Good to know. Thanks for posting that, @Roger_Evans. BTW I got one of those Keyence sensors and a few other cheaper ones to try out. I’m really hoping the cheaper ones work but I’m not holding my breath based on what you’ve told me.
Slow as molasses over here, but steady onwards.
Hi Matt,
It looks great, but I agree, too costly. Happy to wait for one of you guys to come up with something more affordable. I’m loath to put anything pristine through a bunch of gates and cogs, so transport simplicity will be something I’m looking for.
Regards,
Motes
So a few criticisms to make about the Retroscan. I agree with the points that @Owlinsky has made, and I do know some people that own Retroscanners and are happy with them.
Roger has made repeated comparisons to the “$250K” Lasergraphics Scanstation, but this is obviously missing the main competitors for smaller formats which are the Filmfabriek HDS+, and the 16/8 Scanstation Personal (being replaced this year by the Scanstation Archivist starting at $40K). There is also the Blackmagic Cintel 2. None of those machines cost $13K upfront, they all offer a 24 month payment plan (lease plan) at something like 6-7% interest. Any professional company with a full $175K+ Scanstion runs their machines 24/7 so they get 3x the value out of their equipment compared to a business running standard 8hr days 7 days a week, so for a professional company compared to a “home movie transfer” shop or an average film archive it’s more like having 3x Scanstations that only cost $60K each.
It’s one thing if the owner of the machine is scanning their own films and is happy with the quality (usually these people have at least put in their own camera), but it is another matter to be offering a commercial service to people - the “home video to DVD” transfer market. This is a market of yesteryear and there are a lot of end customers who find it very frustrating and difficult to get their films scanned in good quality at a price they can afford. Let’s just say these customers who often get their same films re-scanned elsewhere are not fans of those transfer houses, many such customers have had poor experiences and will use very strong language about this. The core complaint however can be restated that they don’t want to pay for a sub-par service, and for other professional services it’s a lot easier to find a competitor who can provide the service at the quality that the customer expects (for example a new kitchen from a joinery, landscaping, auto mechanics, dry cleaning, photographer, etc).
This is something that we have all experienced at some time or another, there was a few years ago that I had a gas appliance “serviced” paid for the service and it wasn’t even fixed… I had to get it serviced again, this time properly, and pay for the service a second time. My brother took his car to a mechanic to get it fixed - they caused a massive problem with the transmission hose and claimed they needed to replace it but that they’d fixed it temporarily. After it burst off the next day he took it to his regular mechanic, who explained that they didn’t know what they were doing and that he doesn’t need an after-market hose. The problem was fixed, this time permanently by attaching the hose correctly. I reckon everyone here would be able to recall times they’ve had to pay for a service to be delivered twice.
Scanning prices as a service vary extremely widely and in my experience price and quality are not aligned. The prices that most of these transfer houses charge for the service isn’t a single cent less expensive than scanning on a decent quality machine. No matter how good the machine you can get poor results. You also want someone who knows how film should look instinctively operating the scanning machine, otherwise they may not notice when something isn’t right. This of course is true for other services I’ve mentioned as well - painters that don’t do the proper prepwork for example and the paint peels after 12 months and the customer has to pay another painter to do the job properly.
In summary my main criticism is that the Retroscan has flooded the market for a service it is no longer suitable for in 2021. If $40K is too much to pay for commercial equipment (whatever the industry) to deliver commercial services to paying customers, then in my opinion the business model is wrong.
Totally agree @filmkeeper, I get frustrated at the amount of people offering scanning on a retro scan/universal with terrible skills, outputting very poor scans. Unfortunately the customers have no idea that there is a variance with digitising, they just think everywhere that offers “film to DVD” is the same, just like dropping some film off to be developed at the chemist, it didn’t really matter what chemist you chose.
I own a MkII, and thought the scan quality was much better than I was expecting.
The biggest issue I see is people overexposing the scan, not filling the sensor with the frame (they crop in), and only closing an AVI output from the bundled software (very fast), which severely handicaps the scan from any further work. If the operators chooses the TIFF output (which is VERY slow), there’s great range for making adjustments.
I made a post about this on the Moviestuff Facebook page but gave up.
The gamma level is very low by default on the scanning program, when I expose for the highlights it looks like I’m capturing blackness. I think this throws off the users, so they blow the highlights to bring up the shadows.
