What's the point of the exposure compensation setting on the Wolverine 1080p scanner?

[Brian Daniels]

Hey there,

This is actually my first post.

So, like many people here, I've inherited a couple of Super 8 film reels that I'd like to digitize. I'm still debating whether I should buy a scanner--like the oft-mentioned Wolverine--and do it myself; or, whether I should pay a professional to do it for me instead. I've seen some YouTube video reviews of the Wolverine scanner where they go through the settings on the Wolverine, and I'm having trouble understanding what's the point of offering people the ability to adjust the exposure via the exposure compensation setting.

I realize the Wolverine doesn't allow people to adjust the aperture, shutter speed, or ISO, and only has the exposure compensation setting to adjust the exposure, but isn't exposure frame/scene-specific anyway? I mean, as far as I can tell, it's not like the Wolverine allows people to change the exposure, using the exposure compensation setting on the machine, right before a Super 8 film frame is scanned anyway. You would have to decide on a certain exposure prior to starting the scan, regardless of if the scene changes, and stick with that, so, given that exposure is frame-specific, what's the point of offering Wolverine owners the ability to adjust exposure through the exposure compensation setting in the first place?

[Joerg Polzfusz]

Hi!

Yes, a „real“ exposure compensation should be based on a „per scene“-logic. The feature that is implemented in the Chinese Winait scanners (whether they are branded as Wolverine, reflecta, Somikon, Kodak, …) only makes sense for films shot with defect cameras that permanently over-/underexposed all scenes or for „darker films“ (some films like the Tri-X have a more „greyish“ base instead of a clear, transparent base like the Kodachrome 40).

Jörg

[Brian Daniels]

Thanks Jörg. Your explanation makes perfect sense.

[John Bakker]

Yes, it's scene specific but nobody forces you to set an exposure and tranfer a whole reel with that setting. If you have a reel with under- and/or overexposed scenes the choice is up to you to transfer these scenes with an adjusted brightness setting. It takes a bit more effort but in the end will give better results.

[Brian Daniels]

Yes, it's scene specific but nobody forces you to set an exposure and tranfer a whole reel with that setting. If you have a reel with under- and/or overexposed scenes the choice is up to you to transfer these scenes with an adjusted brightness setting. It takes a bit more effort but in the end will give better results.

Sorry for the late reply. After Jörg's reply, I assumed the matter was settled.

I don't have a Wolverine in my possession, so I can't really test this out myself just yet, and that's assuming I were even to get one, but, if I read your reply correctly, how would you go about using the exposure compensation setting on the Wolverine to correctly expose multiple scenes that were under/overexposed? Does the machine give you enough time to use the exposure compensation setting before the frame gets captured, where you can see the frame on the LCD screen prior to capture? Or, do you have to slightly modify the Wolverine with your own feed reel and takeup reel, like some of the setups I've seen out there, where I'm assuming you get to decide when the film advances? If you do decide to use your own feed+takeup reel, can you tell the machine to refrain from scanning the frame, so that you still have a chance to inspect it on the LCD screen, which would allow you to adjust the exposure compensation prior to capture?

Sorry if these appear like newbie questions. This is all still very new to me.

[John Bakker]

Sorry I didn't see the reply Brian. One has to subscribe to the topic to receive notifications (would be nice if you could receive a notification when someone quotes your post).
No, the machine doesn't give you enough time to do that, hence I said it's more time consuming. You would have to record every scene separately, setting the brightness at the level required for that scene (or reel of film if it's over- or underexposed equally all through the reel).

[Brian Daniels]

Sorry I didn't see the reply Brian. One has to subscribe to the topic to receive notifications (would be nice if you could receive a notification when someone quotes your post).
No, the machine doesn't give you enough time to do that, hence I said it's more time consuming. You would have to record every scene separately, setting the brightness at the level required for that scene (or reel of film if it's over- or underexposed equally all through the reel).

Thanks John for getting back to me so soon. I really appreciate it.

So, if I understand you correctly, if one were to use the exposure compensation setting on the Wolverine in order to make sure that each scene is exposed correctly, does the machine still give you the chance to look at each frame on the LCD screen before you start recording, say for an underexposed scene, or do you have to manually inspect the scene yourself with a light source and a magnifying glass, before placing the reel on the machine? I'm assuming the former, but, again, I don't own one of these machines.

[Nantawat Kittiwarakul]

Yes it indeed does give the preview on the screen before starting the scan. But considering its lack of real manual exposure setting I highly doubt if that exp. compensation is of any use.

I can think of only 2 possible solutions.

