Flash Durations - your opinion on these please. UPDATED

HoppyUK

HoppyUK

Messages
23,200
Name
Richard
Edit My Images
No
Appreciate your opinion on these pics if you would please - your estimate of the effective motion-freezing ability of the flash images in post number 2, compared to actual shutter speeds.

The first set of pictures below are for reference. It's a standard office fan with black and grey squares stuck on the blades. The images are taken with ambient/continuous light at actual shutter speeds to freeze the movement. Use these to compare against the flash pictures in the next post down, and give your estimate of what you think the movement freezing looks like in terms of shutter speed equivalents. Thanks.

It's harder than you might think. Flash durations can be difficult to measure, particularly with studio-type flash heads (voltage regulated). Unlike speedlight guns (IGBT regulated) where the flash pulse turns on and goes out very abruptly and is therefore easy to measure, the flash pulse of a studio head turns on quite quickly, but then fades away very gradually. The total burn time at different power settings doesn't change much, but the peak brightness changes a lot - that's what adjusts the exposure and it's basically the height of the peak that determines the movement freezing potential in visual terms. The industry uses a formula to calculate a 't.5 time'* that is supposed to represent the movement freezing ability, but it's well known to over-estimate by as much as three times the shutter speed equivalent. But what do you think? Can you put a number on it, using the actual shutter speed images for comparison?

Camera is Canon 5D2. There is some distortion of the fan blades due to the scanning action of the focal plane shutter, but it doesn't affect the blurring. The fan is spinning clockwise.
*t.5 is the time the flash pulse stays above 50% of peak brightness, as measured on an oscilloscope. There is another similar measure known as the 't.1 time' but very few manufacturers publish it.

Static


1/250sec


1/500sec


1/1000sec


1/2000sec


1/4000sec
 
Last edited:
These four images are from are from three different flash heads, with the manufacturer's quoted flash duration (t.5). Using the images in the previous post for comparison, taken at actual shutter speeds, clearly these t.5 estimates over-state the real motion freezing potential, but by how much? If you can put a number on it, that would be great. Thank you :)

1) 1/4100sec (t.5)


2) 1/2550sec (t.5)


3) 1/2200sec (t.5)


4) 1/1200sec (t.5)
 
Last edited:
I'll have a go to get the ball rolling.

Pic 1, is obviously nothing like as clear as the 1/4000sec shutter speed shot. In fact, going by the darkness of the black and grey squares, it's not as clear as the 1/1000sec shot. I'd put it around 1/800sec in terms of the overall impression.

Pic 2, is not as clear as pic 1. The black/grey squares are a similar kind of tone to the 1/500sec shutter speed image, though visually the blur hasn't smeared so far, yet the blade outlines are less clearly defined. Tricky one, overall impression - maybe 1/500sec.

Pic 3, an odd one. The manufacturer's t.5 time suggests it should be more blurred than pic 2, yet it's definitely a lot clearer. In fact, it's very similar to pic 1, so overall impression 1/800sec.

Pic 4, not as clear as pic 3. The black/grey squares are lighter, and there's more blurring of the blade outlines. Overall impression, I'd say somewhere around 1/400sec.
 
Sorry, but I can't help.
I think that your testing paradigm is flawed. I do have some experience of similar testing, although obviously nowhere near as much as yourself. I tried this a few years ago and this is what I found
1. Because modern FP shutters have a vertical travel, the only meaningful test is of an object moving left/right, so any fan blades moving up/down at the time don't contribute to the result.
2. The guard is a distraction and needs to be removed - remove it though and people start wittering on about H&S
3. The background needs to be white not black, and close, so that there is a clearly seen shadow, which helps with the estimation. I think it's easier to estimate shadow movement than ghosting.
4. Although possibly fairly useful for very short flash durations, fans move much too quickly for longer flash durations, so they need to be slowed down
5. When the fan is slowed down (using a potentiometer) it loses so much torque that its speed becomes inconsistent.

And, (on a practical level) although not really part of the testing process, it's necessary to publish the photos to illustrate the results found, and every shot has to have the fan blade in exactly the same position, which means taking an unbelievable number of shots for each test.

