[bfmelton]

Hi all,

I need some help. I'm using the Pentax 645Z in studio for dance photography, which means that my dancers are engaging in fast movement. The only ambient light is coming from the modeling lamps. The strobes are Alien Bees, both 400s and 800s.

The unretouched pics below show a marked amount of what I assume is flash ghosting on the dancers' feet (which are obviously the fastest moving thing in the frame). Interestingly, other fast-moving objects, such as hands, hems of skirts, hair, and even dust are nicely frozen, but none of that matters due to the unacceptable ghosting with their feet.







The 645Z is notorious for its slow sync speed of 1/125 sec. My reading of the AB800 specs suggests that worst case I'm getting a t.1 flash duration of 1/550 seconds, and given my power settings I'm probably closer to 1/1000.

In an attempt to get the ghosting under control, with the dancer in red, I purchased and used the 75mm leaf shutter lens to get the shutter speed to 1/500, thus bypassing the 1/125 X sync of the Pentax, but her foot is still ghosting.

My questions:

1) Am I diagnosing this correctly? Is this ghosting or simply too slow a shutter speed?

2) if it is ghosting, is it being caused by the modeling lamps, or by the flash duration of the strobes?

If the problem is the strobes, will switching to Einsteins eliminate the problem? Based on my reading of the specs (page 13) and my anticipated lighting requirements, they would let me get me to a t.1 of around 1/4000 seconds as opposed to my current 1/1000 or so. This is a pricey solution, but I'm willing to do it if it would solve the problem (and would be really bummed if I dumped that money into Einsteins and the problem persisted).

Any input would be welcomed. Thanks!

[TonyW]

It is possible that even 1/1000 that you will see some motion blur in the fastest moving areas of the image, however I would want to rule out the effect of the modelling light.

So is it possible for you to test either without the modelling light or masking the light in some way?

Also prior to throwing a lot of money at Einsteins or other it may be worthwhile looking to hire for a day or two?

[Bob L]

It's hard to perceive from the shots how "vigorous" the dance routine is. I shoot a fair bit of football (soccer) and I see similar blur at the extremeties (hands and feet) for night matches when I have to use 1/1000th second to avoid setting the iso to high. I did some contrived action shots with flash and found that 1/2000th was adequate to freeze the player and the ball. I used Priolite HotSync those shots which means that the emphasis was on shutter speed and not flash duration (which is quite long).

Bob

[stevebrot]

Welcome to the Pentax Forums!

My questions:

1) Am I diagnosing this correctly? Is this ghosting or simply too slow a shutter speed?

2) if it is ghosting, is it being caused by the modeling lamps, or by the flash duration of the strobes?

Looks like it is mostly blur to me (confirmed, I suppose by what you saw with the LS lens and the trailing edge of the heel on your top example). If you suspect that your flash duration is simply too long, the question might be, "What light duration would one expect is needed to freeze the dancer's foot?"

Another question might be, "What LV is being contributed by the modeling lamps and will that result in sufficient sensor response to make a ghost against the black background?"

Both questions are open to actual testing with your gear with the second case being testable with a static subject*.


Steve

* No strobe, manual exposure at the same settings you are using for the live shots. The image you get is the brightness of the ghost.

---

Last edited by stevebrot; 07-30-2017 at 08:29 AM.

[AstroDave]

I'm pretty sure you haven't adequately stopped the motion.

Suppose the feet are moving 5 or 10 feet per second (I just tried waving my hand, and I can readily move it 3 feet in well under a second and a foot that is kicking probably moves comparably fast).

For ease of calculation: use 10 feet per second which is essentially equal to 3 meters. In 0.001 second, the total motion of the foot will be 3 millimeters - about what is in your pictures.

Plus, as I have measured for various flashes (see Flash Duration Measurements - AF-540FGZ - PentaxForums.com for one example), there can be lengthy tails on the ends of the flash light output, so the overall duration exceeds the t0.1 times.

You need faster flashes - try lowering the power first. If that leaves enough light for your shots, you will gain some effective speed.

Conversely, I think that little bit of blur gives some realism to your shots (unless your goal is a sharp image of just the dancer's foot!!).

