Good practical analysis of flash vs LED. One more variable to consider is the actual amount of diffusion and the effect it has on contrast, edge sharpness and perceived film grain. Highly collimated light from a point source or through condensers results in higher levels of these characteristics compared to diffuse light sources. This is obvious in darkroom printing where a higher contrast grade printing paper or filter is usually needed with a diffused light source enlarger vs. a condenser enlarger for the same film negative. This is a generalization and a better understanding can be gained by searching the Callier effect.
I've always used flash - an AD200 in a softbox. The only disadvantage I found was that it is harder to focus, but I could use the modeling light if I were to use the flashlight head (I use the bulb) I use an on-camera film holder so I just aim at my softbox (which is already set up) and fire away. I feel like I can get through a roll pretty quickly. I use shutter speed and HSS to control exposure, because adjusting exposure in-camera is faster than on the remote. I calibrate my exposure by shooting without any film, exposing so that the white touches the edge of the histogram.
Color temperature and CRI only tell part of the story when it comes to light quality. The improvement in quality when using flash, especially with slide film, is due to the spectral power distribution of flash vs. LED. I've experimented with measuring the spectral power distribution of LED, fluorescent, flash, and actual sunlight with a spectrophotometer (i1 Studio) to generate graphs of the spectral power distribution of these different light sources, and the graph for speedlite output most closely mimics the graph for sunlight. The output from LED has peaks at certain wavelengths, and the output from fluorescent is spiky as heck. One can get away with a somewhat incomplete spectrum when doing negative conversions (depending on which wavelengths are deficient), but with slide film, reds and skin tones will suffer the most from poor-quality scanning light.
Interesting experiment. One convenient way of reaching a good diffusion would be to bounce the light from the flash off a white sheet of paper. For example, the flash pointing into a small box with a white sheet of paper facing the film holder and camera, or a sheet of paper at 45° and the flash perpendicular to the camera and film holder, as I used to do for duplicating slides with less contrast before digital photography.
I remember back in 2014-2017 when I didn't have access to a lab providing decent scans (I was in Greece at the time), I was improvising with the equipment I had available at the studio I was assisting. I was using big studio flash lights with softboxes on the floor, and on top of that I was rigging a sheet of white plexiglass that acted as a large diffusion and then trying to photograph the negative on top of all that 😂. I'm sweating just thinking about the amount of work I was doing just scan a few frames which I would then have to manually convert in photoshop. I'm so glad analog has made a major comeback and has allowed us to have all these different ideas and products. Everybody is pitching in on it and we are collectively progressing. Quality of light is definitely very important and if the technology is not there on video lights then we could definitely find a way to rig flash lights to do the job.
@@penquito Are there any new labs in Greece or labs that have made a comeback or not really? I’m sure you’ll make a killer setup at home. There’s plenty of resources online to take inspiration from.
I purchased an x-ray light box online from a dental office. I took out the internals on one end I cut an opening for my canon speed light. On the other end, I cut another round opening where I installed a nightlight type bulb socket that could be turned on and off with a flip switch when focusing. I lined the light box with silver foil tape. I, too, am using F 11 with 1/64 power. My color corrections have been excellent!
Hashem from my experience with flash, the key to making this work would be to extend the distance of flash to film plane with adequate diffusion for an even spread across the film plane, while also doing as custom white balance to ensure that the diffusion is not causing a colour shift. I'd also line the box with black matte material that stops the bouncing of light within the box itself. There are clearly other variables, but in terms of flash alone, that would be it, along with the quality of the flash output, which is dependent on the model.
And why not, on the contrary, to cover the inner space of box with reflective material? This would improve the situation with the uniformity of dispersion, it seems to me...
Yeah I figured this would be the case regarding distance, as mentioned in the video. Lining the box makes sense too! But, I had to make due with what I had in terms of materials and time (for the purposes of my test). It would be cool to see people's purpose made setups using some of these techniques 😃
Love this experiment, and thank you for going through the pains of doing it and being generous with the results. I wonder also if bouncing the flash like in a book lighting setup might work (but that also means more faffing around the entire rig), but regardless this is so fascinating. Thank you again!
