Perry Paolantonio

jeez. the color is awful on that. Again. Kodak keeps promoting S8 with these videos that look awful. I don't get it.

"all look pretty good" is different than what you said before - that different machines will result in the same thing, given the same input parameters. They won't unless they're all running identical cameras, firmware versions, software versions, and used the same grading settings in the scanner when scanning. I'm not saying each one can't do a good job, but what I am saying is that we worked extensively with Lasergraphics to try to deal with some of the noise issues, and led the push on getting those problems addressed. Not everyone has up to date systems, so one can't say that they're all going to produce the same thing. Considering these machines have been out there for about 4 years now, there could be dozens of permutations of software/firmware versions, camera versions, etc. And I wish I could post an example of one film we did recently that came from another scanstation, but I can't so you'll just have to trust me. It was night and day, and they even sent me a screen shot of their scan parameters so I could try it on mine. My hunch is that the problem was a combination of camera version, firmware version and software version. because ours had recently had all of those upgraded, and we were able to produce a completely different (cleaner, easier to grade) scan, using mostly default settings.

One would think. But as you said: We have seen scans made on a ScanStation that I couldn't believe were made on the same machine we have. We rescanned the same film and it looked completely different (better). Much like the old days with telecine, where every facility had a different setup and you really couldn't compare the output in an apples-to-apples way even if they were using the same telecine, not all data scanners are the same. There are subtle differences based on firmware revisions, software versions, calibration, and scan settings. While it seems like Scanner A and Scanner B should result in the same thing with the same scan settings and the same film, that's typically not the case in my experience.

We use acid-free artists tape. Paper backing, low tack adhesive (that doesn't leave residue on the film), and it's something you can order on Amazon in a variety of widths. We use 1/4" for 8mm and 3/8 or 1/2" for 16mm/35mm. A roll lasts a really long time, so don't be put off by the cost of the tape ($10/roll or so).

Hi Byron, We haven't had issues with DIY processed film in our scanstation, even film that's got a lot of variable density. When the film is really thin, and it's hard to identify the perfs from clear acetate, there are a few tricks we can try to get it through the machine. Have you tried scanning it on a ScanStation yet? If not, we'd be happy to give it a try. -perry

As a general rule, if you can unspool the film, you can scan it on the ScanStation. You start to have problems when the film is severely cupped (the edges of the film curling towards one another, causing the film to not sit flat in the gate) *and* if the film is dry and brittle. Some cupping is actually ok, because of the natural tension placed on the film in the scanner's gate, which flattens it out, as long as the film isn't rigid. If the film is pliable, it should be scannable, even with moderate cupping. If your budget allows, I would highly recommend scanning at 4k, even if you don't need 4k right now. With UHD televisions under $300, this is clearly the direction things are headed, and if you scan at a lower resolution, you're going to need to scale it up to fit those screens. Upconverting always results in some softening of the image, so you're at a disadvantage right off the bat, if you scan at 2k or lower.

IR flash won't do anything unless the hair is physically on the film, or stuck in the gate of the scanner. If it was in the camera gate, it's part of the picture, and IR won't do anything. An IR pass works by preventing light from hitting the film wherever there's physical stuff on the film. The result is a B/W map of that dust that some restoration software can use to target and remove dirt. But it only works on dust that's on the surface of the film when it runs through the scanner, or hairs in the scanner gate, etc. It won't do anything if there was hair in-camera.

In your budget? No.

I've only heard this from scanner manufacturers, on machines that have fixed camera positions, relative to the film. So smaller gauges would be a crop of the sensor, kind of the way the Blackmagic Scanner or some others do it. In those cases, the really small gauges are referred to as "1k" but only by the engineers and marketing people for the manufacturers of these things, not by the people in the industry actually using them.

I missed that one. Hilarious. We scan 16mm for emerson students all the time.

You can believe what you'd like. It would be to everyone's benefit though, if you didn't go around spreading misinformation.

