DPX files

[Jeff Barklage]

I cannot find my answer on the RedUser's search page...
Does anyone know what program opens the DPX files?
I would like to save my color graded stills in something other than TIFF files.
Since DPX is a listed format, how can I open files saved in it?
I know there is a free download for opening & viewing these files, I just can't remember where it is....

 

[Chris Nuzzaco]

Photoshop can open DPX files.

 

[jbeale]

Photoshop will open DPX. Not free, but I guess most people here already have it as an essential tool.

If you're serious about wanting something free, there are tools such as "Image Magick" which can read, write and convert DPX (and just about every other format you've ever heard of.) It's mostly command-line driven but there is a minimal-function display tool also. http://www.imagemagick.org/script/index.php

 

[Jeff Barklage]

Thanks!

 

[M Most]

I cannot find my answer on the RedUser's search page...
Does anyone know what program opens the DPX files?
I would like to save my color graded stills in something other than TIFF files.
Since DPX is a listed format, how can I open files saved in it?
I know there is a free download for opening & viewing these files, I just can't remember where it is....

On a Mac:
Photoshop.
After Effects.
Shake.
Nuke.
Graphics Converter Pro.
Iridas FrameCycler.
QT Player, FCP, or just about any Quicktime aware application, if you have Gluetools.

On a Windows PC:
Photoshop.
After Effects.
Digital Fusion.
Nuke.
Combustion.
Iridas FrameCycler.
Graphics Magick (free, also on Linux).

...among a number of others.

 

[Lucas Wilson]

On a Mac:
Photoshop.
After Effects.
Shake.
Nuke.
Graphics Converter Pro.
Iridas FrameCycler.
QT Player, FCP, or just about any Quicktime aware application, if you have Gluetools.

On a Windows PC:
Photoshop.
After Effects.
Digital Fusion.
Nuke.
Combustion.
Iridas FrameCycler.
Graphics Magick (free, also on Linux).

...among a number of others.

XnView as well. Free.

Lucas

 

[Paul Leeming]

XnView as well. Free.

Lucas

Seconded. XnView ROCKS!

Paul

 

[Dan Hudgins]

XNVIEW works with Cineon to DANCAD87.EXE...

XNVIEW works with Cineon to DANCAD87.EXE...

Seconded. XnView ROCKS!

Paul

Thanks for the tip on XNVIEW, it seems to work for Cineon to DANCAD87.EXE for editing, color correction, mixing, proxy, filmout, and DVD etc. I guess it would work with DPX as well, does anyone have a DPX frame made from RED ONE (tm) footage (that I can post processed to share here) they could let me download or email me so I can prove that I can CC RED ONE (tm) footage?

I have attached the Kodak DLAD (tm) in my Grading Curves #2 screen and a 1K downres. This was taken from a single input file, I am working on speeding up the FUSE command now to get better color depth from imported frames at higher speeds...

I tried GraphicsMagick 1.2 but it would not run on my machine, it said it needed a DLL, XNVIEW on the other hand runs great so far :waaa:

Maybe I went a little high on the contrast, but the gray blocks on my "widget" seem on target (lower left of the CC screen you cannot see them when the color is balanced, you can see the faint xor outline of my digital density probe on the gray box in the DLAD image, upper left in CC screen.) If I was not trying to get the gray on target and it was a real shot I would try to make her face a little more orange.

Anyway, if you have some DPX frames to share that I can show corrected, I would like to see if XNVIEW works well with RED ONE (tm) frames as a way to go to/from my programs to /from DPX... Thanks...

BTW: the images are for viewing at video gamma 2.4, the gamma block in the CC screen image may be off due to going to JPG, there are two dark and two light blocks near the gamma 2.4 test block in the lower right to set the monitor brightness and contrast... The spots in the brown patch of the 1K image seem to be JPG artifacts (not like that my dithered BMP output viewed directly, which would go to the filmout without compression, that's why I am working on all uncompressed processing to avoid blocking artifacts)?

 

[M Most]

I have attached the Kodak DLAD (tm) in my Grading Curves #2 screen and a 1K downres. ....
BTW: the images are for viewing at video gamma 2.4, the gamma block in the CC screen image may be off due to going to JPG, there are two dark and two light blocks near the gamma 2.4 test block in the lower right to set the monitor brightness and contrast... The spots in the brown patch of the 1K image seem to be JPG artifacts (not like that my dithered BMP output viewed directly, which would go to the filmout without compression, that's why I am working on all uncompressed processing to avoid blocking artifacts)?

You cannot simulate film output with any degree of accuracy with a 2D LUT. Film has significantly different saturation and brightness response, and thus a significantly different color pallette. The only way to simulate it is to use a proper 3D LUT that is built for the particular monitoring system and output path you are going to use, although for non-critical viewing, a "general purpose" version will sometimes suffice.

 

[Dan Hudgins]

Output to Print Stock directly.

Output to Print Stock directly.

You cannot simulate film output with any degree of accuracy with a 2D LUT. Film has significantly different saturation and brightness response, and thus a significantly different color pallette. The only way to simulate it is to use a proper 3D LUT that is built for the particular monitoring system and output path you are going to use, although for non-critical viewing, a "general purpose" version will sometimes suffice.

