FORUMS: list search recent posts

10-bit RGB versus 10bit-RGB LOG

COW Forums : AJA Video Systems

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
Grant Lovering
10-bit RGB versus 10bit-RGB LOG
on Jul 3, 2008 at 11:50:35 am

This might be a question (or 3) for Gary or anyone who has gone through nutting this out...hopefully you are listening :-)

If material that has been shot digitally in LOG, say Sony SR's S-LOG (or Genesis Panalog) does it make any difference as to whether it is captured with the 10-bitRGB codec or the 10-bitRGB LOG codec (using Kona3) or do you get the same result capturing all available data with just a different QuickTime wrapper?

Further to this are the values mapped differently on capture, with the 10bit-RGB doing a linear conversion adjusting gamma and potentially clipping data on capture while LOG maintains the native LOG format recorded by the camera during capture?

Cheers,
Grant



Return to posts index

gary adcock
Re: 10-bit RGB versus 10bit-RGB LOG
on Jul 3, 2008 at 3:35:42 pm

[Grant Lovering] "shot digitally in LOG, say Sony SR's S-LOG (or Genesis Panalog) does it make any difference as to whether it is captured with the 10-bitRGB codec or the 10-bitRGB LOG codec (using Kona3) or do you get the same result capturing all available data with just a different QuickTime wrapper? "

this one is full of pit falls.

simple rule of thumb is
Shoot as video - stay linear (video is considered linear)
Shoot in Log - capture into log.

That being said,
I cannot say enough good things about Bob Monaghan's Glue Tools for DPX workflows-
his tools for working at this level are far beyond anything else on the market- V3 allows users to create everything from low res SD proxies up to 2K output files within the FCS 2 workflow. Additionally he now has direct integration with your Kona cards for playback. ( DPX capture can be handled thru Aja's VTR Exchange app)




gary adcock
Studio37
HD & Film Consultation
Post and Production Workflows
Inside look at the IoHD




Return to posts index

Ramona Howard
Re: 10-bit RGB versus 10bit-RGB LOG
on Jul 3, 2008 at 5:06:42 pm

With Rave, if the material is log it is captured as log. That is the wonderful thing about working with DPX. Nice to see others finally getting the picture as to what is really possible.

>his tools for working at this level are far beyond anything else on the market-

Gary, wouldn't go quite that far. Bobs stuff is very good and I recommend it often, but plenty of us are doing, and have been doing this for years :)

Cheers,
Ramona
Rave - The DDR choice of Colorists worldwide.



Return to posts index


gary adcock
Re: 10-bit RGB versus 10bit-RGB LOG
on Jul 3, 2008 at 6:30:17 pm

[Ramona Howard] "Gary, wouldn't go quite that far. Bobs stuff is very good and I recommend it often"

Ramona,
the Gluetools plugin is an FCP enabler for your (and anyone else's) hardware and when it comes to usability, I do not think there is anything superior when we are talking about handling DPX frames within a Final Cut Pro timeline.

The Gluetools apps allows for an honest to god real workflow with DPX files within FCP, since his tool do not capture frames, only allow for proper handling of those frames in an application that inherently does not like nor can it handle sequential frames correctly.

gary adcock
Studio37
HD & Film Consultation
Post and Production Workflows
Inside look at the IoHD




Return to posts index

Ramona Howard
Re: 10-bit RGB versus 10bit-RGB LOG
on Jul 3, 2008 at 7:26:45 pm

Yup, that's why we recommend it. Makes for a smooth workflow between all platforms and applications. No messy codecs or better yet no worries about what is going on within those codecs :)

You know me, just stirring it up. Wording is very important.

> I do not think there is anything superior when we are talking about handling DPX frames within a Final Cut Pro timeline.

That is more reasonable than the first comment and can't agree more :)

Even though this forum revolves around FCP, it is important for users to know that many solutions exist that use the AJA hardware AND not all are created equal, which is why I usually jump in. Wording, especially if it can be interpreted (which is pretty much everything) can lead people to think they are stuck or that it is the only solution, when in fact it is just the opposite. Clarity of information being shared is extremely important so as to not lend to the already wide spread confusion.....

FCP handling DPXs (natively) is very cool....now if only Avid would jump in, we would all have a much easier, well rounded workflow that translates well for both Film and television.


Cheers,
Ramona
Rave - Working natively with DPXs for 8 years



Return to posts index

Grant Lovering
Re: 10-bit RGB versus 10bit-RGB LOG
on Jul 3, 2008 at 9:35:25 pm

Ahhh....I enjoy a lively thread.

