Robert Houllahan

19 hours ago, Larry Baum said:

Also I should clarify, I don't think it is that all the output samples are in REC709 color SPACE, they just all seem to be in REC709 color GAMUT (solely with regard to the xy locations of the color primaries on an CIE xyY gamut plot).

Again did you have a LAD (Lab Aim Density) or TAF (Telecine Alignment Film) or some other SMPTE control film scanned to reach this conclusion?

5 hours ago, Daniel D. Teoli Jr. said:

 

I've told you and Perry time and again. All scanning options should be tested with comparison samples posted online. 

I have no recollection of being told this.

Nice in theory but really a moot venture as high end clients already know what they want and can afford to do finishing on our Arriscan for example.

Too many variables in any scanner "shootout" and allot of the business of owning these scanners is knowing what is the right machine for the job and client budget.

Also what files are you looking at? Were they. LAD or TAF alignment film you had scanned?

I would just say that high end “finishing” scans from a Scannity / Arriscan / Director / Xena are actual RGB scans either with a Tri- Linear RGB sensor or a Monochrome Area sensor and pulsed RGB LED lamp. 

Scanners with Bayer mask sensors are fast and can make very good scans but as far as I know they are “dailes” or “archival” use machines for more critical film and tv work.

There is some tradeoff with a color sensor, the dyes used in the silicon mask are not perfect and there is crosstalk between color channels. Any scanner maker has to do some math to figure out how to deal with this.

Color sensors are capable of wide gamut and it may just be a choice engineers made building the Scan Station and a possible feature request to LG as how this in interpolated is just how the software works.

Have you asked LaserGraphics? 

I am not able to scan to DNG direct with our Scan Station but maybe that is in a later software release.

I generally think the 2-flash is 14bit precision as the Sony Pregius 6.5K sensor is 12bit.

I do not know exactly what Lasergraphics is doing under the hood on the Scan Station but on the Xena scanner there is a Bayer mask transformation matrix to overcome some of the (quite large amount) of color channel cross talk. It is possible that the Scan Station uses a similar matrix or a 3D Lut to manage this aspect of using a Bayer mask sensor and the color dyes Sony used on the silicon.

I could do a comparative test on our Arriscan in 2-Flash which is 16 bit and true RGB.

What Scan Station? Was it in 2-flash? 5K machine or 6.5K machine?

I do find the scans from our Scan Station to be allot crunchier than the Xena scans or the Arriscan.

The Scan Station is very fast.

10 hours ago, Andrew Wise said:

Why would you want a longer optical path?

And longer focal length? 

I’m really not sure it would improve the design.

If you look into the lens specs and talk to the lens manufacturers you will find that there are optimal magnification ranges for all of these lenses. Imaging what is basically a transparency with the lens the led lamp and the size of the sensor is a bit of a trick with multiple film gauges. If you have a lens which images the film to the sensor and is optimally focused you can still not get the best detail and sharpest scan image if the lens is not best for that magnification range.

On 7/11/2022 at 11:15 PM, Tyler Purcell said:

100% agreed. I think I echoed that as well few days ago. The lens "works", it's just inexpensive. Sadly, the optical path is not long enough to use a decent focal length lens. 

What focal length is that lens? 50mm ? 80mm ? F4 Makro-Iris

Depending on what it is the lens may just be out of it's optimal magnification range.

20 hours ago, Dom Jaeger said:

Hi Rob,

Most of the time an overscan would pick up camera unstability, but there is some room for error because 16mm cameras register the film laterally by the edge of the film rather than the perf. Only vertical registration is controlled by the perfs. In theory the edge of the film and the perf edges should be locked together but of course this depends on the perforation and film width QC and not all cameras use the film edge directly next to the exposed frame perf (so if there is any weave in the film itself it may show up as camera unsteadiness).

So I tend to recommend a double exposure steady test because it's the ultimate way to rule out any other factors, and you can test different speeds and each mag as well (important for 16mm cams that have the pressure plate in the mag). 

But I'm approaching this as a camera tech not a scanning expert, so perhaps I'm being overly cautious? I'm curious to hear your thoughts on using the edge of the perf for lateral registration when scanning 16mm. And how important is camera steadiness nowadays anyway? Are there downsides to post stabilisation? 

All true, when we have had Pro jobs in that have been doing double exposure registration tests we do an overscan as well as that combination gives the most detail and feedback about the camera registration.