I am going to be doing the filmout directly to Print stock from monochrome seperation images, the film recorder program will make adjustments to the curves for the color seperation exposures, but the new print stocks are very good about crossover.

Most of the problems typical workflows have are from the dupe-negative processing, de-sat in the dupe-negative stock, and poor filtration of the light used in the film recorder and the printer used to print from the dupe-negative to the print stock.

Without all that fuss, the color will match well enough without any 3D look up table.

The film will be forced to come close to the CRT monitor look. When you print film a print of 6th generation does not look like 2nd or 4th generation even if you only work with monochrome seperations, so as a grader you learn what you are going to end up with.

The digital monochrome seperations are better than film monochrome seperations, and people printed film for 50 years without any 3D look up tables. Anyway 3D look up tables cannot simulate adjacentcy effects, so are not a "perfect" solution. I have a test frame feature that can be used to make film test strips so that you can check the film print directly in a projector on still fame (we have a edit projector that runs fwd/rev and still) before you run the long rolls.

Anyway, that's off the subject I was asking if anyone had some DPX files I could work with and show the results of.

If a 3D LUT was to be used to improve the matching of the filmout TO my monitor display the 3D LUT would come after the CC screens not before. If the color gambit of the filmout is not big enough to match my CRT monitor, the another lab should be used that has a good light house and uses bleach that takes ALL the silver out of the film stocks. From what I have seen in examples people are simulating bad lab work with their previews using a 3D LUT to muck up the color to match bad lab work. I know how good the new print stocks can get when printed using good equipment, and if you make seperation exposures directly to the print stock the gambit of the printstock should be close to what the CRT monitor can produce if not better. If I have frames from real footage I can play with a 3D LUT later to see if I can get the film to match the monitor better.

A quick walk around the SONY (tm) store watching the same film on LCD, DLP, OLED and such displays shows that the same content on different outputs do not match in any way shape or form, the best you can do is to keep things clear and simple and hope that the end result is pleasing.

With photo pinting one just tries to stay some place in the center of the color plus or minus 5CC, I am sure with digital color seperations better than film to film results can be obtained.

Yes, there would be issues going from my system to a labs workflow because they would muck up the color, but I am most interested in going from the RED ONE (tm) into my system along side the 35mm film scans. The best way to get into my system seems to have the RED ONE (tm) software output DPX files and then pass as much sensor data as I can into my high color depth processing.

If you are going out of my system to DVD there is no major mismatch since the DVD would also be viewed on a TV with video gamma about 2.4. So getting the images in with close to 48 bit data is a bigger issue than what happens after.

So if someone with a camera can shoot short junk shots under various lighting conditions, I can see how bad the noise in the RED ONE (tm) shots are relative to the film scans. And I can see what tone banding and JPEG2000 like artifacts are still in the DPX files from the RED CODE (tm) camera data. Since I do not have a computer that can work with RED CODE (tm) I am dependent on someone giving me some DPX frames.

 

[M Most]

Most of the problems typical workflows have are from the dupe-negative processing, de-sat in the dupe-negative stock, and poor filtration of the light used in the film recorder and the printer used to print from the dupe-negative to the print stock.

Without all that fuss, the color will match well enough without any 3D look up table.

I don't necessarily agree with that, but even if I did, in any practical system, one must accommodate the need for more than one print. In order to do that in any reasonable amount of time and for any reasonable cost, you need to create a negative.

The digital monochrome seperations are better than film monochrome seperations, and people printed film for 50 years without any 3D look up tables.

Well, yes, because there's no such thing as a lookup table in a chemical process.

Anyway 3D look up tables cannot simulate adjacentcy effects, so are not a "perfect" solution.

There is no such thing as a "perfect" solution in film, period. There are always soup variations and other imperfections inherent in a chemical process, no matter how tightly controlled.

If a 3D LUT was to be used to improve the matching of the filmout TO my monitor display the 3D LUT would come after the CC screens not before.

Well, yes, it would be applied as a view LUT, at the end of the processing chain, after any color correction.

If the color gambit of the filmout is not big enough to match my CRT monitor, the another lab should be used that has a good light house and uses bleach that takes ALL the silver out of the film stocks. From what I have seen in examples people are simulating bad lab work with their previews using a 3D LUT to muck up the color to match bad lab work. I know how good the new print stocks can get when printed using good equipment, and if you make seperation exposures directly to the print stock the gambit of the printstock should be close to what the CRT monitor can produce if not better. If I have frames from real footage I can play with a 3D LUT later to see if I can get the film to match the monitor better.

You seem to be going through an awful lot of trouble for a process that by its nature will produce only one print.

 

[Dan Hudgins]

Short run prints are more than enough for 90%+ of indy films?

Short run prints are more than enough for 90%+ of indy films?

in any practical system, one must accommodate the need for more than one print. In order to do that in any reasonable amount of time and for any reasonable cost, you need to create a negative.

For people who want to make a movie that can be taken to a festival you only need to make a few prints, if that many.

