Posts Tagged ‘canon’

I know it might sound childish or even presumptuous, but i feel pretty uncomfortable by this whole Technicolor hype. The “Marvels Cine Style” profiles i’ve developed the past years are much more advanced, better balanced and actually delivering more dynamic range. I really am a modest guy and you’d hardly hear me bashing other software developers, but this whole Technicolor style is actually sad. It’s curve suspiciously follows our Marvels Cine Style, with the blacks dialed up (limited!) and the whites dialed down (limited too!).

Warning to all users of the Technicolor Cinema-style flat-profile! Using the Technicolor style will severely limit your dynamic range!
Blacks are lifted and whites are limited! With only 255 values between black and white for each color, snooping 10% off at both sides is simple “DR theft”! You remove another 10% of DR when using the Technicolor LUT (via LutBuddy); that’s probably two stops total!
With the Marvels profile 3.4 you use all available digital space. You can make the Marvels profile as flat as the Technicolor profile by altering it’s curve, dialing the contrast setting further down.

Marvels Film (Martin Beek, and others) have not commented on this and my opinion reflected here is my own and not that of Marvels Film. Nevertheless, the Marvels Cine Style picture profile for Canon DSLR cameras can be downloaded here: http://marvelsfilm.wordpress.com/marvels-cine-canon/

Jorgen.

The latest version of the Marvels Cine picture style for Canon HDSLR cameras is now out of beta and available on the blog of Marvels Film.

They have dedicated a “sticky page” to the subject: http://marvelsfilm.wordpress.com/marvels-cine-canon/

Jorgen.

I have been woken up by a text message from a desperate Martin Beek (@ Marvels Film), at around half past one in the morning a few days ago. His message was, that all of his testfootage shot in Belgium and Italy had been turned down by his customer because ALL the shots had micro-moire patterns in the hair of people. He was even thinking about selling all his DSLR equipment and giving it all up.

I’ve written such a filter before, as a part of a digital imaging system for testing c-mount lenses, to reduce chromatic abberiation. I’ve now revamped this methodology into a FCP plugin called the “Marvels DSLR Moire Filter”, as a tribute to Martin, who is now in seventh heaven for all day now and as happy as a monkey with seven tails.

I don’t want to bother you with the technical details and point you to my Plugins page (see top menu) to read more (and download) this little gem that i’m pretty pleased with myself.

The plugin is not specifically designed to work with the Canon cameras, but with any other DSLR video camera that uses the line skipping algorithm to reach it’s HD resolution. As far as i know, that’s all of them except for the Panasonic Lumix GH series.

To cut a long story short… If…

  • your footage is spoiled by red/blue streaks ad pixels in natural patterns and materials such as hair, grass, leaves and water…
  • putting a soft filter on the lens did not help…
  • you were not aware of it during shooting (not visible on LCD)…
  • you are using a lens that is much too sharp…
  • grass, leaves, gravel, sand or turf is swarmed with red and yellow worms (could be LSD though) …
  • you’re ready to smash your DSLR camera into the next tree…

Then this plugin is something you’ll love!

Well?! What are you waiting for! It’s almost free… (if you consider the semi-obligatory donation of US$10 if you REALY like it)

PLEASE READ THE REMARKS SECTION ON THE PLUGIN PAGE ! This plugin is not specifically designed to battle aliasing problems and the resulting moire from regular patterns such as brick walls. I don’t say it won’t work, but there are other and maybe easier and better ways to prevent “roof and wall” moire.

Martin has helped writing this article and by the look of it he was in a good mood… 😉

Cheers,
Jorg.

Remark: see update above.

I am in the process of developing a new picture style for the Canon 5D mkII camera, that will be used for a new feature film DSLR production. Base of the style is the Marvels Cine 3.3 style, that i’ve developed for Marvels Film, using my own tools (instead of the Canon profile editor). It is NOT a replacement for the renowned Marvels Cine style, but a new experimental version that will be used to shoot early-autumn exterior scenes in Italy this year.

If you live and shoot in a less-grey/blue environment than northern Europe, consider using this style instead of the Marvels v.3.3.

New in this version is a slightly altered s-Curve to provide better luminance linearity in the 65-75% range (e.g. skin tones) and uses the Neutral style as a basis, instead of the Standard style as in the Marvels Cine v3.3 style. This to move the overall tone a bit more to green instead of the pronounced orange skintone look of the Marvels v3.3.

Everyone is invited to download, test and use this new profile. Please check back regularly if you want to get the eventual release version.
Any comments, testresults, raves and rants are much welcomed..!

Download here: http://mediatube.marvelsfilm.com/marvels_cine_v34beta.zip

By the way… all picture Marvels styles (most syles not by me, except for 3.3 and 3.4beta) can be downloaded as one ZIP archive here.

Jorgen.

I normally don’t like doing this, but i am going to quote a weblog article by someone else. This was a comment on Martin Beek’ s blog article about a Canon Picture Style that i found specially interesting:

I personally have stopped using it (HTP) because it produces vertical banding noise (pattern noise) with almost any ISO number, under some specific conditions. Philip Bloom also reports that he stopped using it. Probably because it produces unpredictable results that are not easy to detect on-set.

