Flash Durations - your opinion on these please. UPDATED

sk66 said:
#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).
Click to expand...

Thanks Steven, I agree with much of what you say. There is not tso much difference between them, certainly not as much as the manufacturers' t.5 times suggest. These are not the same images as I use for actual testing, when I use a much larger disc in place of the blades, spinning at different speeds, and a lot of different markings and detail that help me make better comparisons. These are much more simplified, more 'real-world' images. I try to avoid looking at too much detail though, or attempting to 'measure' the length of the tail/blur etc. I find it's better to stand back a bit and try to assess the overall visual impression.

I also only compare images when the target is in exactly the same position, at the top. The scanning action of the focal plane shutter distorts the shape of fast-moving objects. FP shutters always run from top to bottom of the camera, though since the image is inverted by the lens, that's from bottom to top of the picture. The fan is spinning clockwise as you look at it here, at a constant speed throughout.
 
What has been overlooked is the voltage/amps going into the fan in the first place. 230v-250v in the UK as the voltage can vary which in turn could affect the fan speed. So you would have to add a meter to the supply to be exactly the same
 
Seems to me that this kind of comparative estimation of shutter speed equivalents will only produce the same shutter speeds when the flash light-power/time graphs are of similar shape, and that will depend on a number of technology parameters liable to variation. What a practising photographer will want to know is what the motion freezing equivalence is for his particular flashes. So the simplest most reliable experiment for determining that for one's own flashes will be useful, along with the kind of data you're presenting here to make the case that there are no simple authoritative magic numbers.
 
HoppyUK said:
These are much more simplified, more 'real-world' images.
Click to expand...
I like the idea of "real world" comparisons....

We (I) tend to get too wrapped up in technicalities when what we really care about is *how* an image looks and *what* it conveys... the technical *why* doesn't really matter.

FWIW, at one point I used a PC based "sound card oscilloscope" before I purchased a real one (which has sat unused for ~ a decade now... if you lived in the US I'd send it to you). There are cheap/free programs for all of the OS's and simple to build interfaces (probes) should you decide to try that route.
 
realspeed said:
What has been overlooked is the voltage/amps going into the fan in the first place. 230v-250v in the UK as the voltage can vary which in turn could affect the fan speed. So you would have to add a meter to the supply to be exactly the same
Click to expand...

The fan speed doesn't vary by any amount that is noticeable, even in close examination. And that can get very close when the fan is at max, with a fast IGBT flash, with big enlargement of the images. For my magazine testing work, I only compare against actual shutter speed images taken immediately before, and that is double-checked against a duplicate set of comparisons taken immediately after. I've never seen any difference.
 
#1 1/800
#2 1/600
#3 1/800
#4 1/600

If you want to try a cheap oscilloscope try one of these: https://www.picotech.com/products/oscilloscope the 2000 series scopes start at ~£100, they do a pretty good job for the money. If you really want to measure the flash duration I suggest using a high speed camera. Something like a modest Phantom Miro will run at 100000 fps at reduced resolution easily enough to visualise the flash decay and it would highlight any other little quirks of the flash output.
 
Kaolin said:
#1 1/800
#2 1/600
#3 1/800
#4 1/600

If you want to try a cheap oscilloscope try one of these: https://www.picotech.com/products/oscilloscope the 2000 series scopes start at ~£100, they do a pretty good job for the money. If you really want to measure the flash duration I suggest using a high speed camera. Something like a modest Phantom Miro will run at 100000 fps at reduced resolution easily enough to visualise the flash decay and it would highlight any other little quirks of the flash output.
Click to expand...

Thanks Gareth. It's an oscilloscope I need, that graphs brightness against time, and a probe or whatever that's accurate down to around 1/100,000sec. I know nothing about them though - any suggestions welcome :)
 
Very many thanks to all who posted their estimates. Here are the max and min figures, with the overall average:

1) 1000-700, Av 1/817sec.
2) 700-320, Av 1/470sec.
3) 1000-500, Av 1/783sec.
4) 600-400, Av 1/510sec.

There was broad correlation between the estimates, and I'm pleased to say that the averages work out very close to my own figures which is reassuring. They were all way lower than the manufacturer's claimed t.5 figures, as expected, with the biggest variance being Pic-1 that nobody rated higher than 1/1000sec, compared to the claimed 1/4100sec t.5.