With a modified backlight on the LCD panel you can run the DIY film recorder faster so you can make short runs. I doubt any film I make would need more than 10 prints, and many peoples films would not even need that many, since if you sell the film the distributor can pay to have a printing negative made form your DPX files (and he can pay to have them re-Graded if need be by the lab making the match). There are also commercial film recorders that run at sound speed and output directly to print stock,

http://www.cinevation.net/cinevatio...?company=cinevation.net&page=p_cinevator.jsp&

For my DIY DANCINEL.EXE (tm) the rate would be at most maybe 3 to 5 seconds per frame, but there is no rush for a micro budget filmaker, you just want something that looks good for the few showings you would get no matter how much you pay anyway. The images at your premier will look better than the prints made in the "normal" way, so the audience reaction might be more favorable and help you sell the film, which will then not do as well with the mucked up prints made by the distributor.

XnView should take care of converting Video gamma BMP back to DPX with the values in the right spots for interchange, yes/no?

To get back on the subject of DPX files from RED ONE (tm) are there any links to DPX (tm) frames made from RED ONE (tm) footage that people can play with and re-post here?

 

[Paul Leeming]

Here you go (click on the image below for full 4K 2:1 DPX, 32MB):



This is RAW data converted to DPX 10bit log format using Camera RGB and RedLog curves to maximise the log information.

Also for XnView there's a great forum where Pierre (the author) posts regular bugfixes etc. He's very receptive to changes if needed so if there's any problem with the program re: colour spaces etc I'm sure you can chat to him directly to help solve the problem.

HTH

Paul

 

[Jeff Barklage]

XnView for MAC OSX has been removed from the XnVIEW website!
The webpage states that there were too many problems with OSX..so they removed it's download until further notice!
Damn!
Well...any other free DPX viewers that work well?
ImaigMagik seeme very complicated...
I have the latest FCP-6 with COLOR....it is suppose to have a DPX reader, but doesn't seem to like the DPX files I generate in RedAlert!

 

[jbeale]

Well, Blender ( http://www.blender.org/ ) is free, multiplatform, and it can read DPX. But if you thought ImageMagick was complicated, you ain't seen nothing yet!

 

[Lucas Wilson]

http://www.cinevation.net/cinevatio...?company=cinevation.net&page=p_cinevator.jsp& XnView should take care of converting Video gamma BMP back to DPX with the values in the right spots for interchange, yes/no?

Except that XNView (last time I checked) only dealt with 8-bit images. Any higher bit-depths are opened, but truncated at 8 bits.

This may have changed, since the last time I used XNView for anything other than casual image viewing was about 2 years ago. But at that time - 8-bit.

Caveat Emptor.

Lucas
-----
ASSIMILATE, inc.
LA, CA, USA

 

[Dj Joofa]

On a Mac:
Photoshop.
After Effects.
Shake.
Nuke.
Graphics Converter Pro.
Iridas FrameCycler.
QT Player, FCP, or just about any Quicktime aware application, if you have Gluetools.

On a Windows PC:
Photoshop.
After Effects.
Digital Fusion.
Nuke.
Combustion.
Iridas FrameCycler.
Graphics Magick (free, also on Linux).

...among a number of others.

Don't know if Gimp can do DPX, but Gimp is a very powerful multi-platform (Windows, Linux, Mac) freely available, open source tool. It is almost Photoshop, and the interface is also similar to some extent.

 

[M Most]

I have the latest FCP-6 with COLOR....it is suppose to have a DPX reader, but doesn't seem to like the DPX files I generate in RedAlert!

Color reads DPX files generated from Redcine just fine.

 

[Dan Hudgins]

Dithering 8 bit output

Dithering 8 bit output

Except that XNView (last time I checked) only dealt with 8-bit images. Any higher bit-depths are opened, but truncated at 8 bits.

This may have changed, since the last time I used XNView for anything other than casual image viewing was about 2 years ago. But at that time - 8-bit.

Caveat Emptor.

Lucas
-----
ASSIMILATE, inc.
LA, CA, USA

Thanks for the "heads up", but since I have a dithering of the high color depth internal processing in my CC chain when outputing to uncompressed 24bpp BMP, if that caries over to the DPX then at 4K the result should be closer to 10bit per primary video gamma in the film out.

I am using dithering to overcome the limitations of the 24bpp (video 32bpp is 24bpp plus 8 bits fill) D2A converter in PC video boards used in the DIY film recorder. My work is not finished, but the results so far seem usable.

It would be better to use high color depth output, something I plan to look into when I get more important issues delt with like speed of throughput which I am making good progress with so far... My goal is to make a useful 4K uncompressed DI system that any filmmaker can use to produce a quality film with with a great sound mix, it seems usable now, but some ease of use issues and throughput issues are in progress...

It does seem to work (as far as I have been able to test) under Windows XP (tm) with NTFS, so it should be able to access the required TB of data, and XP (tm) should be able to network to MAC and Linux as needed for frame processing.

 

[Peace Villow]

Just for FUN

Just for FUN

HTH

Paul

Graded Version :innocent: