ProRes from DPX... sharper than DPX?

[Steve Zimmerman]

At first, our scanning facility mistakenly sent us Prores HQ 2K scans from our 35mm 2 perf film, when we asked for DPX. We sent the film back, and they sent us the DPX files. Comparing the two images, the compressed ProRes looks visibly better. There is stairsteping on some high contrast edges on the DPX, but rarely on the Prores, and we can see pores on ProRes faces where the DPX is smooth. Both files are from same lab, why wouldn't the uncompressed DPX look visibly superior? Problem with our prores files are, the have a basic baked-in correction.

I originally wanted to go DPX because there is more latitude to play with in color to grade in case of blown out highlights. Is it possible to ask the lab to squeeze more highlight detail into the ProRes? I guess I'm fine working with ProRes instead of DPX's as long as we get the DPX too for safe keeping, and future use.

 

 

Our output format is Blu Ray disc

 

Thanks in advance,

Steve Zimmerman

Charleston, SC USA

[Phil Rhodes]

My guess, they got cheesed off with having to redo it, and did a quick and dirty job on the DPX.

 

Really your complaint should be with them. Work DPX, not ProRes. Compression is for girls.

 

P

[Steve Zimmerman]

Work DPX, not ProRes. Compression is for girls.

P

 

Thanks for the support, Phil!

 

Would you recommend I use Gluetools? Do I start working with the files in Final Cut (with GlueTools) or Color?

 

I also just heard about Clipfinder. I guess it works for any DPX file, RED or film, and it's cheaper!? Anybody use this for film DPX's with FC and Color?

 

I also figured out they neglected to zoom into the full academy width of the 2-perf image, There was a black bar on the left side (the soundtrack area). That's why the resolution was lower, and we had to zoom into the image in the Color geometry room with the 2K 2.35 wireframe to fill the frame. We had them scan some full height 4-perf footage, so I guess it seems they used the same setting for both(?).

 

Thanks,

Steve Z.

[Thomas Worth]

At first, our scanning facility mistakenly sent us Prores HQ 2K scans from our 35mm 2 perf film, when we asked for DPX.

We need some more information. What type of machine was used to do the scans? Not sending DPX files initially sounds very suspicious. This sounds like a telecine job to HDCAM / SR which was then captured on a Mac with a Kona card to ProRes.

 

We sent the film back, and they sent us the DPX files. Comparing the two images, the compressed ProRes looks visibly better. There is stairsteping on some high contrast edges on the DPX, but rarely on the Prores, and we can see pores on ProRes faces where the DPX is smooth.

It could be that the DPX files were created from the ProRes material or from the HD tape using a different system. Without knowing exactly what type of equipment was used, we won't know for sure and won't be able to predict how much of a pain it would have been for the facility to redo the work.

 

Both files are from same lab, why wouldn't the uncompressed DPX look visibly superior?

They should. If they don't, someone made a mistake.

 

Problem with our prores files are, the have a basic baked-in correction.

Why? Was a grade part of the deal? Was the deal for DPX original scans plus a color-corrected ProRes output?

 

I originally wanted to go DPX because there is more latitude to play with in color to grade in case of blown out highlights. Is it possible to ask the lab to squeeze more highlight detail into the ProRes? I guess I'm fine working with ProRes instead of DPX's as long as we get the DPX too for safe keeping, and future use.

As Phil pointed out, anyone who knows what they're doing will work from uncompressed files. What if you decided later on to grade on a system that didn't support ProRes?

[Steve Zimmerman]

We need some more information. What type of machine was used to do the scans? Not sending DPX files initially sounds very suspicious. This sounds like a telecine job to HDCAM / SR which was then captured on a Mac with a Kona card to ProRes.

 

It was an Arriscan

 

Why? Was a grade part of the deal? Was the deal for DPX original scans plus a color-corrected ProRes output?

 

As Phil pointed out, anyone who knows what they're doing will work from uncompressed files. What if you decided later on to grade on a system that didn't support ProRes?

 

 

No, it was a free test scan. I just thought they had done some sort of best light on our footage. The highlights

 

Very true. We are slowly learning the whole Final Cut to Color workflow business. I guess we need to use Gluetools even though we are doing an offline/online. I also just heard something called Clipfinder can be helpful too? Any help would be greatly appecited. B)

 

Steve Zimmerman

Edited April 3, 2011 by Steve Zimmerman

[Paul Korver]

Hi Steve,

We scan a lot of film at Cinelicious... here's a couple of comments.

