be 'fast' and most similarly priced/powered units are around double those durations. Elinchrom's t.5 figures for the BX-500-Ri are 1/1558sec at full power and 1/1395sec at 1/2 power; BX-250-Ri 1/2762sec and 1/2165sec.

Speedlite durations are different again. At full power they behave like studio heads, but as power is reduced not only does the flash duration get very much shorter, t.5 times and t.1 times are effectively the same and both are very close to actual shutter speed equivalents.

 

[sk66]

So here's a scenario...

I regularly shoot indoor kickboxing. My speeds are never high enough to freeze the action really...even at ISO 12800.

If I got an sb910, could I use this to freeze a kick that would usually require about 1/1000th with 1/15th and use a low aperture & reasonable iso?

It's very hard to freeze motion above ~ 1/250 effective..that's the fastest SS most camera's will do with flash, and it's the fastest T.1 time for most flashes. To get a faster T.1 time from a speedlight you have to reduce it's power and it has to be the primary light source. That means you have to need less flash which means you have more ambient to record blur with. (flash is always 2 exposures recorded simultaneously)... There are ways of doing it, but not generally practical in a kickboxing environment.

http://photographic-academy.com/lighting/86-lighting/138-flash-doesn-t-freeze-motion

 

[sk66]

Speedlite durations are different again. At full power they behave like studio heads, but as power is reduced not only does the flash duration get very much shorter, t.5 times and t.1 times are effectively the same and both are very close to actual shutter speed equivalents.

I've never heard/read that t.5 and t.1 are very much the same for speedlights. In fact, I've read the opposite, supposedly with test results backing it up. I've never done any testing myself, so I'm not saying one way or the other, but my experience w/ high speed subjects and speedlights suggests the T.5 isn't the same as the T.1 time for them... not even really close from what I would expect if they were.

 

[Phil Young]

It's very hard to freeze motion above ~ 1/250 effective..that's the fastest SS most camera's will do with flash, and it's the fastest T.1 time for most flashes. To get a faster T.1 time from a speedlight you have to reduce it's power and it has to be the primary light source. That means you have to need less flash which means you have more ambient to record blur with. (flash is always 2 exposures recorded simultaneously)... There are ways of doing it, but not generally practical in a kickboxing environment.

http://photographic-academy.com/lighting/86-lighting/138-flash-doesn-t-freeze-motion

.....using the advice given in this thread, yesterday I froze my wife shaking her hair at 1/25th and the flash was on full power...

I think I understand though - the flash duration becomes the shutter speed and based on what's been said above, divide it by 3.

So if I need a 1/1000th shutter for example then I reduce the flash to whatever is specified to 1/3000th and just add more units to get enough power at that duration...

Stand to be corrected if wrong but that's now my understanding.

 

[EOS_JD]

Just realised I hadn't actually considered the power setting! I was probably shooting at half to one quarter power so flash duration was just too slow - I should have used my Quadras

 

[HoppyUK]

It's very hard to freeze motion above ~ 1/250 effective..that's the fastest SS most camera's will do with flash, and it's the fastest T.1 time for most flashes. To get a faster T.1 time from a speedlight you have to reduce it's power and it has to be the primary light source. That means you have to need less flash which means you have more ambient to record blur with. (flash is always 2 exposures recorded simultaneously)... There are ways of doing it, but not generally practical in a kickboxing environment.

http://photographic-academy.com/lighting/86-lighting/138-flash-doesn-t-freeze-motion

The flash durations quoted on the link for a Nikon SB800 cannot be converted to t.1 just by multiplying by three. IGBT guns don't behave like studio heads below full power. See post #121 above

.....using the advice given in this thread, yesterday I froze my wife shaking her hair at 1/25th and the flash was on full power...

I think I understand though - the flash duration becomes the shutter speed and based on what's been said above, divide it by 3.

