Dynamic Range - Scanity, Director and ScanStation

[John Paul Palescandolo]

A lot of you have probably seen some of my recent posts on here. I'm working on having my 16mm and 35mm film from my student days scanned, now that the technology is available. I had some test scans of my 35mm film performed at two labs in NYC - Nice Shoes, which uses a Scanity and Metropolis Post, which uses a Lasergraphics Director. In short, both test scans looked really, really nice, although while I had hoped to compare two HD scans, the test scan from Metropolis Post looked nicer. Why? Well, their scan was done at 2K, but also with the ungraded file, a graded file was delivered.

 

Alright, so I realize that both scanners are excellent and I can't go wrong either way, but when a lot of money is being spent to scan film, I like to know some more of the nitty gritty tech specifics. From a previous post about scanning 16mm at 4K, someone suggested I do a search for posts where Paul from Cinelicious makes a great case for scanning 16mm at 4K. While I was unable to find that specific post, I did find other info where Paul mentioned that in his tests, the Scanity was found to have a dynamic range of 3.5, while Kodak Vision 3 stocks generally have a dynamic range of 3.1 - 3.3.

 

So, that got me thinking - what is the dynamic range of the Lasergraphics Director, especially since it has the triple-flash HDR? Curious to see if Paul had tested the Director while deciding which scanner to purchase, I emailed him. In his response, he said that he did not test the Director at the time, but their new triple flash HDR has really stepped things while it is impressive, he has not measured the dynamic range of the scanner.

 

Still curious, I reached out to Perry at Gamma Ray Digital up in Boston, whose facility recently acquired a Lasergraphics ScanStation. He was kind enough to reach out to Lasergraphics and they responded that the dynamic range of the ScanStation is 3.6 and the dynamic range of the Director, with triple flash HDR turned on, is 5.6. Lasergraphics also said that:

[...] it is important to establish that the actual dynamic range of Kodak Vision 3 stocks is actually about 2.3 under the most extreme over-exposure conditions (i.e. highly overexposed negative film). For technical details refer to:

1. The Kodak Vision 3 specification sheet:

http://motion.kodak.com/motion/uploadedFiles/TI2647.pdf

Refer to page 4 Sensitometric Curves. Remember to subtract out the "base" (i.e. the minimum density, DMin) before measuring the total density range. The total range is DMax minus DMin for each color separately.

2. Kodak cooperated with Lasergraphics and other vendors to establish extended density range scanning specifications. Refer to the resulting Kodak white paper:

http://motion.kodak.com/motion/uploadedFiles/ExtendedDensityRangeScanning_022309.pdf

This paper establishes that density range (DMax - DMin) of 1.85 is the DPX standard and is sufficient for most scanning (page 2). Under extreme conditions, the range can be expanded to 2.37 (page 3, paragraph 5)

Basically, what I take away from this is Scanity or Lasergraphics Director, they're both great and the results you get out of it will not only depend on the quality of the film being scanned, but also the person performing the scanning and doing the grading, etc. Now, the question, at least for my 16mm film, is resolution vs dynamic range. I'd say it's better to have more dynamic range than resolution, but is my film so grossly mis-exposed that I need the dynamic range of the Director? Probably not, but who knows.

Anyways, I just wanted to put this information out there, for all to have.

[Perry Paolantonio]

 

the dynamic range of the Director, with triple flash HDR turned on, is 5.6.

 

Actually, it's 5.1, according to Lasergraphics.

 

-perry

[David Cunningham]

And that's largely aimed at print or reversal films which by their very nature have greater range.

[John Paul Palescandolo]

The triple flash HDR is aimed at print and reversal films? So, is the Director more apt to handle print and reversal films than a Scanity? According to Lasergraphics, the ScanStation's dynamic range of 3.6 does exceed the Scanity's of 3.5, which is the dynamic range that Paul reported measuring. Maybe what we need here is a test done between all three machines - Scanity, Director and ScanStation.

[David Cunningham]

I have tested the scan station and director with vision3 50d. My result indicated a superior range and picture "quality" with the director. But, the difference is negligible unless you are talking about highly over or highly under exposed films. Also, despite actual experience yet, my understanding is the the director is by far the best in the market form print and/or reversal films.