Thanks again for you help chaps :)
 
HoppyUK said:
Very many thanks to all who posted their estimates. Here are the max and min figures, with the overall average:

1) 1000-700, Av 1/817sec.
2) 700-320, Av 1/470sec.
3) 1000-500, Av 1/783sec.
4) 600-400, Av 1/510sec.

There was broad correlation between the estimates, and I'm pleased to say that the averages work out very close to my own figures which is reassuring. They were all way lower than the manufacturer's claimed t.5 figures, as expected, with the biggest variance being Pic-1 that nobody rated higher than 1/1000sec, compared to the claimed 1/4100sec t.5.

Thanks again for you help chaps :)
Click to expand...

One important point comes out from this and that is that the t.5 figure has no relation to action stopping ability, question is if any of these quoted t.1 and if so do any of those come close to the panels verdict? i.e. is that a better judge.

Mike
 
HoppyUK said:
Thanks Gareth. It's an oscilloscope I need, that graphs brightness against time, and a probe or whatever that's accurate down to around 1/100,000sec. I know nothing about them though - any suggestions welcome :)
Click to expand...

I think something like this would be fast enough - the 25 Mhz option would give you bandwidth to spare. I am not sure about the resistor, you would need to find one with a very fast response.
 
mike weeks said:
One important point comes out from this and that is that the t.5 figure has no relation to action stopping ability, question is if any of these quoted t.1 and if so do any of those come close to the panels verdict? i.e. is that a better judge.

Mike
Click to expand...

Good point Mike, I should have explained that. For the uninitiated, while the t.5 flash duration is the time the flash pulse stays above 50% of the peak, t.1 is the time above 10% of peak. It's a much longer time obviously, commonly regarded as a close approximation to a real shutter speed equivalent, and in practise it sometimes fits reasonably well with the other common yardstick of t.5 multiplied by x3. In other words, t.1 = t.5 x3 = real world shutter speed equivalent.

However, very few manufacturers ever quote a t.1 flash duration and I have found that, a) while it's better than nothing, it's still not very accurate, and b) it gets less accurate at lower power settings, and can actually understate at minimum output. For the flash units used in the comparisons above, the only one I have a t.1 time for is Pic-1, that has a t.1 of 1/1350sec, compared to t.5 of 1/4100sec and according to the averages here, an actual shutter speed equivalent of 1/817sec.

What I would like to do is spend some time with an oscilloscope and using the data derived from this little experiment, see if I can come up with a formula that accurately relates to real shutter speeds, at all power settings, with all flash types. The problem is the way studio flash 'stamps' a relatively sharp image with the peak, that is then blurred by the varying overlay from the tail. I'm thinking along the lines of t-point-something multiplied by the total burn time (or a % there of), and then possibly a constant factor that pulls everything in line.
 
Kaolin said:
I think something like this would be fast enough - the 25 Mhz option would give you bandwidth to spare. I am not sure about the resistor, you would need to find one with a very fast response.
Click to expand...

Great! Thanks.

£99 sounds a bit too good to be true though - what would the total cost be roughly, with probes and resistors or whatever else I'd need to make it all work?
 
I can't help with probes I have never used any with a oscilloscope, my usage is very specialised. I think a photodiode will give better results than a photo resistor as a diode is faster. A set up this this looks like it will do what you need.
 
Kaolin said:
I can't help with probes I have never used any with a oscilloscope, my usage is very specialised. I think a photodiode will give better results than a photo resistor as a diode is faster. A set up this this looks like it will do what you need.
Click to expand...

Thanks for that Gareth. That's a useful link, and other good links within it too. Maybe I can take this further without spending a fortune or huge amounts of time.

Cheers,

Richard.
 
HoppyUK said:
Great! Thanks.

£99 sounds a bit too good to be true though - what would the total cost be roughly, with probes and resistors or whatever else I'd need to make it all work?
Click to expand...

Been doing some research and found some interesting articles that show what you need, will send the links later

Mike
 
Yes, that's about right as far as flash goes - I've never had any reason to test LED flashes.
The first tests need to be carried out on IGBT flashguns, as they are the most demanding as far as the test equipment goes. But you're likely to need a different resister when testing powerful studio flash heads, something like a x100.
 
Sounds like a silly suggestion... but is there any ambient light effecting the second set of exposures?
 
