Macro Flash Comparison

Just a quick comparison between 3 different flash setups.

1. A Meike MK-MT24 II Twinflash set to ETTL with equal flash outputs from each flash tubes.

2 A Metz 58 AF-1 set to ETTL fitted with a home made softbox approx 3 x 6 inches mounted on a bracket just above the front of the lens.

3 A K & F Concept twin ringflash set to ETTL with equal flash outputs from each flash tube.

Photos below are in that order, originally shot in raw and the same conversion process used for all shots.

Not sure if it’s better or worse but I prefer the middle one. It has more contrast which makes me think it’s more detailed. The middle image appears to have a more neutral white balance. I’m an uneducated hobbyist so not a real technical critique.

I agree, the middle one is my normal setup, I am sure the others could be improved by a bit of tinkering but the Meike definitely needs some diffusion to soften the shadows and a bit of white balance adjustment.

If your normal setup gives you closest to your preferred results then that’s what I would go with.

Oddly, although the first shot has the twin flash units supposedly balanced, the photo clearly shows the Right flash casts a more contrasty shadow than the Left flash. Perhaps the fill affect of one flash only provided light on the right side, reducing shadow contrast only to one side of the subject.

I am in agreement that photo #2 has better balance of contrast. Seeing twin lights flanking the lens makes me not like that particular configuration. Using portrait lighting as the analog, I would set one light right above the lens and -1.5EV in intensity compared to the flanking light...the flanking light serves as Main, the centered light serves as Fill in this configuration.

Yes I agree there is definitely some imbalance between the twin flashes of the Meike. To be fair I only had a limited time in which to test the unit as it was only borrowed from a friend.

I am sure that the Meike is capable of much better results but this would require experimentation with light positioning and probably fitting some sort of diffusers. I was considering buying a twin flash unit but these results have convinced me to stick with my original softbox technique.

Wilt wrote in post #19127646
...Seeing twin lights flanking the lens makes me not like that particular configuration. Using portrait lighting as the analog, I would set one light right above the lens and -1.5EV in intensity compared to the flanking light...the flanking light serves as Main, the centered light serves as Fill in this configuration.

Why would you make the fill (flanking) a full stop and a half brighter than the main/key (centered) flash head? In portrait photography the centered (key, normally at a high angle) light is the main (key) and the flanking (and usually lower) light is the fill (part of three point lighting). Depending on the positioning of the lights you can set them both to fire at equal levels. In my experience with ratio control it only works as expected in manual mode, so it's more suited for studio shooting when you have lots of time to work (the subject is dead).

In E-TTL mode the results aren't intuitive because the E-TTL pre-flash isn't effected by the ratio control settings. Overall exposure is calculated off of an equal E-TTL pre-flash fired by both heads, and then ratio control changes how long the two heads fire. If there was a separate E-TTL pre-flash for each flash head then it would be more reliable, but as is the head that fires longer will just blow out the specular highlights unless there's a lot of negative FEC dialed in. This winter I'm gonna experiment with E-TTL, ratio control, and FEC to see if I can get it to at least be consistent enough to use. Would like to fire my fill at -.5 EV compared to the key while in E-TTL mode cause I don't have time to change flash settings when shooting active critters and changing magnification frequently.

I agree with your view on putting both heads at opposite ends of the lens -just results in flat light.

Dalantech wrote in post #19128338
In E-TTL mode the results aren't intuitive because the E-TTL pre-flash isn't effected by the ratio control settings. Overall exposure is calculated off of an equal E-TTL pre-flash fired by both heads, and then ratio control changes how long the two heads fire.

That is very useful information, I hadn't realised that it worked in that way, if the pre-flash isn't proportional to the settings then it makes it a bit pointless to use E-TTL in this sort of circumstance.

Dalantech wrote in post #19128338
Why would you make the fill (flanking) a full stop and a half brighter than the main/key (centered) flash head? In portrait photography the centered (key, normally at a high angle) light is the main (key) and the flanking (and usually lower) light is the fill (part of three point lighting). Depending on the positioning of the lights you can set them both to fire at equal levels. In my experience with ratio control it only works as expected in manual mode, so it's more suited for studio shooting when you have lots of time to work (the subject is dead).

Classic portrait lighting has a 5:1 ratio (Main source 4x, Fill source 1x, Fill fully illuminates the entire area seen by the lens while Main provides highlights to only a portion of what the lens sees) as 'Moderate Contrast'. That would provide more contrasty lighting of the subject...and the opinions of the sample photos were almost all in favor of the somewhat contrastier photo.
If lighting is 3:1 in relative strength of light on subject (Main 2X, Fill 1X, (Fill illuminates the full subject while the Main illuminates only a portion or the subject), that is 'Low Contrast'.