Short of using an oscilloscope, which is in itself pretty time consuming and which needs a pretty high end one to accurately test short flash durations, I've found the gate method to be the most acccurate, it was suggested to me by a competitor of ours who I respect, and who we both know. Basically, just use a flash meter and set it to a shutter speed that is known to be long enough for the purpose - start at say 1/100th second and the resulting reading becomes the benchmark, let's say it's f/11
Now increase the shutter speed in stages, until the indicated aperture reading changes.
For example, if it is still showing f/11 at 1/1000th then the flash duration is faster than 1/1000th, and if it is showing f/13 at 1/2000th then we know that the flash duration is slower than 1/2000th and it's simple maths to calculate the shutter speed that would match the flash duration, if we had that precise shutter speed as an option on the flash meter.
Of course, nothing is perfect and this method will in reality only tell us how the flash duration compares to a shutter speed of the same value, it won't tell us either the t.1 or the t.5 time but, arguably, it's the most practically useful measurement anyway.

BTW, Bron have (or at least had) a meter that measures flash durations, I have no idea how it works or how much it costs, it's limitation (from memory) is that it only meaures up to 1/5400th (t.1), which is what they need for the fastest of their own heads.
 
Garry Edwards said:
which needs a pretty high end one to accurately test short flash durations
Click to expand...
Since oscilloscopes are used to measure speeds in the micro and nano second ranges, you'll find measuring camera shutter speeds will not be challenging even for a basic oscilloscope
 
redsnappa said:
Since oscilloscopes are used to measure speeds in the micro and nano second ranges, you'll find measuring camera shutter speeds will not be challenging even for a basic oscilloscope
Click to expand...
The task here is not to measure shutter speeds, it's to measure flash pulses, this involves using resistors suitable for the flash energy, and converting light to current via a photo electric cell - my very first efforts involved dissecting a cheap light meter, I later found out that they're available from .Maplins for about 75p :)
 
Last edited:
I'm not sure why I wrote camera shutter speeds when you clearly said flash duration anyway, since flash durations will be in the thousandth of a second range a basic oscilloscope would measure this very easily. Are you using a standalone oscilloscope or one of those that connects to a laptop?
Garry Edwards said:
my very first efforts involved dissecting a cheap light meter, I later found out that they're available from .Maplins for about 75p
Click to expand...
That is something I would do lol. good to see I'm in good company.
 
redsnappa said:
I'm not sure why I wrote camera shutter speeds when you clearly said flash duration anyway, since flash durations will be in the thousandth of a second range a basic oscilloscope would measure this very easily. Are you using a standalone oscilloscope or one of those that connects to a laptop?
That is something I would do lol. good to see I'm in good company.
Click to expand...
I use a standalone one, Hameg HM203.7
 
Thanks Garry. This is not the target I use for magazine testing, though it is basically similar. This is a simplified version, basically as 'real world' as I can make it - imagine trying to photograph a juggler tossing his spinning sticks as fast as he can go, and they would behave very like these randomly positioned fan blades. The guard is removed ;) My magazine tests also address your other comments but not wanting to clog up this thread with any more technical stuff, I'll PM you on that.

The only problem I have with this test is the subjective interpretation of the results. It's easy to get a reasonably close estimate, but I want to do better and as a starting point I would like to try and 'calibrate' my views with those of others, see what they think. The difficultly with studio-type flash is that there are two (or three) components to the flash pulse, and they all change differently. As you know, basically the action-stopping is defined by the peak of the flash pulse (the dark bit in the squares) but that is overlaid and smeared by both the flash pulse ramping up to peak brightness on one side, and then fading away slowly on the other. There is no clean cut-off at either the start or finish (unlike IGBT speedlights - they're easy to measure!).
 
Last edited:
Richard I'm probably having a stupid attack (quite common recently with me) - do you want people to give their opinion of which image in set 1 is closest to which image in set 2?
 
Redsnapper, a good oscilloscope is the answer - the problem is the t.5 formula that the industry uses as a standard measure, and is hopelessly inaccurate. It's been used for many years, presumably because it's easy to apply and flatters the manufacturer. The rule of thumb that photographers use to convert t.5 to real shutter speeds is to multiply by x3 (eg 1/1000sec t.5 = 1/333sec shutter speed) but that is very crude, doesn't work the same at all flash powers, and doesn't work at all with IGBT speedlights.

I'm seriously considering hiring a good oscilloscope, but it'd need a week or two of testing and head scratching to come up with a formula that is a) accurate and realistic, b) applies at all power levels, and c) applies to all flash types.