 

I haven't been able to rest a scanty because I've just been unable to afford it.

 

I sent the same 16mm film around to many places including the director at metro post, the xena at cinelab and a few others on the lower end. By far, I was most pleased with my metro post director results.

[John Paul Palescandolo]

David, when you sent your film around to all those places, were you paying for the scans? I'm sure you could arrange a free test of the Scanity.

 

I liked my Director scans from Metro Post better than my Scanity scans from Nice Shoes, but it wasn't exactly a fair comparison. While I had asked each lab just for a one minute, HD test at 23.98 fps, there were seemingly two misunderstandings, tho neither truly negative (no pun intended).

 

Nice Shoes did deliver an ungraded, one minute, ProRes 422 HD test at 23.98 fps. The 4-perf Super 35mm aperture was also cropped to 1.78:1, which I had wanted, but instead of transferring only one minute, they transferred all 1200' feet I'd given them. I can tell you that the Scanity has a lot of resolution. So much, in fact, that I was getting moire patterns over certain patterns when I played the footage back on my MacBook Pro. However, once I zoomed into 100% in Final Cut, the moire disappeared.

 

Metro Post did deliver a minute test, but instead of just scanning to HD, they delivered a 2K ProRes 4444 ungraded scan and a 2K ProRes 422 HQ graded scan, both at 24 fps. Perhaps it was the 2K scan, or the grading, or both, but what was supposed to be a direct comparison isn't really that anymore.

 

Below, I'll post a comparison between the Nice Shoes and Metro Post scans. The Nice Shoes scan comes first and since it was delivered ungraded, I graded it myself in Final Cut. To make the Metro Post scan, which comes second, match the Nice Shoes scan more closely, I used Compressor to crop it to 1.78:1, I converted it from 24 fps to 23.98 fps and also from ProRes 422 HQ to ProRes 422. The grading that Metro Post did definitely looks more natural than the grading I did to the Nice Shoes footage using Final Cut.

 

You have to watch it on YouTube to view it in 1080p.

 

[Perry Paolantonio]

David - Bring that 16mm film by next time you bring something in! I'd like to add the ScanStation to the mix on that test, and we'd be happy to run a free test scan of it for you.

[David Cunningham]

Hi Perry,

 

A free test with my test roll? Twist my arm. :) I'll probably just have you splice it with the 200Ft of new stuff I have and run it through all together in one scan.

 

John,

 

Unfortunately I have not had a very easy time finding free scan tests. So yes, I paid for all my past tests with the Director, Xena, etc. However, I usually at least get reasonable discounts. Some places did offer to scan 1 minute of film, but that just wasn't worth it to me. I wanted to see the entire 100Ft because I chose this roll for all the factors it had in it. Portions were over exposed, portions were under exposed and portions were perfectly exposed. So, I was able to see how well the scanners dealt with that. Thus far, the director did best, hands down. Now, the operator and grader may have a lot to do with it too. So, that may be another bonus for Meto Post who did my Director tests.

 

I haven't found anyone with a Scanity that would do a test scan and those who have a Scanity have pretty high prices and minimums. So, I just haven't gotten around to adding the Scanity to the test mix.

 

My "test film" for 16mm was the 16mm portions of this film:

 

 

This footage is the Director version. You can see the overexposed and under exposed areas. Both of them are extremely over and under. My one-light print of this film resulted in an almost completely blown out and black bad exposure scenes.

[David Cunningham]

John,

 

Those examples are really nice. You used some nice lenses, eh? Unfortunately I only have older C-Mount lenses for my 16mm work. The results are a bit soft but I do like the retro lens look they give. The color and contrast is nice and please. They wouldn't be good for "feature" work, but they are fun to use.

 

I wish your test scenes were the same ones. The lighting and skin tones are a bit different between the two scenes. But, that's all right. You get the idea. Clearly, both these machines yield excellent results. The only thing that jumped out at me was the color difference that was likely your grade vs their grade. But, the quality of the image (dynamic range and grain structure) looked pretty consistent to me. Unfortunately, Internet compression makes it pretty much impossible to be sure about the grain. But, i think you can get a pretty good idea.

 

I can't wait to see the full feature. :)

 

Dave

[David Cunningham]

John, one last comment that relates back to your grading of the Scanity footage.