Hey, the thing I was keen to know though is when these custom LOG formats (which are designed to protect highlights and lift darks to avoid noise when shooting in the digital world) are coming off tape and into Kona is the codec (10-Bit RGB or 10-bit RGB LOG) doing anything differently for how it is interpreted or is the mapping just layed down 1:1 in each codec so theoretically you can transcode between or does the codec map the data differently effectively doing some conversion on the way in?

G.


Return to posts index


Gary Gowman
Re: 10-bit RGB versus 10bit-RGB LOG
on Jul 3, 2008 at 10:03:26 pm

Also check out the DI Handbook by Steve Shaw and Quantel. google it or it may be on the Qunatel site. There is good info explaining the differences between log and lin and our eyes.



Return to posts index

gary adcock
Re: 10-bit RGB versus 10bit-RGB LOG
on Jul 4, 2008 at 2:04:51 pm

[Grant Lovering] "Ahhh....I enjoy a lively thread. "

Ramona and I have know each other for awhile now, it is all good..


LOG was designed to allow for a greater level of data to be laid down in the same space.
Think Film. with a Negative, the user has a far greater level of control to reach into the shadows or print down highlights. Correctly processed LOG info has that same capability when it gets to a proper post facility that understands what to do with the data as LOG.

Linear -is just that - the data ( as luminance and color) increases in a straight line from black to white - it is mapped in a linear fashion. Video is a linear space by default.

you need to think about the math here.

Linear -- where the shortest distance between 2 points ( white and black) is a straight line.
linear video is usually not greater than 10bit

LOG- from Logarithmic (exponential) where by modifying that distance between 2 points using a mathematical calculation ( think of a curve or parabola) you increase the available number of points on that line since it is now curved line between the white and black points- there is more space between those points -hence more bits. Stills from a LOG capture can contain 10, 12, 14 even 16 bits or more of information, whereas the recording to a video format as LOG limits the data's bit depth.


Transcoding between LOG and LIN spaces is NOT recommended if you are not fully aware of the needs and usage for the final output.
Traditionally LOG files work with LUT's ( look up tables) to shape or redefine the characteristics for Viewing and playback on LIN devices -like a monitor.

Most LOG files are not deemed viewable on LIN devices without some sort of correction or modification


gary adcock
Studio37
HD & Film Consultation
Post and Production Workflows
Inside look at the IoHD




Return to posts index

Grant Lovering
Re: 10-bit RGB versus 10bit-RGB LOG
on Jul 4, 2008 at 11:17:21 pm

Thanks Gary, I've got a good understanding of LOG and Lin and the different scales.

What has me interested is material that has been shot digitally in a custom LOG space (such as Sony's S-LOG) the pictures would still need to go through almost an S-LOG (or proprietary LOG, Sony, Panavision etc) to LOG conversion prior to doing a LOG grade to get them ready for a film out as they each use there own curve to define what their LOG is which is not the Cineon convention everyone typically works with.

I came across this which shows recommended conversion of LOG to LIN (I realise LIN is not what we are looking for here, but stick with me while I use it to explain) for a Panalog source. You can see the values are quite different to a typical LOG to Lin which suggest we have logarithmic data which is off a very different base to the Cineon base typically used when doing a data scan or providing source for a film out (because Panalog has used it's own curve (or maths) to distribute the LOG data and has it's own LUT for previewing it in LIN).

http://prolost.blogspot.com/2008/01/panalog.html

Conversion of Panalog to LIN
10 Bit Black Point: 0
Internal Black Point: 0.0
10 Bit White Point: 681
Internal White Point: 1.0
Gamma 1.480
Highlight Rolloff: 0

Typical LOG to LIN (Cineon Spec)
10 Bit Black Point: 95
Internal Black Point: 0.0
10 Bit White Point: 685
Internal White Point: 1.0
Gamma 1.7
Highlight Rolloff: 0

I realise this example is showing a conversion to Linear which may be confusing the issue, but was just using this data that I could find to show that these proprietary LOG formats that use Custom LUTs to look okay are encoding to a unique LOG space.

In my opinion this has proven that you need to determine the gamma and black and white point variance of these custom LOG formats to the typical Cineon base so you can do a LOG to LOG conversion for grading and then supply of Cineon DPX files for film out.