If you look at the fan pictures, there is in effect an oscilloscope trace in there. Take Pic 2 at 1/2550sec t.5 as an example. The black square at the top starts light grey on the left, quickly builds up to a much darker grey (the peak) and then fades away gradually (the tail) - just like an oscilloscope trace.
 
Last edited:
PaulButler said:
Richard I'm probably having a stupid attack (quite common recently with me) - do you want people to give their opinion of which image in set 1 is closest to which image in set 2?
Click to expand...

Hi Paul, yes, that's about it. Basically look at the pics in the first post and use them for reference, then compare the pics in the second post and try to put a shutter speed equivalent on them, say to the nearest one-hundredth. You can ignore the manufacturer's t.5 time - that's just for interest really.

Edit: all opinions are valid :)
 
Last edited:
I'm also a little confused, ultimately what are you trying to achieve and why ? - presumably the fan has three speeds was this kept at s contestant speed during this images. Are you more concerned about the visual image or the science behind it ?
 
OK, well here goes ...

Pic 1) - closest to the 1/1000 but I'd say slower, so 1/800 sounds about right
Pic 2) - more difficult to estimate as the blur of the squares is greater as is the edge blur. The edge blur is fairly close to the 1/500 image albeit that image has "longer" squares so 1/500
Pic 3) - this looks closest to the 1/1000 image to me, edge blur is very close and the depth of grey is also very close so I'll stick with 1/1000
Pic 4) - this looks the most blurred to me, but the 1/500 looks just a touch blurrier and the edges of both seem similar so I'd go with 1/600 I think.
 
PaulButler said:
OK, well here goes ...

Pic 1) - closest to the 1/1000 but I'd say slower, so 1/800 sounds about right
Pic 2) - more difficult to estimate as the blur of the squares is greater as is the edge blur. The edge blur is fairly close to the 1/500 image albeit that image has "longer" squares so 1/500
Pic 3) - this looks closest to the 1/1000 image to me, edge blur is very close and the depth of grey is also very close so I'll stick with 1/1000
Pic 4) - this looks the most blurred to me, but the 1/500 looks just a touch blurrier and the edges of both seem similar so I'd go with 1/600 I think.
Click to expand...

Cheers Paul. Brilliant! Just the job :)
 
I may be a bit thick but why start up the fan when you get good pictures without it working? What is the final achievement aimed at after testing with the best equipment available? seems to me a big circle getting back to where one started with a non spinning fan
 
realspeed said:
I may be a bit thick but why start up the fan when you get good pictures without it working? What is the final achievement aimed at after testing with the best equipment available? seems to me a big circle getting back to where one started with a non spinning fan
Click to expand...

My thoughts entirely. What do you benefit from by knowing this ?
 
Blank_Canvas said:
I'm also a little confused, ultimately what are you trying to achieve and why ? - presumably the fan has three speeds was this kept at s contestant speed during this images. Are you more concerned about the visual image or the science behind it ?
Click to expand...

The point is, I do product testing for various photo magazines and websites, including flash guns and studio heads. The flash duration is a vital aspect of performance, but manufacturers' t.5 times are hopelessly inaccurate and over-state the effective action-stopping potential, so I am trying to compare them accurately against real shutter speeds, actual fractions of a second. It's not easy with studio-type flash heads because of the way the flash builds up to maximum brightness, and then fades away gradually - but it's very important.

I am only concerned with the final image, and the pictorial effect. I am aware of all the science involved and it's unavoidable, but what I'm asking here is for a simple visual/subjective interpretation of how folks think the flash images compare in terms of real shutter speeds - as Paul has done.
 
Blank_Canvas said:
My thoughts entirely. What do you benefit from by knowing this ?
Click to expand...

What's the benefit? If you photograph anything that moves, it's important to know that the flash duration quoted of say 1/1000sec (pretty quick) actually turns out to freeze action no better than a shutter speed of 1/300sec (pretty slow!). Meaning that the kids you want to be nice and sharp jumping around the studio, will not be.

Questions along these line come up in this forum on a regular basis.
 
Surely if it it pictorial effect your after why not have a ladies hair blowing in front of it sideways on? As a prospective product buyer I would be more interested in seeing how powerful the fan was at different speed in picture form than just looking a fan spinning. Actually doing a video would be the route I would take anyway
 
Last edited:
HoppyUK said:
The point is, I do product testing for various photo magazines and websites, including flash guns and studio heads. The flash duration is a vital aspect of performance, but manufacturers' t.5 times are hopelessly inaccurate and over-state the effective action-stopping potential, so I am trying to compare them accurately against real shutter speeds, actual fractions of a second. It's not easy with studio-type flash heads because of the way the flash builds up to maximum brightness, and then fades away gradually - but it's very important.