This was interesting. I think the convenience, continuous light and low heat output of LEDs make them really good sources. That said, all light sources have differences in distribution of frequencies and LEDs are certainly a quite different light source than a flash. Don't know the details and don't have the equipment to measure different light sources, but (as you mentioned) the shutter speeds you can get with a flash would be an advantage. Maybe using a flash like the Godox V1Pro with a build in led light would simplify the setup. Keeping the flash outside the box might also work well as you could have just a reflector inside the box. Increasing the distance could make it easier to get light distribution perfect and keep the flash cooler under heavy use, but with 1/32 or less power that might not be a problem. Looking at some of the macro flashes the part of the flash that actually generate the light is quite small and could probably be built as a fairly compact light source if someone was interested in making one specially for camera scanning. Flash is a pointier light source than a bunch of LEDs though so getting an even distribution without adding height to the source might be difficult.
I still maintain that the Cinestill CS Lite is the best of any light source I've used that's currently out. That little color setting is a shocking difference maker, and it's cheap. If it could fit a 4x5 sheet of film on it, it might be perfect.
I ve tried a custom tunable RGB set - there was no such big difference in the result on a first look. But if you look at the channels while conversion you will see one significant thing. Because the base layer of the film has its own saturated tint (yellow for example) you get less information in blue channel unless the color temperature of your light is cool enough. Even if it is around 6000-6500k it still gives a lack of info in blue channel. That causes excessive noise in blue channel while converting. RGB or a very cool light (even through a cyan gel filter) gives you slightly better results due to a better base layer exclusion (the more grayish you make it in pre - the clearer will be the result in post). So you can use a cyan gel filter on a flash or make your rgb light as cool as it needs to be for your negative.
I've been using a flash for a while. I started because my scanning light (which was expensive) was not giving me very fast shutter speeds at 100 iso and F 5.6. Unsurprisingly, flash kicks constant lights's butt in terms of brightness. I use a mixing chamber from an old enlarger to diffuse the light. I will keep using the flash for black and white and slide film, but I'm working on building a custom RGB light for color negative film scanning
Great video Hashem! Totally speculation on my behalf, since I don’t use flash, but I think the more neural colours that you’re getting from the flash are due to the flash being a better quality of light source. Much like that which you’d get from the sun or a tungsten bulb. Unless you’re using a really high quality LED fixtures, you’ll find might find a minor green or magenta shift in LEDs. Even if the LED has a high CRI and TLCI ratings that doesn’t really show the whole picture regarding the quality of the light. From my understanding SSI ratings are a much more accurate way to rate LED fixtures. If you want to go on a deep dive the channel @gaffergear has some great videos on LED lighting. As your research suggests using RGB lighting will probably allow you to get better quality scans as you can dial the rgb values in. They typically have a +\- green and magenta function. If you were to go do the route of RGB experimentation I’d suggest giving a light fixture with RGBWW LED diodes a crack. These have an additional warm white LED. Which can help improve the light quality.
LEDs generally do not have a continuous spectrum output of even intensity. This applies to white and color LEDs. Flash will have a more balance spectrum. However, the function differences are often quite small, and they can be adjusted in a digital processing world. On the other hand, the same cannot be said for using LEDs as a light source for color printing on traditional photo paper, which is more sensitive to gaps in the LED spectrum. This was illustrated recently in a carefully done analysis of the Intrepid LED enlarging light source, which found that the choice of LEDs for the light source were so poorly matched to the photo enlarging requirements that a properly corrected print could not be made.
Interesting Experiment. Thanks for sharing! It looks to be more of a hassle than the standard scan flow is… Keeping a consistent scan flow between 35 and 120 6x6 3x4.5 is the dream. Between light sources, the macro lens rig and the software interpretations results swing wildly… Any tips?
This is a great idea but it seems like it requires a dedicated unit with a built-in scanning flash and specialized scanning housing with an extremely powerful diffuser to be practical at all.
I assume a flash is probably the best white light source in terms of color temperature. Ideally you’d use RGB LEDs with really good CRI like the Frontiers use. But I agree both is too much effort.
I use flash, a Godox V1 with a hot shoe trigger. It does take a few minutes to focus initially, using my phone light and focus peaking on my mirrorless camera, but after that it’s fine. I’m setup horizontally though using a tripod not on a copy stand.
I am brand new to film, I have a Valoi easy35 on the way, haven't used it yet, but I can't see why a solution like that with a flash and diffuser on the end wouldn't be viable.
Not a film specific thing, but tungsten lights have the best colour spectrum possible (according to light manufacturers). So probably the best colour capture... but you have to deal with the heat of a tungsten light, so its not super practical.
Surely this is just down to pure science. A flash gun (modern) will have very good light consistency (5000-5600K). Depending on your led light pad you may be using a source with high variability. Better light pad better consistency and better brightness = better results. Pure cost to quality equation. If you're using a 300-500AUD flashgun as opposed to a 50-120AUD led light tablet - buy a high quality light tablet and you'll see the same results surely.
Cool test, thanks for posting! IMO Sunlight>Flash>LED, but LED is super convenient for the reasons you mentioned. Results aren't that different & in most cases the pros outweigh the cons. Sunlight is the benchmark in terms of fullest spectrum, then probably followed by tungsten light, hmi, & flash. LED is lowest in terms of quality. We still haven't figured it out yet even if its 99+ CRI. The spectrum isn't as full and it peaks at different color spectrum values, as opposed to heat based light sources which are fuller/more even. After all LED is working from a binary, like digital cameras, and the gradation of the spectrum is inherently binary as well. Some newer tech like RGBWW & RGBACL is helping to flesh out the spectrum. What this all means in terms of how the digital camera perceives the different light sources? I have no idea. According to your tests, others I've done and seen, the differences are so minimal it kind of doesn't matter. Unless you are trying to raise your shutter speeds, using a really poor quality LED, etc. Using something like the nikon ES-2 and diffused sun should yield best results, but then you have to wait to scan until conditions allow. Also doesn't need tripod or copy stand. I wish there was a 120 model. RGB scanning sounds like it could do a good job technically, but the post needed to make it work... like stitching 120 scans for higher megapixels, I can't imagine it being worth it in most cases.
@Pushing Film ... Most people forget that commercial slide and negative copying in the 80's / 90s were done using a Transflash unit .. Just saying, as someone shooting film in the 70s / 80s / 90s.
Good practical analysis of flash vs LED. One more variable to consider is the actual amount of diffusion and the effect it has on contrast, edge sharpness and perceived film grain. Highly collimated light from a point source or through condensers results in higher levels of these characteristics compared to diffuse light sources. This is obvious in darkroom printing where a higher contrast grade printing paper or filter is usually needed with a diffused light source enlarger vs. a condenser enlarger for the same film negative. This is a generalization and a better understanding can be gained by searching the Callier effect.
Thanks Hashem! You make such great and informative videos! Really appreciate your hard work!
@@segzeeman7356 very kind, thank you!
I've always used flash - an AD200 in a softbox. The only disadvantage I found was that it is harder to focus, but I could use the modeling light if I were to use the flashlight head (I use the bulb) I use an on-camera film holder so I just aim at my softbox (which is already set up) and fire away. I feel like I can get through a roll pretty quickly. I use shutter speed and HSS to control exposure, because adjusting exposure in-camera is faster than on the remote.
I calibrate my exposure by shooting without any film, exposing so that the white touches the edge of the histogram.
Thanks for these tips! I've been scanning with led light but i get white vignettes, presumably due to uneven light. So I want to try with my ad 200.
Yes, trichomatic light scan is beyond good !
Color temperature and CRI only tell part of the story when it comes to light quality. The improvement in quality when using flash, especially with slide film, is due to the spectral power distribution of flash vs. LED. I've experimented with measuring the spectral power distribution of LED, fluorescent, flash, and actual sunlight with a spectrophotometer (i1 Studio) to generate graphs of the spectral power distribution of these different light sources, and the graph for speedlite output most closely mimics the graph for sunlight. The output from LED has peaks at certain wavelengths, and the output from fluorescent is spiky as heck. One can get away with a somewhat incomplete spectrum when doing negative conversions (depending on which wavelengths are deficient), but with slide film, reds and skin tones will suffer the most from poor-quality scanning light.
Interesting experiment.
One convenient way of reaching a good diffusion would be to bounce the light from the flash off a white sheet of paper.
For example, the flash pointing into a small box with a white sheet of paper facing the film holder and camera, or a sheet of paper at 45° and the flash perpendicular to the camera and film holder, as I used to do for duplicating slides with less contrast before digital photography.
I think speedlight is more convenient when we use it with macro bellows and slide/film duplicator setup
I remember back in 2014-2017 when I didn't have access to a lab providing decent scans (I was in Greece at the time), I was improvising with the equipment I had available at the studio I was assisting. I was using big studio flash lights with softboxes on the floor, and on top of that I was rigging a sheet of white plexiglass that acted as a large diffusion and then trying to photograph the negative on top of all that 😂. I'm sweating just thinking about the amount of work I was doing just scan a few frames which I would then have to manually convert in photoshop. I'm so glad analog has made a major comeback and has allowed us to have all these different ideas and products. Everybody is pitching in on it and we are collectively progressing. Quality of light is definitely very important and if the technology is not there on video lights then we could definitely find a way to rig flash lights to do the job.
Yep, we've sure come a long way!
hahaha currently in Greece and facing the same scanning problem! on my way to build a good setup ✊
@@penquito Are there any new labs in Greece or labs that have made a comeback or not really?
I’m sure you’ll make a killer setup at home. There’s plenty of resources online to take inspiration from.
I purchased an x-ray light box online from a dental office. I took out the internals on one end I cut an opening for my canon speed light. On the other end, I cut another round opening where I installed a nightlight type bulb socket that could be turned on and off with a flip switch when focusing. I lined the light box with silver foil tape. I, too, am using F 11 with 1/64 power. My color corrections have been excellent!
Nice!
That’s so smart!
I photographed my negativs with a flash directed to the wall, the film hold in a cardboard. Works really good.
Nice!
I wonder if the plexi glass, that the LED is going through, is giving off a slight color cast? -robb
Just white balance using the film edge.
Hashem from my experience with flash, the key to making this work would be to extend the distance of flash to film plane with adequate diffusion for an even spread across the film plane, while also doing as custom white balance to ensure that the diffusion is not causing a colour shift. I'd also line the box with black matte material that stops the bouncing of light within the box itself. There are clearly other variables, but in terms of flash alone, that would be it, along with the quality of the flash output, which is dependent on the model.
And why not, on the contrary, to cover the inner space of box with reflective material? This would improve the situation with the uniformity of dispersion, it seems to me...
Yeah I figured this would be the case regarding distance, as mentioned in the video. Lining the box makes sense too! But, I had to make due with what I had in terms of materials and time (for the purposes of my test). It would be cool to see people's purpose made setups using some of these techniques 😃
Love this experiment, and thank you for going through the pains of doing it and being generous with the results. I wonder also if bouncing the flash like in a book lighting setup might work (but that also means more faffing around the entire rig), but regardless this is so fascinating. Thank you again!
Glad you thought so! That would probably work well, since you can always shield light between the lens and film holder using a hood, etc.
