FS7: LUTCalc v2.4 - Print charts and levels, Alexa Classic 709, DJI D-Log, etc.

C

cameramanben

Well-known member
Hi All,

I've just uploaded a new version of LUTCalc to the lutcalc.net website, GitHub, the Mac OSX App Store and the Chrome Web Store:

Run LUTCalc page

* New Feature - Print out the Stop vs IRE chart, plus exposure table and IRE / 10-bit values for each stop from -8 to +8.
* New Feature - LUTAnalyst automatically sets output gamma and gamut to the analysed LUT.
* New Feature - Alexa Classic 709 (Alexa-X-2) gamma and gamut.
* New Feature - Amira Rec709 gamma.
* New Feature - Experimentally-derived DJI D-Log gamma for Zenmuse X3 / Osmo.
* New Feature - ProPhoto / ROMM RGB colourspace and gamma.
* Feature Change - Improved LogC calculations from Arri's ACES IDTs.
* Feature Change - Improved luminance coefficient calculation from colourspace primaries.
* Feature Change - Improved flexibility in preset range settings for particular curve / format combinations.
* Feature Change - added 14-bit (16384 point) output option to 1D cube files.
* Feature Change - Linear and conventional gamma options are now listed under 'Linear / γ' rather than 'Linear / Rec709' to avoid confusion.
* Feature Change / Provisional Bugfix - changed the way that Dolby PQ is scaled due to user feedback.
* Bugfix - xy/uv gamut outlines were failing for LUTAnalyst colourspaces.

As ever, feedback gratefully received at contact@lutcalc.net. Very sorry for being rather (really) slow to reply lately, excuses are heavy filming schedule and a wife and daughter keen to share some of my spare time ;-). I really do read all the feedback, it's just that anyone who has seen my forum posts will know that I tend to fall off the deep end in framing answers :-),

Ben
 
A bit more info...


Things have been quiet for a while as I've been busy in the day job (phew!). During the gap I have also been adding to and adjusting LUTCalc, so have now released v2.4.

A piece of feedback that I've had is that it would be handy to generate pdfs or printouts of charts and IRE / 10-bit values. Above the chart I have now added a 'Print Chart' button to do that.

I've done a fair bit of analysis of the Alexa Classic colour space (ALEXA-X-2 in the Amira), outlined in a white paper 'ALEXA LogC Curve - Usage In VFX' from Arri. Basically the matrix parameters (which Arri gives) are applied to a tone map (which they don't) rather than linear data. I'm happy that I've figured the tone map sufficiently accurately, so have now included the gamma and gamut to LUTCalc. Let me know what you think!

The Amira has a slightly different standard Rec709, which as far as I can tell is not a straightforward tonemapped matrix; I have included the gamma curve, but not the colourspace / gamut for that.

Recently I bought a DJI Osmo which includes a 'D-Log' option. I couldn't find any documentation about what D-Log actually is, so I carefully tried to figure it out through testing. The resulting gamma curve is now included.

When the LUTAnalyst tool has analysed a LUT, the resulting Gamma and Gamut output options will now automatically be selected. This now seems like something I should have thought of to do ages ago, but hopefully should clear some confusion I have had from a user or two!

Another user asked for ProPhoto / ROMM colourspace and gamma. Done.

Lastly, alongside some mathematical fine tuning, I have change the 'Linear / Rec709' submenu to now read 'Linear / γ'. It was strictly speaking inaccurate, my intention being to keep things understandable for dabblers as well as experts, but I have now had enough emails from the expert-y side asking where the linear and gamma options are to hope that this is clearer!

Ben
 
Fantastic. Particularly excited about the D-Log - the Phantom 4 has D-Log as well.

Theoretically, does the Alexa stuff you added enable us to achieve a better S-Log3 to Alexa look LUT?
 
Hi David,

thanks for that. I don't have the colour space figured for the DJI as yet, but the D-Log contrast curve is a good fit (to my Osmo anyway ;-) ). One thing I took from it is that the D-Log approach is rather like stills, defining headroom above mid gray as around 3 1/2 stops, with the flat contrast giving access to shadow detail at 8-bit. Video cameras and their log curves tend to base their ISOs around five, six or more stops of headroom - the advantage being that if you cut from one shot to another you should have a good chance of holding the subject's midtones constant without risking white clip.

As such it is worth playing with the 'Knee' customisation in LUTCalc to tune curves designed for lots of headroom down to D-Log's approach.

The gamma and gamut options listed as 'Alexa-X-2' are based on the Arri white paper I mentioned, which prints a set of matrix values to be applied to a tone curve. I figured out a couple of different ways to come up with the tone curve which agree well. This Rec709 approach is what seems to lie behind the online Arri LUT Generator.

However, from the Amira onwards, Arri seems to have taken a more nuanced or customised approach, ie I haven't been able to come up with a matrix and tone curve combination which gives a decent fit. The difference is subtle, but there. That is why the Amira includes 'ALEXA-X-2' as a separate option in camera.

In other words, give it a go and let me know what you think. ;-)

I'm still going to keep plugging away - my intent with the Alexa-X-2 analysis was to use it to figure the underlying maths of LC709A, LC709 and V709, with the hope of dramatically shrinking the file sizes needed to include them in LUTCalc. The vast majority of the filesize of LUTCalc is in LUTs representing complex, nonlinear colourspaces. Plus I love a puzzle!

Ben
 
Anyone else unable to load LUTs created in this version? Premiere says "invalid file" unable to load LUT. Earlier versions worked fine.
 
I've created quite a few 16x16 .cubes with V_2.4 that are all loaded (and even 'embedded') into PP CC_2015 fine.

:-) jason
 
basspig,

if you could upload a LUT which isn't working, I'll take a look and see if I can get to the bottom of it. Thanks,

Ben
 
Hi Basspig,

sorry for the slow response - I managed to miss this thread for a while so didn't see your upload. Someone on the Sony forum also mention the difficulty and emailed me an offending LUT. Ridiculously simple fix and very silly it got in there. Very sorry for that too!

There are two bits of code that the online version uses which I didn't write - filesaver.js and blob.js. They help cross-browser compatability in saving files, and when I uploaded the last version of LUTCalc, which I had tested and tested, for some reason I thought it a good idea to check that filesaver.js and blob.js were up to date. After all my testing. Silly boy! Filesaver.js started adding a 'byte order marker' to the start of the file. If you open a problematic cube LUT it shows up as some weird characters before 'TITLE' on the first line. Delete those and all should be hunky dory.

I've just replaced filesaver.js with an updated version and all seems to be OK again. I just tested it ;-) . Lesson learned - I hope!

Very sorry about that and to anyone else who has had trouble with the online version. The Mac App version uses native OSX file saving and there is no need for filesaver.js or blob.js in the Chrome App, so neither of those should have been affected.

Thanks,

Ben
 
Hi All,

spoke too soon, in my haste I uploaded the corrected filesaver.js file, but forgot that I had also had to make a slight adjustment to the way it is called in LUTCalc (I did the coding on a train when I didn't have decent web access, then uploaded after getting home). Should now work as intended, though if you have been trying since my last post, you may need to clear your web cache to ensure that the new copy of 'lutfile.js' gets loaded.

Sorry again for the mistake; I'm only going to touch fileSaver.js and blob.js if there is a very good reason to from now on!

Ben
 
Excellent, you fixed it! Thank you for taking the time to explore the matter and root out the cause. I just created a LUT and it works as it should now.
 
You must log in or register to reply here.
Back
Top