1. Babysit throughout the scan, keeping your eyes on the preview screen. If any particular shot seems improperly exposed (too dark/bright) then stop the scan, adjust the exp. compensation, move back a bit to the first frame of the shot, then resume the scan. A very time consuming of course...
2. Or just do multiple pass for the entire reel with exposure bracketing. First pass on -1 settings, second pass on 0 settings, third pass on +1 settings, kinda that. Then choose the best looking for each scene from each pass and assemble them together. This is probably even more time consuming than the first option, but at least you don't have to babysit it throughout the whole process.

[John Bakker]

Nantawat, even if it would have full manual control, the process would still be the same.

[Nantawat Kittiwarakul]

Nantawat, even if it would have full manual control, the process would still be the same.

For its VERY limited dynamic range this might be the only way to get decent result. Thank goodness I eventually didn't go this route and headed to DIY path instead.

[Brian Daniels]

For its VERY limited dynamic range this might be the only way to get decent result. Thank goodness I eventually didn't go this route and headed to DIY path instead.

Nantawat, since you're still pretty active in this thread, and since you brought up the topic of dynamic range, I thought I would ask you about a post you made that I stumbled upon a while back, when I was researching the pro & cons of the Wolverine MM Pro. Your post is the second post down.

When you say that the latitude is quite limited on contrasty images, where you can get either highlight or shadow detail, but not both, are you referring to the dynamic range of the Wolverine?

[Nantawat Kittiwarakul]

Definitely YES.
All Wolverine scanner and its variant seems to have just about 4-5 stops of dynamic range - about on par with camcorder from 2000's. Considering today's mirrorless camera would have 10-12 stops even the midrange ones, that's one of reasons why I chose DIY route and wouldn't go back.

[Brian Daniels]

Definitely YES.
All Wolverine scanner and its variant seems to have just about 4-5 stops of dynamic range - about on par with camcorder from 2000's. Considering today's mirrorless camera would have 10-12 stops even the midrange ones, that's one of reasons why I chose DIY route and wouldn't go back.

I'm still confused on whether the bit depth of an imaging device, like the Wolverine, can affect the dynamic range, if at all. I mean, I know what dynamic range and bit depth are, I'm just not sure if bit depth has any influence on the dynamic range of the resulting final image. If dynamic range, which is measured in stops, refers to the range from the deepest black to the brightest white, and bit depth can be thought of as the number of steps between those two extremes, then wouldn't a low bit depth result in some tones near the extremes clipping to pure black or pure white, resulting in a low dynamic range image?

[Nantawat Kittiwarakul]

In layman's term, it's nothing more than a poor image sensor that can't capture extreme white & extreme black at the same time, plus sub-par image processing.

In order to have better understanding let's see these frame grabs.



The first image is from Reflectra scanner (Rebadged Wolverine). Its AE decided to retain highlight on women's shirt, and crushing all detail on men's black shirt. Without comparison this would seems to be as much detail as can be captured from the film.

But let's have another point of view...



The very same film frame, captured by my Panasonic GH5 as still images, then process & output as video file later on. While the highlight level is about the same, but you can see much more detail in shadow area. Overall detail is way better without aggressive sharpening applied too.

(BTW this is just a test film that I had been used/abused for years, hence loads of scratches in the the later image.)


While technically "identical" (both 8 bit depth, with 1080 pixels vertical resolution) the end result is still vastly different. Comparing the resultant video files instead of a single frame like this, the gap is even much wider. So yes, using good scanner will give you a good start - and more likely to result in high quality output.

[Stan Jelavic]

This is a pretty complicated subject since different components of the system change the original image dynamic range. The Wolverine sensor dynamic range is 70 dB or 12bit which is quite adequate.
BTW - every additional bit doubles the resolution i.e. 1 bit 2 states, 2 bits 4 states etc.
Doubling up corresponds to 6dB in the log scale since our eye is nonlinear.
So 70 dB is roughly 12bits ... 70/6 = 12 approximately.
There are two modes, 12bit and 10 bit.
For 10 bit the sensor does compression internally and that degrades the dynamic range depending on the algorithm used.
The Wolverine Novatek chip that processes the sensor data can run in 12bit or 10bit mode and I do not know
what the Wolverine designers picked.
Even if the Novatek chip accepts the 12bit from the MIPI interface it is still a question what the H264 on chip codec does. It says that it supports WDR (wide dynamic range) that jams the 12 bits into 8.
And in the end most of the monitors at that time were 8 bit so the final content is 8 bit per channel.
So it is art how to jams 12 original bits into 8. Some systems do it better than others as Nantawat illustrated.
Combine all this with the auto exposure control...

One more note, there is no dynamic adjustment of exposure in Wolverine MM which really sucks. You have to stop the scan go beck in menu and readjust the exposure. In addition the auto exposure will sometimes "hunt" causing video flicker depending on the scene.