[stevebrot]

Conversely, I think that little bit of blur gives some realism to your shots (unless your goal is a sharp image of just the dancer's foot!!).

...I was wondering about the crop. At normal viewing distances for the entire frame, the slight blur would be almost unnoticeable. As for the "notoriously" deficient sync speed...I was reminded of that video of Bruce Weber rapid-shooting Michael Jackson (in motion) with Pentax 6x7...1/30s x-sync

You Tube

You Tube

Steve

[bfmelton]

Thanks to all for the very thoughtful and useful replies. Based on the input here, while I'll test out the ambient light from the modeling lights, I fear the flash duration is just too long. These are cropped way in, but unfortunately the problem is apparent in more normal views. When the rest of the dancer is tack-sharp and her feet are just a tad blurry, it's not only obvious but also clear that it's a bug and not a feature. :-)

The suggestion to rent is very appreciated. Hadn't even thought about rental: I have just rented some Einsteins for the next shoot and will give them a test drive.

Thanks again!

[Edgar_in_Indy]

I think you will find the following article interesting:

Rob Galbraith DPI: Stopping action with the Paul C. Buff Einstein 640 monolight

It's a review of the Einstein strobe, specifically in regards to its ability to freeze action far better than any strobe that doesn't cost near $10,000 or more. Here's an image of a sprinter from the article. They froze the action perfectly with Einstein strobes, but they said that in order to get the same amount of action freezing ability using daylight they had to shoot at 1/5000 of a second:

[Gerd]

Hello bfmelton,

the blurred light edges on the feet are correct. The light beam is created by your backlight (which is already slightly outshone) and the blurred by the motion blur.

With techniques such as HSS or SuperSync, PowerSync etc. you can not freeze fast movements by "LIGHT". You can shoot with such techniques only over the shutter speed a picture absolutely sharp. Both techniques produce only one kind of continuous light.

Another possibility is to expose the subject only via the flash time of the flash. Then, however, the flash must be correspondingly fast and the ambient light must not play any role. If you only over the burning time exposes, you can the 645Z synonymous to 1 / 125s pose and the picture is still sharp.

Greeting Gerd

[photoptimist]

It's also worth noting that the excellent examples of people frozen in mid-motion all caught the subject at or near the apex of motion when the fastest-moving parts (arms & legs) have reached their limit of travel and are about to reverse direction. At that moment, the body is relatively still with respect the camera (with the camera perhaps smoothly panning to track overall body movement as in the case of the sprint hurdler). Capturing the apex of motion requires lots of practice so that one learns to anticipate the apex and hit the shutter button a fraction of a second before the apex or it requires spray-and-pray shooting (with all that implies about the performance of the camera and lighting to handle long fast bursts).

The other strategy might be to embrace the blur and use it to increase the dynamism of the still image by using: a much slower shutter speed; brighter continuous lighting of the fastest-moving parts; and selective rear-curtain flash on other parts (e.g., face). If one side of the dancer's body is almost entirely lit by continuous lighting and the other side is lit almost entirely by flash, the effect could be quite stunning. The challenge is that it requires a lot more control over both the lighting and the dancer's motions.

[Edgar_in_Indy]

If the problem is the strobes, will switching to Einsteins eliminate the problem?

OP, if you happen to be located somewhere near central Indiana, I have a couple Einsteins I could let you borrow to experiment with. I believe that they would allow you to get the results you are after.

It's also worth noting that the excellent examples of people frozen in mid-motion all caught the subject at or near the apex of motion when the fastest-moving parts (arms & legs) have reached their limit of travel and are about to reverse direction. At that moment, the body is relatively still with respect the camera (with the camera perhaps smoothly panning to track overall body movement as in the case of the sprint hurdler).

That's good advise, from both a technical and aesthetic perspective. But in the case of the Einstein photo of the sprinter, I believe that moment was chosen for purely aesthetic reasons. The article I linked to above uses a spinning wheel to demonstrate that at 40ws, the Einsteins freeze action better than a shutter speed of 1/8000s. The picture of the sprinter was taken at 20ws, where the freezing power would be even greater.