I had someone that had some film scanned recently by a place in Syd Australia and they only delivered to her on a hard drive in SD. it’s crazy.
If you join the Moviestuff universal owners Facebook group, you’ll see the lack of scanning and film knowledge most of the owners have. Don’t get me wrong, there are plenty that know what they’re doing. But so many just buy the scanner with no idea about film, and expect they can make a good buck doing home movie scans.
This is so true, here in Australia, you can get a scan on a HDS+ in Canberra for the same price (if you’re willing to splice to a larger reel to keep the cost down). I got a sample reel scanned when I was considering what scanner to buy.
So, guys, Roger wouldn’t have been in business for 20 years if he didn’t know his market.
His market is mom-and-pop shops that do film transfers for customers who are happy simply knowing what’s on the film. Their standards are not your standards. Your standards are higher. That doesn’t mean the customers of these mom-and-pop shops aren’t happy. Clearly they are, or else the mom-and-pop shops would not keep buying Roger’s machines.
I begged Roger to step up to a 4K/10-bit camera, but he is sticking with his 2K solution. So I had to move on. But I still think Roger’s Retroscan machines serve a valuable purpose. He knows his market, and you are not his market.
Indeed that’s a massive problem with those services, and not just them but even most film archives and universities with film scanners. They can have the latest and greatest gear, but if their operators don’t have much knowledge about film or aren’t properly trained to operate them they will get less than optimal results that are possible with their equipment.
I respectfully disagree. There may be some end-customers who are happy with that quality point, but there are others who are not. These are people who generally want to transfer their old home movies edit them in Resolve and burn them to DVD and/or Bluray to share with their families. Not people who simply want to know what’s in their deceased grandfather’s attic.
If you’re running a commercial business there is no valid reason to use equipment that is “consumer grade”, and especially when commercial equipment is so cheap as it is. The proper tools of the trade cost money, that’s true for any industry.
Yes that’s right, that’s basically a quote straight from Roger, and I believe him.
I don’t agree with this, I think if the owners of the machines think this, they need to get a new job. There’s a certain responsibility that comes with handling peoples photo and video memories. They aren’t just poorly shot boring video. Someone is entrusting you do do the best, expects you to know how to use the machine to the best ability, and create a great looking output. If you look at everyone’s website, they constantly rave on about “we use frame by frame, best quality la la”, but yet many don’t.
I do a lot of scanning for families, and 50% of my clients are very fussy on quality. I guess times are changing!
I think Roger made an excellent choice with the camera. 2K looks great for 8mm formats, and as he knows about his user base, they’d be well out of their depth with 4K files and the additional problems that come with it. But sure larger formats would like more resolution, but I have never scanned 4K.
If you’re smart enough, it’s an easy swap to change the camera on Rogers machine. The machine outputs a positive trigger on frame detection, just have to splice the new cable on the new cam you choose, but then you’re going to have to capture with spinview or whatever program comes with your camera. Hopefully the new version of the Moviestuff software allows easy integration of other FLIR machine cams.
I normally deliver clients a .mp4 1080p file, and a 2k 16:9 Apple ProRes 422.
I get old clients asking me originally to put their film on a DVD, but I explain why they should have it as a digital file, and how it will work.
The MP4 files are great for playing on all devices like their smart TV.
And sure, they have no idea what Apple ProRes is, and will never use it. But - you’re not just scanning the film for the old people that shot it, you’re scanning if for their children and grandchildren who could know a heap about video formats! and will be glad to find some high res files there. Doing a proper job of preservation incase they throw it out or loose it.
And my posts really have nothing to do with the Moviestuff machine, I think it’s a good machine with excellent frame detection and very stable output.
This is more so about people running a business scanning peoples 8mm home movies and outputting a very low quality output, regardless of what scanner they use. It just comes down to how much effort they want to put in, and if they are willing to convince and old person to get some digital files along with a DVD!
I currently use Rogers Universal Mark 1 and have ordered the Mark II. I use Premier Pro for touch ups and editing. I’d proudly compare my scanning product with those scanners with any $40,000 plus machine for the average consumer. They wouldn’t be able to tell the difference and they would pay less having my service verses those with the big buck machines charging clients big bucks. I’m very proud of Rogers machines. It’s not just the machine that makes the final product great, it’s the passion put into the final product. Just because I have $40,000 golf clubs does not mean I’m a good golfer.
The lowest price I know of for commercial scanning of 8mm film on a 2K CCD Scanstation in Europe is roughly $4/minute (18fps) or 30c/ft ex-tax going off the stated price on their website. It’s not a company I’m recommending hence why I’m not linking to them (those sensors have some issue that develops over time and requires you get them serviced by LG and a lot of companies don’t keep up with maintenance), but I do know someone who has used them in the past and any person off the street can take their film there at that price. Lowest price I know of in the US is also 30c/ft for 8mm 3K scans on a 6.5K Scanstation, and that is a professional company I’d have no problem recommending. So that’s what I man when I say that I don’t see much evidence of commercial scanning houses using Retroscans that charge substantially less than is possible to do on more professional equipment.
I feel the “average customer” has an information deficit in any professional industry. Speaking personally I don’t like to hear the line “the average customer doesn’t know …” as it seems every time I hear it to be used an excuse to deliver a lower quality product or service to what the customer expects and I can give examples from other industries. As an example I used to know a joiner who did good work, but he refused to throw out bad timber so any warped or rotten timber he would use anyway because “the customer doesn’t know”, he would sell a bookcase with rot all over it for the same price as one without any rotten timber in it. You can push the Retroscanners and improve the results a bit, but for commercial work (including for individuals) I don’t see how it can compete against the HDS+ or the other commercial scanners in 2021.
I’d like to clarify that I agree with Roger that there’s a place for cheap scanning machines to be used for the archival purposes of identifying/sorting film material. For that kind of purpose the Retroscan is fine. A lot of professionals use discovery scans to identify what they need to do to prepare the film for a proper transfer as well.
Funny how people brag about knowledge but not skills. Tend to be similar to my old law enforcement days. My high school educated self could out perform all of the law degree fellow enforcers. In fact several would consult me before moving on with an investigation. The proof is in the pudding. Again give me Rogers machinery, and I’ll make a digital film that to the naked eye aka typical clients tv look as good as any $40k machine. Sure if put on some type of computer program the stats won’t be equal. But I don’t care about graphs etc I care about clients happiness at the best price. I’ve play music on at $500 stereo and have convinced so call audio junkies they were listening to “bose” $5,000 stereo. LOL. There is more to to idea than a “name brand”
I don’t doubt you can do a nice scan. My criticisms are not primarily directed at the Retroscan users that are getting the most out of their equipment, it’s at those who are using them to produce sub-par commercial work for clients.
One of my friends has been putting a lot of work into improving his RUMkII scanner. He is currently testing prototype gates in it, I won’t share the full details here because he intends to patent his design and then produce it. The idea will be that existing customers with the RUMkII can buy his aftermarket gates, and possibly do some other modifications, to make those things a bit more usable. This scanner has its place in the market and some people like it how it is. If that’s you that’s fine. If this isn’t you please reach out and we’ll be happy to assist you - there are even some mods you can do yourself which are quite easy. If you make any hard modifications then Moviestuff will void your warranty and that will deter many users from making improvements. I do not intend to misrepresent Moviestuff in any way, so if anything I’ve written in this post is incorrect feel free to correct me and I’ll edit this post.
What I can share about this is that we couldn’t find any information on the light that Moviestuff supply, but it’s terrible. It’s too dim to do proper scanning with leading to a large exposure time and consequently motion-blur. If you’re a technical guy we recommend putting in a high CRI YUJILED, you will need to attach a heatsink and then diffuse the light. The cheapest and easiest diffusion method is opal diffusing glass, but the drawback is that less light goes through compared to a true integrating sphere. Putting such a terrible light in a machine that is so expensive is a rip-off.
The scanner comes with a Chameleon 3 CM3-U3-31S4C-CS which is not a bad camera for scanning but it is old and obsolete. Putting it into a brand new commercial scanning machine that costs north of $8K in 2021 is ridiculous. It’s a 2K camera, and all claims about it being “near 3K” are false and misleading, I will be asking Moviestuff to correct their marketing.