HoppyUK said:
No! Dark studio.
Click to expand...


Like I said... it may have been a silly question.. It wouldn't be like you to make such a rookie mistake :)
 
i'll play:

1) 1/800
2) 1/400
3) 1/800
4) 1/400
 
The23rdman said:
i'll play:

1) 1/800
2) 1/400
3) 1/800
4) 1/400
Click to expand...

Thanks Dean. Broadly in line with most others, so not a big change to the averages, that are now:

1) 1/815
2) 1/463
3) 1/785
4) 1/499
 
Denyerec said:
I get fed up with blurry feet on dancers, so do let me know your recommended replacements for Elinchrom BXRi 500w/s heads when you're done ;)
Click to expand...

Some say that a little bit of blurring like that is actually preferable, conveying movement. As long as faces etc are sharp, I might tend to agree but if you want them dead sharp, that can be tricky.

Elinchrom BXRi-500 is far from the slowest head in that class, though the less powerful BXRi-250 with it's smaller capacitors does usefully better, roughly 50% faster at full power. Staying with Elinchrom, the ELC Pro-HD 1000 and 500 heads are capable of durations around the 1/3000sec mark (real shutter speed equivalents) at optimum settings, and close to that at quite a good range of power (depending on how they configure their multiple capacitors). That should do it, acceptable visually, unless you look very close.

The fastest heads I've tested are the Broncolor Siros (it has Bron's clever ECTC technology) and Lencarta Superfast that is IGBT. Both will do the job to a very high standard and have the added advantage of very 'clean cut' flash durations. Unlike conventional studio flash heads (voltage regulated) where the sharpest part of the image is always overlayed with a certain degree of blurring as the tail of the flash fades gradually, with these heads using IGBT control or similar, the flash pulse rises very rapidly to a peak and is then cut off sharply - there is no tail to blur anything. IGBT flash durations are easy to measure with good accuracy (at least at lower power settings like 1/4 power and less where the IGBT gets into its stride) as the t.5 and t.1 times are very similar and both are close to actual shutter speeds too.
 
It may be a bit late but, I'd put the figures as...
1 = 1/800
2 = 1/500
3 = 1/750
4 = 1/500

Pretty interesting stuff. I've actually been looking into this myself recently.
I've got a few sets of BXRi-500 and they are ok for shooting when someone reaches the apex of their leap, but whipping hands and fingers are a blur.
I was thinking of getting a couple of 250's to try to give me a bit more flexibility without worrying about the fingers too much but it may be worth looking at those lencartas.

I'm going off topic here but....
Interestingly another figure thats hard to come by is the workload. i.e. how heavy duty an item is. Only once you start speaking to a manufacturer/supplier do they start telling you what you bought isn't rated for such a workload and you need something else instead! Are these lencarta lights suitable for 3 day weeks of 1500 shots a day?
 
Deadlock said:
It may be a bit late but, I'd put the figures as...
1 = 1/800
2 = 1/500
3 = 1/750
4 = 1/500

Pretty interesting stuff. I've actually been looking into this myself recently.
I've got a few sets of BXRi-500 and they are ok for shooting when someone reaches the apex of their leap, but whipping hands and fingers are a blur.
I was thinking of getting a couple of 250's to try to give me a bit more flexibility without worrying about the fingers too much but it may be worth looking at those lencartas.

I'm going off topic here but....
Interestingly another figure thats hard to come by is the workload. i.e. how heavy duty an item is. Only once you start speaking to a manufacturer/supplier do they start telling you what you bought isn't rated for such a workload and you need something else instead! Are these lencarta lights suitable for 3 day weeks of 1500 shots a day?
Click to expand...

Thanks Deadlock. Once again, pretty much in line with others' estimates. Cheers.

I've never seen any workload performance figures, as in some measure of durability. Probably too many variables, eg 1500 shots a day could be as few as one every minute for 24hrs, or once every five seconds for two hours, and most users will never get anywhere near that. Not to mention power setting, modelling light on or off, ambient temperature etc etc. The chances of getting the industry to agree on a set of standard measures is low. Some models warn against over-use, notably those without cooling fans, to avoid the safety cut-out shutting things down. I usually test that but I've always found them to be very cautious and they go on working way longer.
 
Last edited:
Hi Richard.
Good points.