Not here, and not with any of the clients we've authored DVDs or Blu-rays for who handled their own transfers and/or color or restoration. I could easily list 50-60 titles off the top of my head from the past couple of years that I know for a fact to not have had any noise reduction applied anywhere in the pipeline from film to Blu-ray and/or streaming. My point is that making blanket statements about this kind of thing is not a good idea, because what you're saying is demonstrably untrue. -perry

I've been authoring DVDs since 2000, and Blu-rays since 2007, with about 1000 titles under my belt. There may be a few labels that still do this, but saying "almost every movie" is way over the top. In 17 years, we've never used noise reduction on any discs we've authored. We don't do it on film scans or when grading unless the client insists, and in those cases, we've requested that we not be listed in the credits. -perry

If you're talking about brand new film, you're not going to gain much from an ultrasonic cleaning, vs a non-immersion cleaning. Basically, there are two types of machine cleaners: non immersion cleaners use buffer rollers saturated with a solvent of some kind to clean and dry the film as it passes through. This will eliminate virtually all of the dust on the film, which is the main concern with freshly shot/processed film (dust picked up in the lab, when the film is exposed to air). Film is a dust magnet, and no matter how clean the lab is, it's there, flaking off of us all day long. Dust looks way worse on negative than positive film, because it shows up as white spots. An ultrasonic cleaner adds an additional step: the film is threaded through a pool of solvent liquid, where ultrasonic vibrators agitate the solvent. This "scrubs" the film clean at a molecular level. This is very effective, but it's really most useful for film that has a lot of stuff on it: prints, film that's been through a printer a few dozen times, any situation where caked on debris of some sort needs to be removed. The major downside of the ultrasonic cleaners (with a few exceptions that use the newer more environmentally friendly, but ridiculously expensive HFE solvent) is that they require pretty nasty chemicals to do their thing. Perc is basically cancer in a barrel and can cause neurological issues if not handled properly (ventilation, respiration, gloves, etc), not to mention environmental concerns with proper disposal. All of the non-immersion cleaners I'm aware of use pure Isopropynol - 99.8% or purer rubbing alcohol, which is significantly more benign. That being said, they're both pretty effective methods of cleaning film, with ultrasonic being the better choice for film that's really messy. Fresh film doesn't need much to clean it up.

Modern encoders for DVD and Blu-ray don't require the kind of brute-force noise reduction or brickwall filtering that was necessary 15 years ago. By "modern" I mean encoders from the past 10 years or so. Yes, in the beginning of DVD, the encoders weren't efficient and would choke on anything random: video noise, film grain, etc. But the past 10 years brought in a wave of high quality encoders for MPEG2 (DVD) and AVC (Blu-ray) that don't decimate the grain and don't require that you do "extensive" (or really any) noise reduction before encoding. In a theater, you're not going to be seeing Super 16 projected, and you're pretty unlikely to see regular 16 projected (in most theaters). So are you talking about comparing it to a 35mm blow-up? Because that's not apples-to-apples either. It's also worth bearing in mind that the photochemical print process itself is responsible for a fair bit of grain reduction. We hear this a lot from clients who have only ever seen their films projected, but when they're scanned at 2k or 4k, they're surprised by how much grainier they look than what they remember. This is because they're not used to seeing the grain that's always been there, because of the natural grian loss that happens when you print the neg. The type of transfer, the resolution of the transfer, the method of viewing all come into play here. An SD transfer is going to be less grainy looking than an HD transfer, simply because of the higher pixel density of HD, which makes a more detailed representation of the film. Same goes for 2k or 4k. Modern film stock isn't grainier (quite the opposite in some cases), but high resolution scanning tends to highlight that grain more than low resolution scanning.

16mm optical is mono. Always. The way the tracks work, a light is shone through the film and hits a photocell that turns the variations in the amount of light hitting it into variable electrical signals. This is then turned into a sound wave (oversimplifying here, but this is the gist). It doesn't matter how many tracks you see when you look at it, because they're all doing the same thing, and it's just about varying the light levels that hit the sensor. There are no discrete tracks that are read separately, it's all mushed into one, thus it's always mono. Stereo existed in Mag audio, but that was more for output to video or blowup to 35. I used to have a steenbeck with stereo heads (which was a real pain, since my stuff was in mono. It worked, but there was a lot of unnecessary noise from the unused second track.

Since you're working with Black and White, either tool will work fine, but Resolve is going to be much more suited to the task, because that's what it's built for.

I don't believe this is shipping yet, but I could be wrong.