 

At first, our scanning facility mistakenly sent us Prores HQ 2K scans from our 35mm 2 perf film, when we asked for DPX. We sent the film back, and they sent us the DPX files. Comparing the two images, the compressed ProRes looks visibly better. There is stairsteping on some high contrast edges on the DPX, but rarely on the Prores, and we can see pores on ProRes faces where the DPX is smooth. Both files are from same lab, why wouldn't the uncompressed DPX look visibly superior? Problem with our prores files are, the have a basic baked-in correction.

I originally wanted to go DPX because there is more latitude to play with in color to grade in case of blown out highlights. Is it possible to ask the lab to squeeze more highlight detail into the ProRes? I guess I'm fine working with ProRes instead of DPX's as long as we get the DPX too for safe keeping, and future use.

 

ProRes 4444 is a great codec. We think of it as file-based HDCAM-SR because it's signal to noise ration (PSNR) is equal to HDCAM-SR 444... but the great thing is that it's resolution independent unlike SR which has resolution limitations. DPX is best of course if you don't mind the 8x data size and data rate. However, I can say with certainty that when comparing DPX to ProRes 4444, you should not be able to tell which image was DPX and which was ProRes 4444 without some serious sub-pixel sampling tools. You'd never be able to see it with the naked eye.

 

If your ProRes was lacking highlight info that is present in the DPX then a choice was made when the "look" was baked into the ProRes... likely a LOG-to-LIN LUT that was not carefully supervised... that or it was an artistic choice of the colorist. But it's not because of the ProRes codec. If you're lacking info in the highlights from an ARRI scanner then it could be a bad scan or if it was a very high-dynamic range scene it's possible that the ARRI lacked the dynamic range to be able to scan all the data. An ARRI scans at about 2.3 film density when there a possible 3.2-ish film density available on kodak vision 3 negative. What stock was it?

 

As Phil pointed out, anyone who knows what they're doing will work from uncompressed files. What if you decided later on to grade on a system that didn't support ProRes?

 

Sort of true... however there are a lot of big budget commercials that are using the Alexa and opting for ProRes 444 LOG C over RAW/Uncompressed. We've graded a lot of this footage and you can tweak the crap out of it. A properly scanned ProRes 444 master from film has a similar characteristic. Again if you have the time, money, disc-space and data throughput to deal with uncompressed the by all means do it. But in certain instances ProRes rocks. For instance... 4K ProRes 4444 has a similar data rate and file size as 2K DPX.

 

 

I also figured out they neglected to zoom into the full academy width of the 2-perf image, There was a black bar on the left side (the soundtrack area). That's why the resolution was lower, and we had to zoom into the image in the Color geometry room with the 2K 2.35 wireframe to fill the frame. We had them scan some full height 4-perf footage, so I guess it seems they used the same setting for both(?).

 

This is sort of confusing and lame but technically speaking a "2K" scan of 35mm film is 2048 across the entire neg (perf-to-perf). Meaning the only way to do a 2K scan from 35mm and get 2048 across is to shoot 2-perf, 3-perf or 4-perf Super 35mm (all formats that expose perf-to-perf). This is due to the way the scanner sensors are built (they don't "zoom"). So the actual pixel resolution of your scan 2K scan depends on what portion of that negative area it's taking up. If you shot 4-perf anamorphic then you're really dealing with an Academy aperture which leaves space on the left of picture for sound track. The resulting resolution of a 2K Academy scan is 1828 not 2048. Sucks I know... especially since you're going have to stretch/uprez those pixels out 2048 for a 2K DCP or 1920 for and HD Blu-Ray release. For that reason we always recommend we scan all our clients 4-perf anamorphic material at 4K which would yield a stunning 3656 pixels of resolution. Or... if you want a 2048 2K scan of your Academy area then you need to ask for a 4K scan and any good scanner can output an on-the-fly over sampled 2048 pixel resolution across the Academy picture area.

 

Hope that makes sense.

 

-Paul

[Phil Rhodes]

This is sort of confusing and lame but technically speaking a "2K" scan of 35mm film is 2048 across the entire neg (perf-to-perf)

Euh. Well. It's a complex area, but there are a few caveats with this, as I'm sure you're aware.

I'd expect them to provide cropped 1800-wide frames or whatever if the picture didn't fill 2K.