This was interesting. I think the convenience, continuous light and low heat output of LEDs make them really good sources. That said, all light sources have differences in distribution of frequencies and LEDs are certainly a quite different light source than a flash. Don't know the details and don't have the equipment to measure different light sources, but (as you mentioned) the shutter speeds you can get with a flash would be an advantage. Maybe using a flash like the Godox V1Pro with a build in led light would simplify the setup. Keeping the flash outside the box might also work well as you could have just a reflector inside the box. Increasing the distance could make it easier to get light distribution perfect and keep the flash cooler under heavy use, but with 1/32 or less power that might not be a problem.
Looking at some of the macro flashes the part of the flash that actually generate the light is quite small and could probably be built as a fairly compact light source if someone was interested in making one specially for camera scanning. Flash is a pointier light source than a bunch of LEDs though so getting an even distribution without adding height to the source might be difficult.
Yeah something like that Godox would be more suitable I think!
I still maintain that the Cinestill CS Lite is the best of any light source I've used that's currently out. That little color setting is a shocking difference maker, and it's cheap. If it could fit a 4x5 sheet of film on it, it might be perfect.
*And I've been camera scanning for about 10 years now. Not new to it at all.
@@JamieMPhoto nice, yeah I've heard good things about that one
I ve tried a custom tunable RGB set - there was no such big difference in the result on a first look. But if you look at the channels while conversion you will see one significant thing. Because the base layer of the film has its own saturated tint (yellow for example) you get less information in blue channel unless the color temperature of your light is cool enough. Even if it is around 6000-6500k it still gives a lack of info in blue channel. That causes excessive noise in blue channel while converting. RGB or a very cool light (even through a cyan gel filter) gives you slightly better results due to a better base layer exclusion (the more grayish you make it in pre - the clearer will be the result in post). So you can use a cyan gel filter on a flash or make your rgb light as cool as it needs to be for your negative.
Interesting!
I've been using a flash for a while. I started because my scanning light (which was expensive) was not giving me very fast shutter speeds at 100 iso and F 5.6. Unsurprisingly, flash kicks constant lights's butt in terms of brightness. I use a mixing chamber from an old enlarger to diffuse the light. I will keep using the flash for black and white and slide film, but I'm working on building a custom RGB light for color negative film scanning
Great video Hashem!
Totally speculation on my behalf, since I don’t use flash, but I think the more neural colours that you’re getting from the flash are due to the flash being a better quality of light source. Much like that which you’d get from the sun or a tungsten bulb.
Unless you’re using a really high quality LED fixtures, you’ll find might find a minor green or magenta shift in LEDs. Even if the LED has a high CRI and TLCI ratings that doesn’t really show the whole picture regarding the quality of the light.
From my understanding SSI ratings are a much more accurate way to rate LED fixtures.
If you want to go on a deep dive the channel @gaffergear has some great videos on LED lighting.
As your research suggests using RGB lighting will probably allow you to get better quality scans as you can dial the rgb values in. They typically have a +\- green and magenta function.
If you were to go do the route of RGB experimentation I’d suggest giving a light fixture with RGBWW LED diodes a crack. These have an additional warm white LED. Which can help improve the light quality.
You should give the Nikon ES 2 a try. Use it with natural light, and no need to worry about camera shake or alignment.
Yeah. The sun is a pretty underrated light source 🌞
LEDs generally do not have a continuous spectrum output of even intensity. This applies to white and color LEDs. Flash will have a more balance spectrum. However, the function differences are often quite small, and they can be adjusted in a digital processing world. On the other hand, the same cannot be said for using LEDs as a light source for color printing on traditional photo paper, which is more sensitive to gaps in the LED spectrum. This was illustrated recently in a carefully done analysis of the Intrepid LED enlarging light source, which found that the choice of LEDs for the light source were so poorly matched to the photo enlarging requirements that a properly corrected print could not be made.
Interesting Experiment. Thanks for sharing! It looks to be more of a hassle than the standard scan flow is…
Keeping a consistent scan flow between 35 and 120 6x6 3x4.5 is the dream. Between light sources, the macro lens rig and the software interpretations results swing wildly…
Any tips?