BTW, I did not suggest the Main/Key be centered...I would put the Fill centered over the lens so it casts no visible shadow of its own; the Main is the flanking light to provide the highlights and the visible shadows.

Ramon-uk wrote in post #19128379
That is very useful information, I hadn't realised that it worked in that way, if the pre-flash isn't proportional to the settings then it makes it a bit pointless to use E-TTL in this sort of circumstance.

The way eTTL should work (not to say this is actually HOW it behaves!) is that both lights cast a uniform amount of preflash, and the camera computes what franctional output to ask the flash to give in order to illuminate the subject to obtain 'perfect exposure'. The command it given to the flash units a moment before the shutter opens, and the flash units apply the ratio that was commanded via FEC settings and or the ratio requested. So if the Fill was set for -1EV output compared to the Main (which provides 0EV), when the camera commands the Main to give 'half power' it gives half power and the Fill gives 1/4 power (-1EV relative to Main)
While my suggestion of -1.5EV for the Fill would not result in 5:1 subject lighting, it provides a bit more contrast than if using 3:1 source ratio.

I don't have experience with a ratio'd output from eTTL flash units, as I never trusted eTTL to provide consistency...every once in a while it would shoot Full Power although the eTTL command from the camera had asked for a lower power output (...the electrical contacts in the hotshoe to hotfoot were too subject to poor contact interrupting the communication between body and flash.) So I defer to Dalantech's critique in finding eTTL was not intuitive in its result.
My own approach to ratio'd flash uses two photosensor Auto flash units, and one (on camera) is set to higher ISO than actually shot to ratio it down in power output, while the other (the Main, located off camera) fires assuming same ISO as set on camera.

I will readily admit that lighting a portrait will be very different inherently than a macro subject! But the principles of ratio'd lighting on the subject applies, even if the application techniques present different issues (like the 'background' behing the subject bug is very close, and therefore much more reflective of light back to the subject, than if a background were located 5-6' behind a human subject.)

Wilt wrote in post #19128559
Classic portrait lighting has a 5:1 ratio (Main source 4x, Fill source 1x, Fill fully illuminates the entire area seen by the lens while Main provides highlights to only a portion of what the lens sees) as 'Moderate Contrast'. That would provide more contrasty lighting of the subject...and the opinions of the sample photos were almost all in favor of the somewhat contrastier photo.
If lighting is 3:1 in relative strength of light on subject (Main 2X, Fill 1X, (Fill illuminates the full subject while the Main illuminates only a portion or the subject), that is 'Low Contrast'.

It sounds like you're swapping what I'd call (and what every article I've read calls) the key and the fill. The key is usually the main light source, sometimes centered above the subject (doesn't have to be) and at a high angle. The fill is off to the side and is used to "fill in" the shadows to keep the key from crushing them (that's why it's called the fill).

In a studio setting it's pretty common for the key (what you're calling the main, and it should be the primary light source) to be further away from the subject than the fill, and that's why the fill is usually a lower output light source than the key. With a macro twin flash the key and the fill are pretty much at the same distance, and with the limited angles between them and the subject (just not enough space for creative positioning) they can both be the same intensity. I'd like to reduce my fill by half a stop to keep the specular highlight created by it to a minimum. You can see the specular created by the fill in the female ladybug's shell, just to the left of her head:

IMAGE LINK: https://flic.kr/p/2jG4​3DE

Wilt wrote in post #19128559
BTW, I did not suggest the Main/Key be centered...I would put the Fill centered over the lens so it casts no visible shadow of its own; the Main is the flanking light to provide the highlights and the visible shadows.

I always use the key as the higher light source relative to the subject with the fill off to one side. I'm also back to using an adjustable flash shoe to change the angle of the key so that it's almost firing straight down. Here's an example of how the specular highlights change, although not the best comparison to the first shot because the angle and mag are not the same:

IMAGE LINK: https://flic.kr/p/2jF2​1ty

Ramon-uk wrote in post #19128379
That is very useful information, I hadn't realised that it worked in that way, if the pre-flash isn't proportional to the settings then it makes it a bit pointless to use E-TTL in this sort of circumstance.

For a multi flash system in E-TTL mode each flash head would have to fire it's own metering pre-flash, adjusted by the ratio settings, and I think that it might be too slow to be practical.

E-TTL can also be confused by the reflectivity of the scene. If there isn't anything behind the subject to reflect the pre-flash back into the lens the metering system won't see very much of that light so it will turn the main flash on longer than necessary and cause the subject to be over exposed. The opposite can happen with a very reflective scene -a lot of the pre-flash makes it back into the lens and the main flash won't fire long enough and the resulting image will be under exposed. Also happens when shooting in natural light.