P

[Chris Burke]

 

This is sort of confusing and lame but technically speaking a "2K" scan of 35mm film is 2048 across the entire neg (perf-to-perf). Meaning the only way to do a 2K scan from 35mm and get 2048 across is to shoot 2-perf, 3-perf or 4-perf Super 35mm (all formats that expose perf-to-perf). This is due to the way the scanner sensors are built (they don't "zoom"). So the actual pixel resolution of your scan 2K scan depends on what portion of that negative area it's taking up. If you shot 4-perf anamorphic then you're really dealing with an Academy aperture which leaves space on the left of picture for sound track. The resulting resolution of a 2K Academy scan is 1828 not 2048. Sucks I know... especially since you're going have to stretch/uprez those pixels out 2048 for a 2K DCP or 1920 for and HD Blu-Ray release. For that reason we always recommend we scan all our clients 4-perf anamorphic material at 4K which would yield a stunning 3656 pixels of resolution. Or... if you want a 2048 2K scan of your Academy area then you need to ask for a 4K scan and any good scanner can output an on-the-fly over sampled 2048 pixel resolution across the Academy picture area.

 

Hope that makes sense.

 

-Paul

 

 

paul,

are there any scanners that do zoom in or should someone oversample if you want a full 2k width?

[Paul Korver]

paul,

are there any scanners that do zoom in or should someone oversample if you want a full 2k width?

 

Hi Chris,

To be clear your question is only referring to Acadamy / 4-Perf Anamorphic right? (since all the other formats do expose perf to perf and would yield a 2048 scan). In that case of Academy picture area there are no scanners I'm aware of that "zoom" to focus the 2K sensor area on the smaller Academy region. That's why technically speaking a 2K Academy/Anamorphic scan has always been 1800 & change pixels. If you are getting scans done at a post facility and you want 2048 across an Academy region I would first be sure ask if they can scan at higher than 3K resolution. If they say yes than they should be capable of outputting 2048 across the academy region by way of oversampling. If its purely a 2K scanner and they say they can deliver 2048 academy then there's likely some uprezzing going on.

 

Paul

[Steve Zimmerman]

Thanks for all the great information and tips, guys.

 

They scan 16mm at 2k, don't they? They can't be using one setting and scanning that stuff at one quarter resolution? I am surprised there is no middle setting for 2K scanning 35mm 2-perf. Is this limitation the same for Arriscan and Spirit 2k?

 

Thanks,

Steve

[Paul Korver]

Thanks for all the great information and tips, guys.

 

They scan 16mm at 2k, don't they? They can't be using one setting and scanning that stuff at one quarter resolution? I am surprised there is no middle setting for 2K scanning 35mm 2-perf. Is this limitation the same for Arriscan and Spirit 2k?

 

Thanks,

Steve

 

Steve - they do have different "settings"... they're called "gates" which is a physical gate that positions the film so that the sensor scans from perf-to-perf for a given stock (in most cases either 35mm or 16mm). They typically don't make a specific gate for Academy 35mm which is why there is a decrease in resolution for that format only. This is all very confusing if without a very clear picture in your mind of the exact dimensions the negative area used for each format... don't mean to be patronizing if you already know this stuff but here's a helpful chart for 35mm: http://en.wikipedia.org/wiki/Super_35.

[Steve Zimmerman]

I understand how Super 35 is wider than academy. It's too bad there is no setting for academy width to get better scanning resolution.

 

Our scans look fine, but it's too bad we would have to get a really expensive 4k scan to actually get full 2k pixel width, when a smaller format like S16mm can.

[Phil Rhodes]

I would be astonished if you found a scanner that couldn't produce 1920-wide scans from Academy.

 

 

P

[Will Montgomery]

I can say with certainty that when comparing DPX to ProRes 4444, you should not be able to tell which image was DPX and which was ProRes 4444 without some serious sub-pixel sampling tools. You'd never be able to see it with the naked eye.

This has been my experience on several different tests. ProRes 4444 is an amazing codec.

[Dirk DeJonghe]

paul,

are there any scanners that do zoom in or should someone oversample if you want a full 2k width?

 

 

Sorry to be a bit late on this thread. Yes, there are scanners that allow optical zooming in on an Academy frame to give full 2048 or whatever resolution you require. I use one since 16 years and it is called an Oxberry Cinescan.