Strange vignette or bleed halfway through a roll - A7CR

[Timmy]

Hi, after some help with camera scanning my negatives.

Camera: Sony A7CR - using 4 and 16 shot mode - ARW → DNG with PixelShift2DNG
Lens: Sigma 70mm Macro, 7Artisans 60mm Macro (APS-C)
Film holder: Valoi easy35v2
Conversion: NLP, basic colour profile
I usually shoot tethered, but have also shot some untethered, and the copied the files over.

I have been chasing an issue where when I get near the end of a roll of film I start to get some strange vignette seemingly bleed into the frame. On the colour films it comes up as red. I originally thought I had cooked the development - they were on the end of the roll, on the outside of the reel, would have ‘moved quicker’ with rotary processing. It also seemed to comes from the sprocket holes. So I thought was that.

But … I couldn’t see it clearly on the negative on a light table, but I just put that down to my eyes and it being really hard to pick up that kind of shift in colour in a small negative.

Also, the start of the film was fine, no issues at all. It is only when I get at least half way through, maybe more, that I start to see it.

Fast forward a bit, and I am seeing it on more rolls of film where I was sure I hadn’t cooked the development - including black and white film:

So I went back to some of the negatives and tried scanning them again.

And they came out perfectly clean. No colour shift vignette on bleed on the edges.

I thought it might have been the Sigma macro lens, so did some tests with an APS-C macro lens, but then also re-did some shots on the Sigma - all which turned out perfectly fine. Tried changing the settings. I check the lens with a test negative, F11 seemed to have the best performance in terms of sharpness and colour fringing on high contrast edges.

I uploaded some images taken with different settings here.

It strangely feels like something is overheating when shooting an entire roll - hence the ones near the end of the roll start to exhibit the behaviour. I haven’t noticed it as much when shooting a 24 exposure roll.

When shooting a roll of Pentax 17 frames I just do two frames at once, so still 36 shots per roll.

There are some colour differences in the images above, but that is just me being lazy with rescans and leaving things as default when clicking convert.

Any ideas on why about halfway through a roll of 36 exposures, when taking 4 or 16 shot pixel shift images, I start to get strange colour fringing / bleeding?

Thanks

[Digitizer]

Welcome to the forum, @Timmy

@Timmy

Any ideas on why about halfway through a roll of 36 exposures, when taking 4 or 16 shot pixel shift images, I start to get strange colour fringing / bleeding?

Have you tried scanning film … beginning at the end? Doing a scan from frames 1->36 and one from 36->1 would reveal if we’re dealing with an issue that builds up over time. If the bleed is with the same frames, it’s due to an issue with the film.

Repeat the test with a different holder and backlight, e.g a phone or tablet with some extra diffusion or distance - or simply a lit wall… Do you still get the issue?

By changing parts and timing of your scanning gear and routine, you can find the cause of the issue. And If I remember correctly, there were issues with the valoi 35, but I don’t remember whether what version. Search for post about it. This also helps to promote you as a new member.

[Timmy]

Hi,

Going back and scanning the single frame of the kangaroos just in its own and there were no issues at all. Same with the others. Camera off for a while, turned it on. Scanned some individual frames for a second time, no issues, so I’m pretty sure it’s not the film.

Thanks.

[Digitizer]

@Timmy

I’m pretty sure it’s not the film

Looks like it.

Scan single frames, one per hour? Not really. Is your light running hot after a while? LEDs like cool and, depending on how the light is made, some of the LEDs might turn off for protection.

Can you remove the Valoi light and use something else instead?

Cameras don’t like heat either, but I can’t imagine a camera running hot will create that falloff…unless it’s engineered to do so. I’d not expect such a thing from Sony though.

[Harry]

The first version of the easy35 had potential issues with vignetting, or at least the edges of the holder could cause issues with certain lenses. This latest v2 has a completely different light source, a kind of mini ‘integration sphere’ that is supposed to have addressed this.

"The Universal Light Source (patent pending): A leap from 81% to 99% light uniformity. This ensures edge-to-edge illumination even with wide-angle lenses, effectively eliminating the “orange corners” and vignetting issues common in home scanning."

[ArnoG]

I would contact Valoi about this. Their head (Arild Båsmo) is very responsive. Indeed, as @Harry already mentioned, their version 1 had light fall-off towards the edges, similar to all finite planar light sources (basic physics: https://www.arnogodeke.com/Blog/Homogeneity-of-planar-light-sources-for-film-scanning). In their version 2, they introduced an integrating half-dome which addresses and fixes this issue. In general, LED output does depend on temperature and color can drift as the temperature changes. I notice this in my own custom built trichromatic light source which has to warm up a while before it becomes stable. Maybe it’s related to that, maybe not, but I’m sure Valoi would appreciate the feedback if it turns out that their version 2 has a potential problem (I’m not saying that this is the case though)

[Digitizer]

@ArnoG

LED output does depend on temperature and color can drift as the temperature changes.

…and these shifts would necessitate different exposure and white balance settings.

Only if spacial characteristics changed could a spacial shift as your bleed be caused. This could be the case if there were/are several LEDs with different temperature dependencies. Whether the Valoi light is built with several LEDs, I cannot say though.

Planar light sources have some falloff, and with a panel that contains multiple leds, the total intensity curve looks like this:

(imagine LEDs in positions -7, -6, -5 … 5, 6 and 7)

Here’s the formula:

(contribution per source fades with the square of the distance via Pythagoras)

[Harry]

I think that there is also an ‘edge’ effect that is often difficult to diagnose but seems to be associated with masking and/or the position of the diffusor, the shading follows the line of the negative mask. From the image posted above of the car and garage that looks to be what is going on there rather than any vignette effect from light fall off. I have no idea why it might come and go though as seems to be the case here.

[Digitizer]

@Harry

the shading follows the line of the negative mask.

If some light is reflected back by the negative holder, the dark “inner frame” will appear with the conversion. Therefore, a test should be made with the same negative scanned without a holder.

I tested scanning negatives lying directly on the (Kaiser Plano) panel with only rough cut cardboard or paper masks to reduce stray light. None of these scans had such edge effects.

If we look back at enlarger film holders, we see book style items made from sheet or cast metal parts and none of the ones I have exceeds 2mm thickness. But that is more of a general idea to experiment with than a direct solution for the OP.