I am only concerned with the final image, and the pictorial effect. I am aware of all the science involved and it's unavoidable, but what I'm asking here is for a simple visual/subjective interpretation of how folks think the flash images compare in terms of real shutter speeds - as Paul has done.
Click to expand...

Thanks for the explanation, that now makes perfect sense and puts what you're doing into context. Like you say it's not an easy one to remedy. I did wonder if you're ultimate objective was to freeze the fan with flash and determine an optimal shutter / aperture / ISO combination, I have a scientific background particularly with forensic photography and wondered where you wanted to go with this in order that I could possibly help.
 
Blank_Canvas said:
Thanks for the explanation, that now makes perfect sense and puts what you're doing into context. Like you say it's not an easy one to remedy. I did wonder if you're ultimate objective was to freeze the fan with flash and determine an optimal shutter / aperture / ISO combination, I have a scientific background particularly with forensic photography and wondered where you wanted to go with this in order that I could possibly help.
Click to expand...

Cheers Nick. You may live to regret your last sentence :D though I'm really not sure how far I want to ride this particular hobby horse. Many thanks for the offer.

In terms of the visual effect of the flash pulse, everything we need to see is actually there in these images of the fan. I really can't think of a better method or more controllable subject, and the fan relates very well to real world subjects - such as a juggler tossing his spinning sticks/batons/skittles, or active kids jumping around with flailing arms and legs, wind in their hair etc etc.

What I may end up doing is publishing comparison pictures in the magazine reviews and letting readers judge for themselves, but (apart from the fact there is rarely enough page space for that) I would like to be more helpful - people buy magazines to get these kinds of answers!
 
realspeed said:
Surely if it it pictorial effect your after why not have a ladies hair blowing in front of it sideways on? As a prospective product buyer I would be more interested in seeing how powerful the fan was at different speed in picture form than just looking a fan spinning. Actually doing a video would be the route I would take anyway
Click to expand...
Good idea for a one-off, but poor Richard has to do this on a regular basis, it would be expensive and time consuming, and he would really need to get the same model every time, and she would have to always have her hair at the same length...

Actually I've just done some shots (for a totally different purpose) where the model was tossing her head and her very long hair was flying. I used an "ordinary" (not IGBT) flash head for that and you know what? I like the blur.
 
realspeed said:
Surely if it it pictorial effect your after why not have a ladies hair blowing in front of it sideways on? As a prospective product buyer I would be more interested in seeing how powerful the fan was at different speed in picture form than just looking a fan spinning. Actually doing a video would be the route I would take anyway
Click to expand...

Thanks for the thought Baz, but I'm not sure that would work.

What do you think of the images in post #2?
 
mike weeks said:
Richard,

1 point that I would raise is that there seem to be 2 different understandings of t.5 as I have seen both 50% of peak power and 50% of total power listed on websites, now that could lead to some differences. Spent more than a few hours in my youth having to calculate 5CR

Mike
Click to expand...

Thanks Mike. I've seen others too, though it may be just the wording that's ambiguous. That method doesn't make much sense though, and it would make t.1 a shorter time that t.5 whereas in fact it's much longer.

Correct definition - t.5 is the length of time the flash pulse stays above 50% of peak brightness, and t.1 is the same at 10%. I just had a random look via google images and this graphic came up, Fig 1. There's some other relevant info on this link too https://photonemesis.wordpress.com/2009/10/13/speedlight-vs-studio-strobe/

BTW, some say t.1 is a good method for estimating effective shutter speed equivalents, but it's not actually - it's neither accurate, nor consistent across different types of flash.
 
HoppyUK said:
Thanks Mike. I've seen others too, though it may be just the wording that's ambiguous. That method doesn't make much sense though, and it would make t.1 a shorter time that t.5 whereas in fact it's much longer.

Correct definition - t.5 is the length of time the flash pulse stays above 50% of peak brightness, and t.1 is the same at 10%. I just had a random look via google images and this graphic came up, Fig 1. There's some other relevant info on this link too https://photonemesis.wordpress.com/2009/10/13/speedlight-vs-studio-strobe/

BTW, some say t.1 is a good method for estimating effective shutter speed equivalents, but it's not actually - it's neither accurate, nor consistent across different types of flash.
Click to expand...