So.......we have some data for Panalog from Stu Maschwitz's blog. Anyone got data for Sony's S-LOG? Apparently Sony are releasing a LUT for S-LOG soon which will probably contain the maths for those inclined to work out the translation from S-LOG to Cineon LOG or S-LOG to Lin.

Unless someone has data, I'm guessing the best thing is to do your own tests with a LOG to LOG translation of the S-LOG with the film LUT that previews your target and play with the gamma, black and white point until you get something balanced that you can then grade. Any better ideas or has someone already done the maths for S-LOG?

Alternatively I've tried making this way harder than it needs to be and just load in your Panalog or S-LOG material and start grading with your LUT that is previewing your target film out and it will be all good.

:-)


Return to posts index


Stu Maschwitz
Re: 10-bit RGB versus 10bit-RGB LOG
on Jul 5, 2008 at 7:20:29 am

Hey all, just one quick note about those panalog settings: They convert panalog to true linear light values. They are not designed for video preview. Video is NOT linear, it is gamma encoded. To go from Panalog to video, you could use my panalog-to-lin settings and then apply a gamma 2.2 or sRGB conversion to the linear data.

-Stu Maschwitz
http://www.theorphanage.com
http://www.prolost.com


Return to posts index

Ondra Mayer
Re: 10-bit RGB versus 10bit-RGB LOG
on Feb 2, 2009 at 8:05:36 pm

Hello All,

So to be sure with panalog capture. The capture process using kona 3 with 24psf DL / 10-bit log RGB setup > then importing this qt mov to AE and using panavision plugin for AE to display preview is ok?

What I want know if kona 3 log codec is capable store other log formats (panalog, sony for ex.)?

Many thanks,
Ondra.


Return to posts index

gary adcock
Re: 10-bit RGB versus 10bit-RGB LOG
on Feb 3, 2009 at 4:27:50 pm

[Ondra Mayer] "capture .. kona 3 with 24psf DL / 10-bit log RGB setup > importing qt mov to AE and using panavision plugin for AE to display preview is ok?"

In my testing that workflow was fine with the FCP and CS3 suite- I have not tested with CS4- and the FCP/ AE setup has changed- so test, test, test.





gary adcock
Studio37
HD & Film Consultation
Post and Production Workflows

Inside look at the IoHD
http://library.creativecow.net/articles/adcock_gary/AJAIOHD.php




Return to posts index


Mark Visser
Re: 10-bit RGB versus 10bit-RGB LOG
on May 8, 2009 at 5:48:46 pm

Apparently Sony are releasing a LUT for S-LOG soon which will probably contain the maths for those inclined to work out the translation from S-LOG to Cineon LOG or S-LOG to Lin.

Has anyone heard any news on this? We're working with plates from an F-35 and need to convert to linear for VFX. Google finds this thread, but not much else.

thanks!

-Mark


Return to posts index

gary adcock
Re: 10-bit RGB versus 10bit-RGB LOG
on May 11, 2009 at 12:49:17 pm

[Mark Visser] "Has anyone heard any news on this? We're working with plates from an F-35 and need to convert to linear for VFX. Google finds this thread, but not much else. "

Not a lot of F35 users on the cow.

the app you are looking for is from steve shaw in the UK http://www.lightillusion.com/cubebuilder.htm
the Cubebuilder tool offers users the ability to create and modify Sony S-log LUTS.



gary adcock
Studio37
HD & Film Consultation
Post and Production Workflows

Check out
http://www.aja.com/kiprotour/

Inside look at the IoHD
http://library.creativecow.net/articles/adcock_gary/AJAIOHD.php




Return to posts index

Jendra Jarnagin
Re: 10-bit RGB versus 10bit-RGB LOG
on Aug 28, 2009 at 3:00:29 am

[Mark Visser] "Has anyone heard any news on this? We're working with plates from an F-35 and need to convert to linear for VFX. Google finds this thread, but not much else. "

[Gary Adcock} Not a lot of F35 users on the cow.

the app you are looking for is from steve shaw in the UK http://www.lightillusion.com/cubebuilder.htm
the Cubebuilder tool offers users the ability to create and modify Sony S-log LUTS.

You can also download Sony's CVP file editor, (which has instructions on how to use it) and export a LUT that is the reverse of S-LOG.

Jendra Jarnagin, DP
http://www.floatingcamera.com


Return to posts index

<< PREVIOUS   •   VIEW ALL   •   PRINT   •   NEXT >>
© 2017 CreativeCOW.net All Rights Reserved
[TOP]