35mm scans by taking a photo on every sprocket detection and then the Moviestuff software decimates frames. The maximum speed of the camera is 15fps, so this is why the 35mm scanning speed is about 4fps. 35mm is therefore not properly supported by this scanner, a fact that is not made clear in the advertising. As far as I’m aware we don’t have a design to fix this problem - my idea would be to make an electronic circuit to do it, so if anyone here would like to give it a crack please do. Given the price of the scanner is $10K if you buy it for 35mm (as you’ll need the 2000ft extension), I think Moviestuff should be the ones to offer a free remedy for this undisclosed problem.
I’ll let the users themselves do a full overview of the pros and cons of this scanner, but I would summarise it this way: it’s better than anything cheaper, but it’s very expensive and not good value for the price. It’s the genuinely cheapest scanner that can do 8/16/35mm BUT you may be be better off with a single-format scanner if you intend to scan a particular format (the full tri-format price is almost $12K). Some of the claims made in the advertising are false and misleading and I feel are likely to deceive. If you feel you have been deceived you should reach out to Moviestuff and see if they will offer you a remedy, and go from there. I would recommend before buying one that you think carefully about your needs, and if it doesn’t suit your needs then get a product that does. Think also about whether you need to buy a scanner, you may be surprised at the commercial rates charged by competitive scanning companies these days.
It has been a good while since I visited and commented here. At the time, I was investigating using a Steenbeck 16mm flatbed as a transport for a machine vision camera triggered by interrupts created by a wide slotted disk. I set up optical solutions which worked into a 2/3" sensor camera (SI2K).
Rolling shutter was going to bring me undone unless I recovered the image through the original prism path to freeze the image and trigger the SI2K to be there for each best image through the prism. For that, the timing of the frame trigger was not going to be as critical as for freezing a moving film with fast shutter speed.
Early on I discovered that it was going to be a dead loss for scanning neg film. Variations of image brightness through the prism and internal reflection artifacts were massively amplified by the inversion of the neg and digital recovery of the contrast and colour densities.
I eventually bit the bullet and restored the Steenbeck to its normal function. I bought Roger’s Mark II scanner and have been pleased with the results. More dynamic range from the camera to deal with high-contrast reversal would be nice. Neg film comes through nicely. The camera seems to have enough dynamic range to deal with it. I need to practice the dark arts related to inversion and reproduction of the colours.
Whilst still in the learning phase, very recently I was asked to scan some 16mm film someone had got hold of the see what was on it. As chance would have it, the vision had been shot by someone who could afford a 16mm camera for “home movies”. It must have been a Bolex because the frame rates though stable were all over the place as settings were apparently changed often.
It was all shot around about 1968-1970. One piece had been torn off by a projector and stuffed back into the reel head out. It had weaved and folded over between the edge of the roll and the side of the reel and had some hideous creases. There were other breaks and tears which had to be respliced. to save frames I joined across tears. I forgot best practice from the days of projecting film of putting black ink across the tear joins which let a flash of light through set off loud pops from an optical sound reproducer. They seemed to set off a false trigger.
That damaged piece of course was the most historically significant but that is the hand the fates offer.
I am satisfied with what the machine does. In the real world, a 4:3 16mm film print is not the sharpest and does not meet HDTV (1920 x 1080) resolution. Super16mm only came up to HDTV resolution in real world terms. I understand that the Blackmagic Cintel scanner achieves framing of 16mm by sensor crop to around about 2K resolution. 16mm reduction prints of 35mm feature films for screening on ships and oil rigs were not sharp. The Mark II camera and its lens can recover available resolution adequately. With a higher resolution scanner you will get finer grain detail which some propose as being a desirable aesthetic which confers a crisper look through that fine artifact being evident.
For sake of interest here is a link to a recovery of some footage which was inverted, balanced for blue hue due to incorrect camera filter having been used but no effort at colour grading or correcting of overexposure. The weather conditions were hot with a hard sky, more grey than blue with dust haze.
It is not representative of the best that Roger’s scanner can yield.
From my experience with 8mm and Super 8, the film content certainly on the small format exceeds HD (1920x1080). More over, some are well worth 4K specially for having a higher resolution for postprocessing. I haven’t gotten to 16mm yet, but if 8/S8 handle more than HD, so should 16. There are plenty of examples of 16mm 4K scans on youtube, so I would certainly encourage considering a higher resolution/different camera for historically significant material material.