I am with you as the majority go for the 50% of power and 10% of power. The t.1 is good to know as that is just over 3 stops from peak (I know that that is not a completely technically accurate statement) but 3 stops difference will be recorded normally. Did that article we were talking about via pm get printed yet? if so could you send a link

Mike
 
mike weeks said:
I am with you as the majority go for the 50% of power and 10% of power. The t.1 is good to know as that is just over 3 stops from peak (I know that that is not a completely technically accurate statement) but 3 stops difference will be recorded normally. Did that article we were talking about via pm get printed yet? if so could you send a link

Mike
Click to expand...

I'm not quite following you here Mike. Three stops from peak? The t.1 and t.5 times are unrelated to power or exposure, only flash duration.

The article you mention must be the reviews of location flash units I've just finished. Should be out in a couple of weeks I guess, in Digital SLR Photography magazine. It includes the Lencartas Atom and Safari-2, Bessel WP6-600, and iLux 600C :)
 
HoppyUK said:
I'm not quite following you here Mike. Three stops from peak? The t.1 and t.5 times are unrelated to power or exposure, only flash duration.
Click to expand...

t.5 is measured at 50% power and t.1 is measured at 10% power i.e. related to power unless I am missing something, whilst they are quoted in time it has a relevance to power
 
mike weeks said:
t.5 is measured at 50% power and t.1 is measured at 10% power i.e. related to power unless I am missing something, whilst they are quoted in time it has a relevance to power
Click to expand...

Yes sure, power/exposure and flash duration are related, though t.5 and t.1 are not measures of exposure. It's the three stops reference I don't quite get, but I'm sure you know what you mean!
 
Lots of interest and comments, 220 thread views to date, but actually only one poster has had a go at answering the question! Thanks again Paul.

Come on chaps, don't be shy - all opinions are valid :)
 
Brilliant. Thank you chaps :)
 
Why not ask this question on one of the Facebook photography groups, such as "The Light Side"?
They have plenty of footfall, and you should get a lot of responses
 
Garry Edwards said:
Why not ask this question on one of the Facebook photography groups, such as "The Light Side"?
They have plenty of footfall, and you should get a lot of responses
Click to expand...

There are plenty of good posters right here, and with views that I respect. A few more would be good though :)
 
Richard,

I know you understand this but for those that do not due to the way shutters operate at at above sync
View: https://www.youtube.com/watch?v=ZE8VRhHiCSs
you would really need to photograph something moving across the image in its longest dimension. If markings wer applied to the edge of the shutter blade and you were to photograph not flat on to the blade as you have been but to the side on the test would be equally valid for flash duration and shutter speed comparison

many many years ago when spec testing teleprinters we used to use a tuning forks and painted marks on a rotor to adjust the speed, essentially this is a similar operation however as Garry said you might need to take more than one image to get the markings in the right place to assess

Mike
 
Last edited:
#1: 1/800
#2: 1/320
#3: 1/500
#4: 1/400
I't's interesting that if you only look at the primary squares they all appear to be in the 1/500-1/800 range comparatively. And there's really not a lot of difference in the tails (it seems the exposure shifted slightly in the last image). I find it interesting that there seems to be no apparent relevance of T0.5 to T0.1, or rather "stopping power." It seems the characteristic of the tail curve is flat enough that it makes little difference.

I do not understand the comments about direction of travel (subject/shutter)... how does it matter?
I do understand that a subject moving w/ the shutter will leave a longer blur, but I fail to see the relevance. In the test images there are three points of motion moving in three different directions. IMO that serves to give an average impression of the SS/flash stopping capability of an average moving subject (i.e. multiple directions of motion).
 
sk66 said:
I do not understand the comments about direction of travel (subject/shutter)... how does it matter?.
Click to expand...

most shutter slits travel top to bottom so when measuring the blur it all depends upon which direction the blade is moving relative to the travel of the 2 curtains. Essentially what is being compared is chicken and eggs because the stopping power of flash can be sen in an open shutter situation in a dark room i.e. cutain travel has no effect whereas using the shutter to stop action is obviously affected by curtain travel and not the light.

Mike
 
You must log in or register to reply here.
Back
Top