 

If they delivered ungraded HD Prores 422, then grading after-the-fact is hard. You are missing a lot of the light information in Prores422 HD. You really need the Prores4444 to be able to properly grade, especially for shadows and highlights. So, you were up against a challenge from the get-go.

 

Dave

[John Paul Palescandolo]

David,

 

Thank you for your compliments! I remember that we shot using a Moviecam Compact, tho I don't remember the brand of lenses we used, nor the focal lengths. My DP just selected everything and I wrote the check :rolleyes: Back then, I wasn't into cinematography or still photography, thus my role in just writing checks. The camera reports say we shot on 5248, which was EXR 100T.

 

I do wish my scans were the same scene, but the train to NYC is $22 round trip, so I was trying to minimize the trips back and forth, which may have worked against me as well. The Nice Shoes test scan came back really quickly tho - the following afternoon! But, I'd already given Metro Post different footage to scan. No biggie tho because as we can see, both scanners are great.

 

I do have several rolls of 16mm black & white and color reversal film. It's been a while since I've looked at them, but I'm not sure so much if the exposure varies as much as reversal film has less latitude. Also, back in the day, I would transfer my 16mm reversal film to mini dv tape myself. We had a projector at school which had a CCD or some electronic device in it that plugged into the switch board in the control room. There was a dial on the projector to adjust the exposure, so I'd have to stop the tape when an exposure adjustment was needed.

 

Below, I'll post a clip from the original SD transfer vs. the Scanity transfer. You'll be able to see that when I graded the Scanity transfer, I was trying to match what the original colorists did. The original transfer was done at Moving Images in NYC and was done in SD to mini dv in 2004. To try and make the clips match a bit more closely, I unconverted the SD clip to 1920x1080. I also converted the frame rate of the Scanity transfer from 23.98 fps to 29.97 fps.

 

[John Paul Palescandolo]

The SD transfer comes first and the Scanity comes second.

[John Paul Palescandolo]

David, one other question - is there a big difference between ProRes 4444 and ProRes 422 HQ? I once ripped the big knife fight scene from the Blu-ray of V for Vendetta three times and did one encode as ProRes 422, a second as 422 HQ and the third as 4444. I didn't see any difference between 4444 and 422 HQ, but both were noticeably better than 422.

[David Cunningham]

That transfer comparison is an excellent example. Look at the increased detail in his face (the bright face) and in the dark shadowed section of the blue coat! Perfect examples of superior transfer.

 

You won't really see much difference in overall quality between ProRes 422 and 4444 if you do not mess with the grading. It's when you try to mess with the grading because that missing data doesn't allow you to adjust, especially if trying to bring out highlights or shadows. Generally speaking, with 422, if you have blown out highlights or crushed shadows, there isn't much you can do. If you get a 4444 file you should be able to adjust those levels in grading.

 

This guy has excellent tutorials on 422 vs 444 and color in video as a whole.

 

http://blogs.adobe.com/VideoRoad/2010/06/color_subsampling_or_what_is_4.html

[Michael T Gardner]

I recently had some film scanned by Metropolis and I've noticed their website says they're scanning with an Oxberry 3k Cinescan.

 

I'm wondering what machine my film was scanned on. Is the information published on the site out of date?

[David Cunningham]

I recently had some film scanned by Metropolis and I've noticed their website says they're scanning with an Oxberry 3k Cinescan.

 

I'm wondering what machine my film was scanned on. Is the information published on the site out of date?

 

Yes. You are looking at their old site. I'm not sure why it was never taken down. I have been meaning to ask jack why it was never updated or replaced.

 

He does not use the Oxberry anymore. He only uses the Director these days.

[Michael T Gardner]

Thanks for clearing this up for me! I'm very happy with the work they've done. It was far better than what I expected from an old NPR.

[John Paul Palescandolo]

I'm wondering if anyone knows if the Director can output 4K Quicktime files. I know it can scan 4K DPX files, but Jack told me he isn't sure if it can output 4K Quicktime files.

 

Ok, so suppose it can't output 4K Quicktime files and it outputs 2K Quicktime files. Will the film be scanned in at 4K and then be converted into 2K Quicktimes?