So if I need a 1/1000th shutter for example then I reduce the flash to whatever is specified to 1/3000th and just add more units to get enough power at that duration...

Stand to be corrected if wrong but that's now my understanding.

Yes, you've got the general idea But not quite, eg you would get close to 1/1000sec effective shutter speed equivalent at half power with a typical hot-shoe gun.

Most hot-shoe guns are quite similar and commonly quote a flash duration at full power of around 1/1,000sec. This is t.5, and equates to a shutter speed equivalent of around 1/400sec in my testing. When the power is turned down though, things change a lot. At half power, the effective flash duration will be around 1/1,000sec, at quarter power 1/2,000sec, at 1/8th power 1/4,000sec and this linear halving relationship holds good, more or less, down to perhaps 1/30,000-1/40,000sec or thereabouts at 1/128th minimum power.

 

[Garry Edwards]

I've never heard/read that t.5 and t.1 are very much the same for speedlights. In fact, I've read the opposite, supposedly with test results backing it up. I've never done any testing myself, so I'm not saying one way or the other, but my experience w/ high speed subjects and speedlights suggests the T.5 isn't the same as the T.1 time for them... not even really close from what I would expect if they were.

The discharge of (any) electronic flash roughly follows a bell curve. The flash very quickly reaches full power, then the power gradually declines.
The t.5 time represents the amount of time that it takes for half (50%) of the flash energy to be discharged.
The t.1 time represents the amoujt of time that it takes for 10% of the flash energy to be discharged.
Now, with a hotshoe flashgun (or an igbt studio flash head such as the SuperFast) the full power discharge follows the typical bell chart shape.
But as you reduce the power setting, instead of actually reducing the power, the flash is still firing at full power but is quenched, or cut off early. Therefore, the graph shows the initial rise to full power, but not the fall, and the t.1 time replaces the t.5 time. They are, in effect, the same thing, especially at lower power settings

 

[sk66]

instead of actually reducing the power, the flash is still firing at full power but is quenched, or cut off early. Therefore, the graph shows the initial rise to full power, but not the fall, and the t.1 time replaces the t.5 time. They are, in effect, the same thing, especially at lower power settings

Makes complete sense...
TBH, I took/based some of the information from other resources that seemed factually based and with empirical test data... But I suppose testing methods etc etc...

 

[sk66]

This is one example of results I've seen for a speedlight/IGBT (SB800 in this case).
1/1 -- 257.73 -- 825.08
1/2 -- 759.3 -- 821.02
1/4 -- 1615 -- 1797.3
1/8 -- 2979.7 -- 3620.6
1/16 -- 5112.5 -- 7363.8
1/32 -- 7874 -- 13351
1/64 -- 11338 -- 22371
1/128 -- 13661 -- 29674

certainly not a factor of 3 though...

 

[HoppyUK]

This is one example of results I've seen for a speedlight/IGBT (SB800 in this case).
1/1 -- 257.73 -- 825.08
1/2 -- 759.3 -- 821.02
1/4 -- 1615 -- 1797.3
1/8 -- 2979.7 -- 3620.6
1/16 -- 5112.5 -- 7363.8
1/32 -- 7874 -- 13351
1/64 -- 11338 -- 22371
1/128 -- 13661 -- 29674

certainly not a factor of 3 though...

Steven, as I've already said...

The rule of thumb is that t.5 times for studio-type heads should be increased by 3x to get actual shutter speed equivalents, and that will also be close to t.1 times. I disagree with that, and my own estimates are that a conversion factor for t.5 at full power of between 2x and 3x will get close to actual shutter speed. But at lower power settings this will under-estimate shutter speed equivalents. T.1 times are very rarely quoted and IMHO they only serve to muddy the waters further. Bottom line - both t.5 and t.1 are almost useless, and it is high time that the industry came up with something better, but since that would make everything look much slower, it won't happen.