Here's how the shot was setup. For reference, the photographer was using a Canon 1D Mark IV with the 70-200mm F/4L. I'm not sure what the ISO was set at, but I would like to know as the lights are not very close to the sprinter and the aperture is obviously no larger than f/4.

[photoptimist]

OP, if you happen to be located somewhere near central Indiana, I have a couple Einsteins I could let you borrow to experiment with. I believe that they would allow you to get the results you are after.



That's good advise, from both a technical and aesthetic perspective. But in the case of the Einstein photo of the sprinter, I believe that moment was chosen for purely aesthetic reasons. The article I linked to above uses a spinning wheel to demonstrate that at 40ws, the Einsteins freeze action better than a shutter speed of 1/8000s. The picture of the sprinter was taken at 20ws, where the freezing power would be even greater.

Here's how the shot was setup. For reference, the photographer was using a Canon 1D Mark IV with the 70-200mm F/4L. I'm not sure what the ISO was set at, but I would like to know as the lights are not very close to the sprinter and the aperture is obviously no larger than f/4.

Yes, it may have been for aesthetic reasons, but if you run the math of how fast the hurdlers go, how fast their feet must move, the frame size of that photo, and that camera's resolution, then you can show that there would still be maybe 5 pixels of motion blur even with proper panning and a 1/10,000 second flash or shutter if the shot were taken when the feet were moving their fastest (for no panning, the body would have about 4-5 pixels blur and the feet maybe 10).

----

What's interesting to me is that "acceptability" of blur seems to follow a bathtub curve. Really small amounts of blur (say a few pixels) are either not noticeable or not bothersome (OK, pixel peepers might have tighter standards). None of these high-speed techniques totally freeze motion -- they just slow it down to the point of negligibility. On the other end of the blur spectrum, large blur (maybe 20 to 200 pixels or more) can also be acceptable and even pleasing in conveying motion (e.g., background blur behind panned objects, waterfalls, ocean waves, streaming fireworks, etc.). It's the middle ground where blur ruins sharpness but does not do enough to convey motion that just looks bad.

[stevebrot]

None of these high-speed techniques totally freeze motion -- they just slow it down to the point of negligibility.

So true! One of the maxims of measurement is that the closer you look, the less happy you are with your methods.*


Steve

* ...cannot resist adding my favorite cartology anecdote, that coastlines grow longer as the measuring stick grows shorter.

[Edgar_in_Indy]

Yes, it may have been for aesthetic reasons, but if you run the math of how fast the hurdlers go, how fast their feet must move, the frame size of that photo, and that camera's resolution, then you can show that there would still be maybe 5 pixels of motion blur even with proper panning and a 1/10,000 second flash or shutter if the shot were taken when the feet were moving their fastest (for no panning, the body would have about 4-5 pixels blur and the feet maybe 10).

I'm sure you're right, but I suppose that for subjects moving at non-superhuman speeds and for images viewed at any reasonable screen or print size, it would be close enough to frozen as to not be noticeable.

For the benefit of the OP wondering if Einsteins will perform better than his Alienbees, I just did a quick and dirty comparison between an Einstein and an Alienbees B1600, both of which are rated for a maximum output of 640ws. I used a spinning fan for my test subject.

Click each image for the full-size version.

First, here is the fan turned off, for a baseline.



And here's the fan running, lit by the B1600:



And here it is with the Einstein, at 40ws. While it is not perfectly frozen, it is at least close, and is much, much better than the B1600. The sticker spinning quickly in the center of the fan is essentially frozen and is sharp, while the stickers at the ends of the blades are soft but can still be easily made out. Overall, it's a dramatic difference.



And finally, here's the fan shot under direct, noon-day sun. I used my K-30's maximum shutter speed of 1/6000 sec, but in order to match the same f/6.3 aperture of the studio shots, I had to bump the ISO to 800 from the base of 100 used in the studio. As you can see, the Einstein froze the spinning blades better than the 1/6000 shutter speed could.

[BostonUKshooter]

This link may also help you?
Xplor/Godox600, eVOLV200 and Pentax 645Z, HSS!!! – Mark Kitaoka Photographs
Seems like a relatively "cheap" way to achieve HSS with the 645Z!