This is a great idea but it seems like it requires a dedicated unit with a built-in scanning flash and specialized scanning housing with an extremely powerful diffuser to be practical at all.
I assume a flash is probably the best white light source in terms of color temperature. Ideally you’d use RGB LEDs with really good CRI like the Frontiers use. But I agree both is too much effort.
I don't know - your examples showed the scans look imperially better. I think someone mentioned a standard strobe and modeling light.
Some looked slightly better, but some looked indistinguishable.
I use flash, a Godox V1 with a hot shoe trigger. It does take a few minutes to focus initially, using my phone light and focus peaking on my mirrorless camera, but after that it’s fine. I’m setup horizontally though using a tripod not on a copy stand.
Nice!
The flash examples seemed to look sharper? Was that your opinion?
In some examples I did notice that, most likely due to the higher shutter speed
I am brand new to film, I have a Valoi easy35 on the way, haven't used it yet, but I can't see why a solution like that with a flash and diffuser on the end wouldn't be viable.
Not a film specific thing, but tungsten lights have the best colour spectrum possible (according to light manufacturers).
So probably the best colour capture... but you have to deal with the heat of a tungsten light, so its not super practical.
I'd be interested how good scanning the film when held and lit by a colour enlarger would look.
I've tried using an enlarger negative holder, it can work well. Just slow, and prone to more dust + newton rings if you use a glass type.
Not dealing with cycles in the LED, coupled with the non-instantaneous shutter speed is my theory (stroboscopic effect).
I shoot with a negative holder in front of the camera and the off flash bounced to a white wall… And works totally fine for me :)
nice!
Interesting! But I do not like the hassle of setting up such a situation.
Yeah! I think if you can get past the initial setup and space dedication, it could be quite smooth.
If you have a flash with a modeling light, that would make things a bit easier
Surely this is just down to pure science. A flash gun (modern) will have very good light consistency (5000-5600K). Depending on your led light pad you may be using a source with high variability. Better light pad better consistency and better brightness = better results. Pure cost to quality equation. If you're using a 300-500AUD flashgun as opposed to a 50-120AUD led light tablet - buy a high quality light tablet and you'll see the same results surely.
Not necessarily the case, since these comparisons were with a Negative Supply Light source in that price range.
A slide duplicator and a flash has worked for decades. Sunlight will work as well.
Cool test, thanks for posting!
IMO Sunlight>Flash>LED, but LED is super convenient for the reasons you mentioned. Results aren't that different & in most cases the pros outweigh the cons. Sunlight is the benchmark in terms of fullest spectrum, then probably followed by tungsten light, hmi, & flash. LED is lowest in terms of quality. We still haven't figured it out yet even if its 99+ CRI. The spectrum isn't as full and it peaks at different color spectrum values, as opposed to heat based light sources which are fuller/more even. After all LED is working from a binary, like digital cameras, and the gradation of the spectrum is inherently binary as well. Some newer tech like RGBWW & RGBACL is helping to flesh out the spectrum. What this all means in terms of how the digital camera perceives the different light sources? I have no idea. According to your tests, others I've done and seen, the differences are so minimal it kind of doesn't matter. Unless you are trying to raise your shutter speeds, using a really poor quality LED, etc.
Using something like the nikon ES-2 and diffused sun should yield best results, but then you have to wait to scan until conditions allow. Also doesn't need tripod or copy stand. I wish there was a 120 model. RGB scanning sounds like it could do a good job technically, but the post needed to make it work... like stitching 120 scans for higher megapixels, I can't imagine it being worth it in most cases.
to me with flash looks way better if you gonna scan with a camera and not a scanner
Interesting, in my tests I didn't find it "way better" at all!
@Pushing Film ... Most people forget that commercial slide and negative copying in the 80's / 90s were done using a Transflash unit .. Just saying, as someone shooting film in the 70s / 80s / 90s.
You can probably 3d print a rig for easier setup