Wilt wrote in post #19128562
The way eTTL should work (not to say this is actually HOW it behaves!) is that both lights cast a uniform amount of preflash, and the camera computes what franctional output to ask the flash to give in order to illuminate the subject to obtain 'perfect exposure'. The command it given to the flash units a moment before the shutter opens, and the flash units apply the ratio that was commanded via FEC settings and or the ratio requested. So if the Fill was set for -1EV output compared to the Main (which provides 0EV), when the camera commands the Main to give 'half power' it gives half power and the Fill gives 1/4 power (-1EV relative to Main)

Actually that's pretty accurate, but ratio control seems to shift both flash heads based on the ratio that's dialed in. So if the ratio is 2 to 1 and the normal E-TTL exposure would be to fire both heads at 1/4 power you might get the key at 1/2 power and fill at 1/4. The ratio is still 2 to 1 but now the key is over exposing by a full stop. I'm exaggerating the flash head power levels to keep the math simple, but you get the idea. It doesn't work like flash exposure compensation, which E-TTL would calculate the exposure and then adjust both flash heads based on the FEC. With ratio control the flash is still trying to fire long enough to give you a proper exposure based on the pre-flash meter reading, but now one flash head is gonna turn on longer than the other and quite possibly over expose part of the scene. The last time I played with E-TTL ratio control I was dialing in around -2 FEC when I normally don't go above -2/3 FEC without ratio control. Gonna have to experiment with it again cause the light meter in the 80D I'm shooting with is pretty good, so maybe I can get more consistent results now.

Wilt wrote in post #19128562
I will readily admit that lighting a portrait will be very different inherently than a macro subject! But the principles of ratio'd lighting on the subject applies, even if the application techniques present different issues (like the 'background' behing the subject bug is very close, and therefore much more reflective of light back to the subject, than if a background were located 5-6' behind a human subject.)

It's different, but it does scale pretty well. The biggest issue is having a limited amount of room to work, so not many options for how the lights are positioned. I keep my backgrounds to no more than the same distance behind the subject as the flash is in front of the critter (varies with the mag, with lower mag requiring more separation to avoid recording too much detail in the background). So I think that the light reflected back into the lens from the background is reduced by two stops maximum at most (inverse square law). Not an issue since those backgrounds are pretty reflective anyway.

First we should keep in mind that artificial lights simply mimic what occurs in nature...

1. The concept of Fill is not its location but its PURPOSE...to provide an overall level of illumination analogous to ambient light (light which bounces up from the ground, what bounces off nearby cliffs/walls, etc. It reduces the contrast of shadowing cast by the Main, and the Fill casts no shadow of its own It is impossible to NOT cast a shadow of its own when it is off to the side of the lens!
   You can fill shadows to make them less contrasty, but cannot eliminate them once they exist...this is proven even by photo 1 in original post of this thread! If one thinks of Fill as 'weak' and 'soft' it will not cast shadows of its own, even if off to the side.
   
2. The Main is the higher level of illumination which is more directional, as if from the sun behind a cloud...its specularity is high if sun with no cloud, medium specularity if sun behind a somewhat small cloud, but nevertheless if provides highlights.
   

Dalantech wrote in post #19128805
It's different, but it does scale pretty well. The biggest issue is having a limited amount of room to work, so not many options for how the lights are positioned. I keep my backgrounds to no more than the same distance behind the subject as the flash is in front of the critter (varies with the mag, with lower mag requiring more separation to avoid recording too much detail in the background). So I think that the light reflected back into the lens from the background is reduced by two stops maximum at most (inverse square law). Not an issue since those backgrounds are pretty reflective anyway.

Good fundamental technique! If the flash was a point source, if the subject were illuminated for f/4, the background a similar distance away makes its illumination inherently -2EV in level (double the distance, 1/4 the light intensity) But since your technique uses softboxes close enough to be considered 'large' relative to the subject, the falloff of intensity is -1EV (inverse linear, not inverse square), but even -1EV for backgroud draws the viewer eye to the subject more than the background.

Interesting for me to learn that eTTL controlled ratio'd lights essentially keep the fill level fixed, and the Main level rises, so that overexposure can result! That tells me the eTTL photographer really needs to outthink eTTL and deliberately dial in a greater amount of FEC for greater ratios of output light (e.g. 7:1 vs. 3:1, or 5:1 vs. 2:1) so that shadows get darker rather than highlights get brighter.
In my own use of photosensor TTL flash for ratios, I do that simply by telling the on-camera Fill that 2*ISO or 4*ISO is being used so that the fill deliberately underexposes the appropriate amount, and the off-camera Main knows the real ISO I have selected.