I would agree. And if these are home movies, they’re not prints, they’re camera originals. The reason you’re not seeing as much detail is probably more the scanner than the film.
We have scanned countless 8mm, Super 8, and 16mm home movies shot on kodachrome, ektachrome, and various black and white film stocks and there is a ton of detail there - way more than HD can resolve.
Well, I am not so sure. It probably depends on what you are looking for.
Let me explain. The image on the Super-8 film is tiny - as far as I know, the camera frame is specified with 5.690 x 4.220 mm, whereas the slightly smaller projector frame features 5.430 x 4.010 mm as dimensions.
Now, imagine just the perfect optical image cast by the lens of the camera onto such a frame.
The resolution of this optical image will be limited in two ways: for large apertures, various lens defects are ruining the image, for smaller apertures, the image is going to be diffraction-limited. The keyword here is “Airy disk”. Put simply, the best focused spot of light a perfect lens is going to cast onto the film has a diameter of approximately
d [ in micrometer] = f-stop
That means the image of something like a distant star (which would focus within classical optics onto a single point) will end up as a tiny disk (as the wave characteristics of light are taken into account).
Most lenses have their maximal sharpness around f/5.6 to f/8. Assuming such an f-stop for the moment, our image will be composed of tiny disks with a diameter of about 5 microns. That would be around 1140 x 840 disks placed side-by-side on the full camera’s frame dimension, which is slightly less than full HD-resolution.
Remember, that is assuming a perfect lens; certainly most of the zoom-lenses used at the time Super-8 was en vogue were not that perfect.
Next, consider the resolution of the Super-8 material by itself. I have found different data for Kodachrome, with numbers ranging from 53 lines/mm to 100 lines/mm. Taking the best value into account, we end up with only 569 x 422 pixel per camera frame…
So, in case you are interested in the film’s content (which is, the image as seen by the lens), I think, for the reasons elaborated above, that the Super-8 format does not exceed HD resolution or even reach HD resolution in most cases. Certainly not the old Super-8 material I have available for scanning.
However, if you want to capture the intrinsic properties of film grain, you might want to go even beyond 4k, as film grain has very fine, detailed textures.
Film grain, however, is not a structure related to what is filmed, but rather a structure related to the material and chemical processes used to develop it.
While film grain was primarily viewed as a nuisance in the days of Super-8, today it has turned into an aesthetic statement, often even added in post to grain-less digital material.
And of course, higher scan resolutions give you a lot more latitude for post-processing, which is an important advantage. So I do also scan at resolutions that exceed HD.
Well aware of the dimensions, see this post.
SMPTE Recommended Practice Specification for Registration Test Film (for 8, and 16) specifies
“The rosette in the center shall indicate measurement from 60 to 240 lines per millimiter”
See RP19-1982 for 8mm, RP 20-1982 for 16 mm.
For clarity, these lines are not the dimensional unit lines.
@cpixip one cannot resolve/render 100 lines with 100 pixels.
Here is an actual DIY scan I made, suffering from youtube compression, and one can see the difference just by switching the youtube resolution to HD, look at the power lines and the increase details of the monument writings.
I know that 4K for 8mm is ongoing debate at Kinograph, count me on the side that more-is-more, And for 16mm, HD resolution is less than what the film can render…
PS. Manufacturers typically expressed resolution in cycles/mm or lp/mm = line pair per mm.
Found this very detailed overview on the subject of film-lens resolution, will give it a detailed read at later time.
The MTF of Kodachrome at 30% is listed about 50 cycles/mm 2009 datasheet.
One important consideration in determining the cutoff of pixel resolution is that even for a lens-film low percentage response, these would still represent information and details in the picture (like the power lines or small-writing on the sculpture).
The main problem with the stock RUMkII is the light, not the camera. Building a better light is cheap and relatively easy, we should have a full write-up on how to improve that machine soon. There’s a few of us involved behind-the-scenes there. Changing the camera will also help, but changing the light is more important and less expensive.
Most 16mm prints are soft and low-resolution, but they can still scan a LOT better than off a stock RUMkII. If you get your hands on a TV print or a GOOD 16mm print then it’s a different story, they can be sharp and detailed - it just depends on how they were printed etc.