We recommend having the lab prep for transfer. The film will be consolidated onto a larger reel, the appropriate amount of leader added, and the film is cleaned. This is less expensive than having us do the prep and clean on our end. In terms of mailers, we usually recommend not doing that, for the reasons Nick outlined. When the film arrives from the lab consolidated, it likely will take up more space than a small stack of separate reels. We don't mark up shipping, so it'll be included on your invoice as the actual shipping cost we pay to USPS or UPS, or if you prefer, FedEX. Turnaround time on our end is typically 1-2 days from the time the film arrives, depending on how busy we are at the time.

The problem is that that misses the point: Yes, at a certain scan size you hit a point of diminishing returns. If you're shooting on a Logmar, that point is much higher resolution than something shot on a cheap camera with low quality lenses. Good transport, good glass, pressure plate = more ability to pull resolution out of the picture. But it's largely academic once you're past 2k - you're not going to see more detail at higher resolutions in most cases, in the way you would if you scanned a film at, say, SD and compared to a 2k scan. What you are gaining with a 4k scan is future proofing. You're ensuring you won't need to (or the television/media player won't need to) scale the image up to fit a 4k screen. While upscaling algorithms have gotten better, a file that's been scaled up from 2k to 4k will not look as good on a 4k display as a native 4k scan would. No matter how good the algorithm, it still has to create picture where there wasn't picture, and that's always going to result in a softer image.

Sure - though back then you could buy a roll of Super 8 in just about any store that sold film (and that included grocery stores). Try finding a physical store that stocks film these days...

Yes. Sort of. If you're scanning from neg, it's log. If you're scanning from positive, it's linear, but (at least we) typically scan it flat to ensure that nothing is crushed or clipped in the scan. That is, the contrast is lowered, and we ensure that none of the highlights blow out or the blacks get crushed. this gives you maximum flexibility later in grading.

We don't do any grain management at Gamma Ray Digital. We can scan 35mm on our ScanStation or our Northlight scanner. Grain is what makes up the image, so you don't want to do anything to alter it at the time of transfer if you can help it. If you really don't like grain, take it out later, not when you're scanning. That defeats the purpose of a data scan, which is to get you the most flexible digital files from your film, so you can do what you want with it later. By the way - there's so much brutal compression on a YouTube video, including brickwall filters and various things that soften the image to reduce randomness (which makes for more efficient compression), that you really can't use it as an example. For one thing - everyone will see a different version, depending on the speed of their internet connection, and it will always look less grainy on smaller screens than on big ones. You'll get good looking scans from all the machines listed above, but transfers from the Spirit will likely come to you color corrected and possibly grain managed (you'll need to ask what they do). All the rest of the scanners you list are data scanners, so you'll get flat (log) scans from those, and then color correction happens post-scan. -perry

Scanning to 720p as your master format doesn't make a ton of sense, except as a low-res access copy for quick viewing. The actual image area for 720p is 960x720 if you pillarbox (which should be the case, because HD is always 16:9 and that's not an aspect ratio that matches any film format, unless you're cropping to 16:9 in the camera's viewfinder when shooting). We recommend 2k as a minimum, because it can match the aspect ratio of the original film. This gives you much more flexibility to crop and reposition the image, if your output format is HD. For more flexibility, 3k or 4k are even better. The primary argument, to my mind, for scanning small gauge film at 4k is not that you're going to extract more picture out of the film, but that you're going to avoid scaling an image up later. If you scan at HD or 2k, but want to view it on a 4k television (which are now under $500, and will be completely replacing 1080p displays in the coming years), something needs to scale that image up at least 4 times. Could be you, in software, or it could be the television or a media player of some kind. But something needs to scale that up, and that device or software has to make up picture that was never there, interpolating new pixels. The side effect of this is always that your image gets softer. An alternative workflow: Buy film from Kodak: $26 Process film at Cinelab: $18 Scan film at Gamma Ray Digital (2k): $27.50 $71.50 total cost not including shipping. If you ship via Priority Mail from Cinelab to us, and then from us to you, you're looking at another $7 each way, so $85 with shipping.

Yeah, but I think what the original question was about was basically making an R8->16mm print, which I'd assume means a blowup. Even if you did a contact print and then slit that, you couldn't project it in an R8 projector, because you wouldn't have enough perfs. Of course, I might have misundersood what Luigi was asking about.