Just did one for an Arricam LT today which was double exposure and a 3K overscan 4-perf camera.

I would get a overscan done so you can see the camera gate relative to the perforations.

If the perfs are not moving but the gate has this movement then it is in the camera or loading.

We do a 2.5K overscan on our Scan Station for camera tests it really helps to see the perfs.

We do it but it becomes pretty blocky and contrasty.

I would suggest a Push +2 on 7222/5222 XX Negative instead.

With the right developer that can look very nice.

Crazy stuff

We have not seen any higher percentage of jammed S8 cartridges for 100D than any other S8 cartridges.

We did a large order (80-90?) S8 Ekta cartridges for a NetFlix show earlier this year and they all went through the camera(s) ok.

It is likely that the Spirit operator set the lamp output higher for those scenes with bright hilites (i.e. very dense) and as jack said you get some bleed into the scan from the sprocket area with that higher lamp setting.

Not bad for 20 Euro but no Arriscan either.

12 hours ago, Tyler Purcell said:

Yea, I've got a quote to get the reverse engineering and prototyping done for $50k, but who is going to pay $4k a board when the camera isn't even worth that much working?

There are allot of off the shelf motion control systems now and a XTR is worth $25-30K and 416's are worth 3-5x that.

Well there it is....

I really loved my LTR54 and it was a great camera for me, it was a very late one and even had Aatoncode which the Abel tech said he had not seen on a LTR before.

I am glad I traded it and got an XTR Prod...

I would think there must be some way to replace the older Aaton motor and electronics with something modern.

The LTR is a simpler camera compared to the XTR and when I had an LTR it was very reliable.

I think quite a few mechanical parts are the same but the electronics and motor are different.

No set AIM for a partial bypass it is a test to suit process and thus kind of labor intensive.

We ran a series of tests for the client and they settled on the one they wanted, it was not exactly linear in terms of number of racks bypassed as the ECN2 process has "extra" bleach in the process design most likely for consistency in process as the bleach gets more exhausted.

We have done this in the past, ECN2 has allot of bleach and if you run some tests you can bypass enough racks to get partial retention.

It takes most of the bleach racks out if I remember correctly from when Bob did it so the film runs through about 1/5 of the normal bleach time.

I think it always looks like it is too smooth or the contrast is not right compared to real B&W stock.

Maybe if you really worked some of the curves along with the de-saturation you could get it closer.

When I first saw the HDS+ I thought that there really wasn’t enough space for the lens - gate - camera. 
 

Here is the lens and camera on the 4K xena I run small gauge on:

Yeah one of the less looked at or understood specs on these lenses is the magnification range that they work in.

I assume the lens on the SSP was chosen because it is fixed and not a moveable.

The Sony Pregius 4K IMX-253 1.1" sensor is pretty small with the same 3.45u photo-sites as the bigger 6.5K IMX-342 but a smaller piece of silicon to image. The 80MM Makkro-Symmar is great for the 8mm magnification to the 1.1" Pregius but definitely falls far short on the 6.5K camera for 16mm and 35mm.

It is pretty remarkable that the Scan Station can image all three gauges well considering the distance of the camera to the film plane for 8mm, seems sub optimal distance and magnification.

Here is what is used in the scanners we have:

4K Xena with the 4112x3008 Sony Pregius (same sensor as the HDS+) Schneider 80MM F4 Makkro-Symmar has excellent results for 8mm with this setup.

6.5K Xena is a Printing Nikkor 95MM F2.8 Incredible lens tack sharp and extremely flat field. There is a Schneider which is comparable that runs in the $4K range 85mm Macro-Varon.

Arriscan uses a Printing Nikkor 95mm F2.8

Scan Station P uses a Schneider 80MM F4 APO-Rodagon-N (Or APO Digitar) 

The "rule" is one stop per decade generally.

As Perry said a positive film is a Linear scan and we use basically the same sensor on one of our Xena scanners to do small gauge scanning it is 4112x3008 12-bit and I can almost always scan color and B&W reversal stocks with a full grade-able range to 10bit linear DPX so it is a good sensor for this work.

I am not sure what control the HDS+ has over RGB balance etc. so it is possible that the scanner show did some corrections in Resolve to fix color balance etc they could not get done in the scanner.

If you look at the scans in Resolve etc. and open the scopes are the RGB levels balanced? Is there any clipping in the shadows or hilites?