Something you mentioned was power setting.
Am I correct in assuming a generalisation whereby a lower power head (BX250Ri) operating at near full power would have a lesser rated workload than the same but higher power model (BX500Ri) operating at low power?

With regards to variables, the industry got their act together with battery life and CIPA standard testing. Agreed there are a lot of variables it probably doesn't take into account such as Live view, continuous focus, etc but it gives an industry wide standard baseline which is very useful. That way I knew for my recently purchased D810 that I'd need to have 2 batteries with me (+ spares just incase) for a single days shoot. And it turned out to be correct.

With regards to the workload of studio lights, I had 2 other sets of lights before settling on the BXRi (500's as I needed the power for certain shoots), but both would cut out mid shoot, and I'd have to replace 1-2 heads for the rest of the day to allow them to cool down and try to introduce breaks etc, etc, which would reduce productivity and increase frustration.

And this flash duration issue could have been a very valuable piece of information for guiding my purchasing choice.

So it is possible for the industry to get their act together - and damn useful when they do!
 
Garry Edwards is better placed than me to comment really, though he's not been on-line for the last few days which could mean he's away.
 
HoppyUK said:
Garry Edwards is better placed than me to comment really, though he's not been on-line for the last few days which could mean he's away.
Click to expand...
Yes, I'm away, with just a tablet...
will join in this conversation next week
 
Deadlock said:
Hi Richard.
Good points.

Something you mentioned was power setting.
Am I correct in assuming a generalisation whereby a lower power head (BX250Ri) operating at near full power would have a lesser rated workload than the same but higher power model (BX500Ri) operating at low power?

With regards to variables, the industry got their act together with battery life and CIPA standard testing. Agreed there are a lot of variables it probably doesn't take into account such as Live view, continuous focus, etc but it gives an industry wide standard baseline which is very useful. That way I knew for my recently purchased D810 that I'd need to have 2 batteries with me (+ spares just incase) for a single days shoot. And it turned out to be correct.

With regards to the workload of studio lights, I had 2 other sets of lights before settling on the BXRi (500's as I needed the power for certain shoots), but both would cut out mid shoot, and I'd have to replace 1-2 heads for the rest of the day to allow them to cool down and try to introduce breaks etc, etc, which would reduce productivity and increase frustration.

And this flash duration issue could have been a very valuable piece of information for guiding my purchasing choice.

So it is possible for the industry to get their act together - and damn useful when they do!
Click to expand...
I don't speak for the industry but a couple of points.

In theory, it's always better to have a big machine working at less than full capacity than a smaller machine working at full capacity, but I've never found that it makes any difference with studio lighting.
We used to have a customer who had one of our cheap and cheerful original SmartFlash heads, designed for home studio use, in a fixed position on a ride, lighting a shot of the victims customers 5,000 times every day, they burned out a flash tube every 3-4 weeks but the unit kept going for nearly 3 million flashes before it gave up the ghost, which I would say was impossible. One of our current flash heads is rated at 600Ws but the flash tube itself is rated at 2,400Ws, and I only know of one of those that has had to be fitted with a new flash tube, that particular customer has a busy studio and hammers it hard.

Back to topic, I don't think that the industry is anywhere near big enough, well organised enough or honest enough to agree on standards, let alone to stick to them. It's very common for some manufacturers not to actually understand the products that they make, and to describe their performance "optimistically". In my experience, Bron are streets ahead of everyone else in this respect, there are other players who can also be trusted, but many/most can't. Often, even the safety standard certification is false, although that applies mainly to lights sold on Ebay/Amazon
 
Thanks Garry.
I've a growing number of questions, but it would perhaps be best to gather them this evening/tomorrow in a new post.
 
Niwdog said:
Shouldn't you be measuring left to right rather than a rotation?
Click to expand...

There's nothing to 'measure' here. What I'm asking is for a subjective impression of how flash durations compare visually to shutter speeds.

Most people's estimates appear to be quite similar to my own, but are also a long way from the industry standard t.5 duration figures quoted by manufacturers. That's the problem I'd like to resolve and the answer is hidden somewhere in an oscilloscope trace, so I'd like to come up with a formula that works for all kinds of flash and accurately reflects real world use.
 
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...

I don't think he's testing the fan's performance :)
 
You must log in or register to reply here.
Back
Top