HTP further limits your ISO range and gives you an extra amount of noise. Remember than Highlight Tone Priority (HTP) is essentially underexposing all your photos by 1 stop. ISO 200 with HTP is essentially an underexposed ISO 100 RAW data pushed up to look like ISO 200. That’s how it preserve highlights. You’re essentially pushing an underexposed ISO 100 RAW to look like ISO 250 in that image (since you raised exposure in ACR another +1/3 stop).

I personally don’t use HTP as said, but i follow another rule of thumb as an alternative, specially for commercial work that doesn’t allow for experimenting:
1 – use ISO 160, 320, 640, 1250 and 2500 for low contrast / low light shots
2 – use ISO 200, 250, 400, 500, etc. (ISO 100 and 125 are on their own) for well lit / high contrast shots

1, because the broken ISOs have considerably less noise, but also have less latitude (highlights clip at raw 12650)
2, because the whole ISOs have more noise, but retain highlights much better (highlights clip at raw 15300)

Together with Marvels Cine 3.3 this will make your day… ;-)

Cheers!

Martin

I think that Martin is mostly right, although i have to look up and compare the numbers he quotes. I must add one important remark: this is typically targeted towards shooting video, not RAW photography.

Update: quick link to PP: here (zip file).

I have developed a new type of flat picture profile for the Canon “d” series video DSLR cameras. This profile has been devised and tested using the Canon 5DmkII, a MacBeth colour card, two different calibrated light sources (3200k & 5600k soft floods), Adobe Color, Adobe Photoshop  and a few software tools i have developed myself. This picture profile uses 10 S-curve node points and mathematically wraps correctly around the existing build-in Standard Profile S-Curve.

Goals were striving for correct colorimetric reproduction (no weird chroma artifacts), no gaps or bumps in the resulting curve, linear behaviour in the skin-color-exposure range and a few more points that are described in more detail by Martin Beek on his weblog http://marvelsfilm.wordpress.com. There you can download the picture profile and read about it’s uses for video shooters (i’m a colorist and mathematician, not a film maker…).

Go have a look and use it for free in your camera –  happy shooting!

Jorg.

In contradiction to what you might assume, the Canon DSLR cameras video conforms to the ITU-R BT.601 standard for SD video recommendations, instead of the ITU-R BT.709 HD standards.  This means that several suppliers of software for video- and film-makes try to convert the color decoding matrix of the Canon DSLR video footage. An example is the new 5DtoRGB software that assumes (in it’s helpfile) that the Canon 5D (1D/7d/550d/T1…) uses the ITU-R BT.601 color decoding matrix (“Primary chromaticities”). THIS IS NOT THE CASE!

Canon conforms to the ITU-R BT.601 standard, but NOT in regard to the “Primary chromaticities”. The Canon DSLRs exclusively use the sRGB color space for video, which shares the ITU-R BT.709 primaries. So, no color space conversion or matrix conversion is needed for Canon DSLR video footage and ITU-R BT.709 should be maintained throughout the workflow as “Primary chromaticities”.

Confused?! Read on..!

In order to convert the “real world” colors recorded by the camera’s image sensor to electric (and digital) signals, the camera uses a Decoding Matrix to “map” colors to numbers. All digital cameras do this and the broadcast industry has provided several standards to the camera manufacturers, to make sure (this was the idea…) red is red and blue is blue… and sixpence is sixpence…  So we ended up with tons of “standards”.

These standards are laid out in several ITU-R recommendations. They don’t only describe decoding matrices (“Color space” or “Primary chromaticities”), but also Pixel count, Frame rate, Digital representation, Standards Conversion, Luma coefficients and  Transfer characteristics.

Manufacturers, leading broadcasters and standards institutes (i used to work for one…) flooded us with recommendations and Decoding Matrices that the industry and video producers should adhere to. With the age of digital HD video came a few new ITU-R standards that are generally in use for digital video cameras:

Primary chromaticities

BT.709:

 

CIE 1931 xy chromaticity diagram showing thegamut of the Rec. 709 color space and location of the primaries. The white point is shown in the center.

RGB color space parameters[2]
Color space White point Primaries
xW yW xR yR xG yG xB yB
ITU-R BT.709 0.3127 0.3290 0.64 0.33 0.30 0.60 0.15 0.06

Note that red and blue are the same as the EBU Tech 3213 primaries while green is halfway between EBU Tech 3213 and SMPTE C.

The trained eye can see from the chromaticity diagram above that the ITU-R BT.709 decoding matrix (“color space” or or “Primary chromaticities”) is equal to that of the sRGB color space used by the Canon (video) DSLR cameras. The images are from different sources and are only here for illustration; it’s the numbers that count of course.

So, why do Canon DSLR cameras use the BT.601 recommendation for SD video, instead of the BT.701 recommendation for HD video, while using the latter’s color space?! The answer is simple… The Canon DSLR video can not comply with the BT.701 recommendation in regard to frame rates and maybe even pixel count (read the BT.709 specs via the link above)! Canon DSLR cameras can not be considered HD cameras following the BT.701 recommendation and that might also explain why broadcasters such as the BBC do not accept Canon DSLR footage for HD broadcast as such.

Hope that answers some of the questions i’ve received by mail the last few weeks.

Kind regards.

Jorg Escher