Nikon SB800 is an IGBT-controlled hot-shoe gun (obviously!). As the graphs show, it behaves quite differently to conventional studio-type heads and as I said above, in the mid-range below full power both the t.5 and t.1 times become very similar, and both are close to real shutter speed equivalents. The graphs are also pretty close to the rough estimates of flash durations I gave for typical hot-shoe guns. I don't know how accurate the graphs are; I expect good down to about 1/16th power but I have to say I'm sceptical about most oscilloscope traces at very short flash durations. They're there, or thereabouts, though.

Do you have a link to the source please? Be interested to see what else might be there

 

[sk66]

I have to agree that the given numbers, and estimating/converting them is very inaccurate... The only real purpose it serves it for a general understanding of what's happening...

Those numbers and graph was just the first I found in a quick search and came from flickr (the image is linked) edit: should have been linked HERE). A more reliable source is probably Speedlights.net. The link I provided is to their SB700 test/review and these are the durations reported (power/spec t.5/ tested t.1).

1/1 1/1042 1/305
1/2 1/1136 1/1000
1/4 1/2857 1/2300
1/8 1/5714 1/4000
1/16 1/10000 1/5250
1/32 1/18182 1/8000
1/64 1/25000 <1/8000
1/128 1/40000 <1/8000

 

[sk66]

It makes sense that the t.1 times start to spread at lower power settings seeing that the graphs climb hard to max brightness... at lower power settings the hard climb and the hard fall become more equally part of the duration.
It doesn't seem there is any clear way to even estimate t.1 for an IGBT...

 

[HoppyUK]

Thanks for the link Steven

 

[coldpenguin]

Just one other idea which doesn't appear to have been explicitly said. It might be worthwhile not using 160 for the ISO, but 100 or 200. I am not sure about the Nikons, but it used to be with the cheap Canons that 160 was using the ISO 100 circuitry with gain added. This caused images to be noisier (which could be mis-interpreted as blur if the camera or jpeg export is auto-smoothing.

 

[Richard King]

learn to shoot still things first.

i.e. learn the aperture for each lens you haver that has the best IQ, learn how to light the background and the object separately, learn how to shoot a small, large and medium object sharply and with the correct exposure. figure out what DOF you have at what apeture in your working space

Rule things out, so to start off with, use a tripod, use the timer on your camera, use a single flash head, use manual focus (use fully zoomed in live view to focus on still objects)

Learn the difference shutter speed and flash power ( does and doesn't) have on a moving object (say a pendulum, or a fan)

 

[JonathanRyan]

...Ken Rockwell was right lol

Wash your mouth out.....

 

[sk66]

.....using the advice given in this thread, yesterday I froze my wife shaking her hair at 1/25th and the flash was on full power...

I think I understand though - the flash duration becomes the shutter speed and based on what's been said above, divide it by 3.

So if I need a 1/1000th shutter for example then I reduce the flash to whatever is specified to 1/3000th and just add more units to get enough power at that duration...

Stand to be corrected if wrong but that's now my understanding.

Sounds right to me... (if using studio strobes, more complicated w/ IGBT/Speedlights).

 

[coldpenguin]

learn to shoot still things first.

i.e. learn the aperture for each lens you haver that has the best IQ,

+1 Although usually it is thought that decreasing the aperture (i.e. smaller hole) gives you a sharper image as the DOF should in theory increase, there does come a point where it can make the image fuzzier. Someone on here described it as the circle of confusion.
Try taking your shots at f/8 for a while and see what happens? Reduce your flash/shutter time to compensate of course (if dropping the ISO to 100 isn't enough).

http://en.wikipedia.org/wiki/Circle_of_confusion

 

[Phil V]

... Reduce your flash/shutter time to compensate of course (if dropping the ISO to 100 isn't enough).

That's about the most complicated instruction you could have written given the discussions above, alternatively it's complete nonsense, unless you can give a concise and simple instruction to the OP for reducing (studio) flash duration that's not already been covered.

 

[coldpenguin]

That's about the most complicated instruction you could have written given the discussions above, alternatively it's complete nonsense, unless you can give a concise and simple instruction to the OP for reducing (studio) flash duration that's not already been covered.

Fair enough, just wanted to make sure that the OP didn't just reduce the aperture number and then wonder why everything was so bright.

 

[Garry Edwards]

+1 Although usually it is thought that decreasing the aperture (i.e. smaller hole) gives you a sharper image as the DOF should in theory increase, there does come a point where it can make the image fuzzier. Someone on here described it as the circle of confusion.
Try taking your shots at f/8 for a while and see what happens? Reduce your flash/shutter time to compensate of course (if dropping the ISO to 100 isn't enough).

http://en.wikipedia.org/wiki/Circle_of_confusion

No, the circle of confusion is best described as a group of photographers talking about depth of field
What you're talking about is actually diffraction limitation, which the same photographers usually refer to diffraction.
Diffraction limitation starts to raise its ugly head at apertures smaller than about f/11 on APS-size sensors, f/16 on full frame sensors

 

[sk66]

Diffraction limitation starts to raise its ugly head at apertures smaller than about f/11 on APS-size sensors, f/16 on full frame sensors

That used to be a pretty good rule of thumb; not so much anymore.
D800 ~ f/5.6, D7200 ~ f/5.6, 7D ~ f/8, etc..

That said, I'll push well into the diffraction limits if necessary (due to lens sharpness/DOF requirements).

 

[HoppyUK]

No, the circle of confusion is best described as a group of photographers talking about depth of field
What you're talking about is actually diffraction limitation, which the same photographers usually refer to diffraction.
Diffraction limitation starts to raise its ugly head at apertures smaller than about f/11 on APS-size sensors, f/16 on full frame sensors

That's a pretty good rule of thumb IMHO Diffraction gets talked about a lot (done it myself a few times ) but the reality is it's hardly noticeable in practise until you go higher than those f/numbers. Unless you go pixel-peeping...

That used to be a pretty good rule of thumb; not so much anymore.
D800 ~ f/5.6, D7200 ~ f/5.6, 7D ~ f/8, etc..

That said, I'll push well into the diffraction limits if necessary (due to lens sharpness/DOF requirements).

It's possible to see diffraction effects at those kinds of apertures, sometimes down to f/4 if you have a very (very, very) sharp lens and blow up an image to ten feet wide. Then you might just detect a slight softening of micro-detail, but that's simply invisible at normal output sizes.

If you're thinking about diffraction related to higher pixel counts, that's just rubbish. Seriously, it's nonsense, just a theory with no practical relevance.

 

[coldpenguin]

That used to be a pretty good rule of thumb; not so much anymore.
D800 ~ f/5.6, D7200 ~ f/5.6, 7D ~ f/8, etc..

That said, I'll push well into the diffraction limits if necessary (due to lens sharpness/DOF requirements).

The OP has the D7100, I am not up on Nikon cameras, but at 21MP then at f11, (whatever you wish to call it, circle of confusion, or diffraction (being a physicist by training then I was going to call it diffraction effects, but thought I would use a term that could be referenced easily on wikipedia with reference to photography) could be causing an issue.

 

[Richard King]

There are 2 ways of looking at the sharpness thing... from a physics point of view or hands on point of view. Since the real physics will take more than few years to learn. The best thing to do is sit down with a notepad and each lens you have and make real world subjective decisions about what limits you like and dont like.

Leave resolving power, T and S shelll errors, COC, Dof, index, asphericity, form, CA, diffraction, the difference between 'thin' and 'thick' lenses, lens system design, entrance and exit pupils, coating thicknesses...... to those who do or think they know about it.

 

[Phil V]

The OP has the D7100, I am not up on Nikon cameras, but at 21MP then at f11, (whatever you wish to call it, circle of confusion, or diffraction (being a physicist by training then I was going to call it diffraction effects, but thought I would use a term that could be referenced easily on wikipedia with reference to photography) could be causing an issue.

No amount of diffraction has created the blurred hand, that was caused by subject movement.

 

[Hell on earth]

There are 2 ways of looking at the sharpness thing... from a physics point of view or hands on point of view. Since the real physics will take more than few years to learn. The best thing to do is sit down with a notepad and each lens you have and make real world subjective decisions about what limits you like and dont like.

Leave resolving power, T and S shelll errors, COC, Dof, index, asphericity, form, CA, diffraction, the difference between 'thin' and 'thick' lenses, lens system design, entrance and exit pupils, coating thicknesses...... to those who do or think they know about it.

Yeah blinded by science, is what this thread has turned in to, it's a bloke thing
Where is the for ladies version ? short and sweet

H

 

[pingu666]

Not sure if anyone has mentioned, but if the camera has teleconverter digital zoom thing, you might be able to eak a little higher shutter speed by using that, as it crops the sensor.
Probably better not to give up sensor area tho

 

[Phil Young]

Yeah blinded by science, is what this thread has turned in to, it's a bloke thing
Where is the for ladies version ? short and sweet

H

Read the thread a pink monitor lol

 

[huffy]

Not sure if anyone has mentioned, but if the camera has teleconverter digital zoom thing, you might be able to eak a little higher shutter speed by using that, as it crops the sensor.
Probably better not to give up sensor area tho

The shutter speed will be limited by the flash sync speed.

 

[pingu666]

If you crop the sensor, its easier to get a higher speed, all things being equal

 

[Garry Edwards]

If you crop the sensor, its easier to get a higher speed, all things being equal

As we established very early in this thread, the shutter speed is irrelevant.

 

[sk66]

If you crop the sensor, its easier to get a higher speed, all things being equal

I don't understand that statement...

 

[sk66]

If you're thinking about diffraction related to higher pixel counts, that's just rubbish. Seriously, it's nonsense, just a theory with no practical relevance.

Not higher pixel count per-se, but rather smaller photo-sites. It is kind of the same thing... but the smaller the photo sites the smaller the airy disk size that will cause diffraction.

IMHO diffraction is like moire... it *can* occur and it *can* cause image issues, but it's not typically a big concern unless you really push things or have certain subject/image requirements. That said, I don't ever use f/22 (I've no need to).

 

[HoppyUK]

I don't understand that statement...

It's a reference to an obscure strobist technique. If the subject is restricted to the upper half of the frame, the image will avoid being clipped by the second curtain at higher sync speeds. Another trick is to do the same with the camera held upside-down, so you can flash the foreground subject at high sync speed, and the sunset behind remains unaffected - that kind of thing. But since the OP's issue has nothing to do with shutter speed, it's not relevant here.

Not higher pixel count per-se, but rather smaller photo-sites. It is kind of the same thing... but the smaller the photo sites the smaller the airy disk size that will cause diffraction.

IMHO diffraction is like moire... it *can* occur and it *can* cause image issues, but it's not typically a big concern unless you really push things or have certain subject/image requirements. That said, I don't ever use f/22 (I've no need to).

References to sharpness that only mention resolution/pixels and Airy disc size, are flawed because they don't consider image contrast that is the main contributor to perceived sharpness. So even when the sensor's resolution/pixel limit is reached, a 'very, very sharp' lens will always look better than a merely 'very sharp' lens, as it's delivering higher contrast at that limit.

 

[sk66]

It's a reference to an obscure strobist technique. If the subject is restricted to the upper half of the frame, the image will avoid being clipped by the second curtain at higher sync speeds. Another trick is to do the same with the camera held upside-down, so you can flash the foreground subject at high sync speed, and the sunset behind remains unaffected - that kind of thing. But since the OP's issue has nothing to do with shutter speed, it's not relevant here.

Ah, I know that technique...but I don't think sensor size has anything to do with it.

References to sharpness that only mention resolution/pixels and Airy disc size, are flawed because they don't consider image contrast that is the main contributor to perceived sharpness. So even when the sensor's resolution/pixel limit is reached, a 'very, very sharp' lens will always look better than a merely 'very sharp' lens, as it's delivering higher contrast at that limit.

Ok.
It could also be argued that contrast, COC, and everything else "subjective" is flawed. Simply because it is subjective it will be different for different people and in different situations.

 

[HoppyUK]

Ah, I know that technique...but I don't think sensor size has anything to do with it.

It's just a different way of looking at it. If the sensor is cropped (ie with digital zoom or by whatever method) and the bottom of the frame isn't used, you can push the x-sync up a bit. Marginal benefit though, and some obvious downsides.

Ok.
It could also be argued that contrast, COC, and everything else "subjective" is flawed. Simply because it is subjective it will be different for different people and in different situations.

Nothing subjective about it. 'Sharpness' describes the relationship between resolution and contrast (higher resolution always means reduced contrast) and that is the basis of MTF lens analysis. I do that stuff for work on an almost daily basis, using Imatest, and it's very accurate. The problem is that most folks think sharpness is all about pixels, and that's far from true.

 

[sk66]

It's just a different way of looking at it. If the sensor is cropped (ie with digital zoom or by whatever method) and the bottom of the frame isn't used, you can push the x-sync up a bit. Marginal benefit though, and some obvious downsides.

Got it, crop off the darker portion; not sensor size related.

(higher resolution always means reduced contrast)

If you had you said "with a given lens" I would agree. I'll grant that more (smaller) pixels require greater MTF from a lens to deliver the same contrast. The opposite side of this is that "contrast" is something you have quite a bit of control over after the image has been captured (contrast/clarity/sharpening). Actual detail recorded you don't any control over after the fact.


Are we far enough off topic yet?

 

[HoppyUK]

Got it, crop off the darker portion; not sensor size related.


If you had you said "with a given lens" I would agree. I'll grant that more (smaller) pixels require greater MTF from a lens to deliver the same contrast. The opposite side of this is that "contrast" is something you have quite a bit of control over after the image has been captured (contrast/clarity/sharpening). Actual detail recorded you don't any control over after the fact.


Are we far enough off topic yet?

Far enough

'A given lens' doesn't make any difference to the principles of physics - if you demand more resolution from any lens, contrast reduces. Though I take your point about post-processing and it's true that you can at least do a bit to improve a lower contrast image, though of course you can do the same with a higher contrast image and get an even better result - and the perception of improved sharpness applies to any size output/print, whereas higher resolution needs increasingly larger size outputs to see.

The other thing is that tweaks in post-processing are not the same as real lens sharpness, and you can only go so far. With artificial sharpening for example, if you push that too far it looks unnatural and actually less sharp. The Clarity slider in LightRoom is effective, though typically an increase of +20 will only lift MTF about 5%, and further increases are progressively less effective. To put that in context, the difference between full-frame and APS-C is around 12% on average, based on a 1.5x resolution increase. It's not really possible to bridge that gap. Further, attempting to improve sharpness in PP doesn't work on aberrations like coma, and generally makes a bit of a mess of spherical aberration. Basically, there's no substitute for really good glass.

 

[coldpenguin]

No amount of diffraction has created the blurred hand, that was caused by subject movement.

Ah I didn't realise we were discussing the obvious motion blur in the hand, I thought we were talking about the 'blurry'/out-of-focus look around the eyes and the rest of the 'focal' point.
However, looking at the exif data again, the focus appears to have been on the waist of the child (if I am reading the data correctly, I do canons).
Personally, I wouldn't be worried about the motion blur in the hand. I prefer items which are in motion such as moving children spinning/catching/playing to not look slightly blurred, but to look 'properly' blurred.
The head is another matter.