Welcome to our community

Be a part of something great, join today!

  • Hey all, just changed over the backend after 15 years I figured time to give it a bit of an update, its probably gonna be a bit weird for most of you and i am sure there is a few bugs to work out but it should kinda work the same as before... hopefully :)

Understanding ISO with the RED ONE

I've been using the zone system for years with digital cameras, but I take a different approach. I base all measurements off of the highlights, and then work my way down into the shadows. Its a different way of thinking for sure, but I find it far more helpful to work based on the sensors point just before clipping, or more specifically, my images clipping point after color correction, if that makes any sense... Again, this is my own personal take on it, and its served me well.

I don't recommend shooting at high ISO though, just shoot at the lowest you can afford and set the exposure to favor highlights. I can see a big amount of confusion surrounding the reds selectable ISO settings. You can reproduce the ISO 800 results at ISO 320, just underexpose. Thats all you are doing when you shoot with a red at high ISO, underexposing. I sort of wish Red didn't have this selectable ISO feature really, because its just confusing people.

Another big issue is people not understanding light metering in general, both reflective, and incident. I can also see a lot of confusion surrounding the post processing end of things, because a TRUE Ansel Adams zone system takes post processing into account as well, not just on set exposure.
 
Chris,

I agree with you completely. Use the lowest ISO setting. Test for your clip point, expose for that and determine necessary fill. The point being, if you test and know the clip point you can start with metering the highlights, and if you expose for the highlights you have just protected them. If you can not come up with the required fill you will have to compromise the highlights with over exposure. If you know the zone system and know how to use incident and reflective meters you can achieve almost zero noise. My most noisey shots have been night scenes where I blindly rated at 320. I now rate at 160 and work my way down from the clip point. Exposing to the right is dangerous for those who do not stay completely on top of the exposure. But staying on top of the exposure is a good DP's job.
 
Could someone enlighten me what is the difference between dynamic range of the source and dynamic range of the final product.
Mr. Sontag mentioned that maybe it is latitude, but here is the description of latitude:
http://en.wikipedia.org/wiki/Exposure_(photography)

Latitude is the amount of light in stops you could over or under expose and still recover the highlights or shadows. Most photographic materials have 1/2 stop latitude.

I am repeatedly coming out with more stops of dynamic range after the grading then I have in the source material.

It is almost impossible to squeeze histogram below the source DR and have your frame looking good. Other way around, as soon as you bring all to look good you get extra stop or two.
How do you call the finale dynamic range of the camera? PRODUCTION DYNAMIC RANGE?
 
I fully agree with Mac's and Hans point of view.

We've been discussing this already months ago. In the end the we need to wait for build 16, which will have a true sensor ample and true sensor histograms (at least thats what Jim wrote us).

Thats the barely missing "feature", seeing you are safe within your shooting.

I'd prefer a EVF / LCD grayscale display (happily we are almost there since build 15) - sensor neutral - while feeding a "white balanced" color display to the HDMI / HD-SDI outs for the video village. So the DOP can concentrate on framing the image and maintaining enough dynamic range for post.

Most DOPs want to "make" the colors with the colorists, some are colorists themselves now. So it makes just sense. Its their responsibility. Right now the correct tool is missing, but we know its on its way, so all we need to do is choose a work around.

The discussion if REC709 is misleading or not is one of your point of view. If you treat RED ONE as a video camera, so you aim at downconverting 12 bit RAW to 10 or 8 bit YUV HD-SDI, you'll be probably fine with the REC709 approach. Welcome in Sony/Panasonic/JVC/... land (of confusion?) RED's 12 bit are not exceedingly much for this purpose (many systems have e.g. 14 bit sensors nowadays).

But most of us tend to use RED ONE as a camera system with a wider dynamic range, so we want to keep the full 12 bits RAW data. But that inherits that we do not get a white balanced result, as this is an artificial processing step that can be done for preview (via display LUT) or simply in post (which many prefer). Shooting RAW is about the same as shooting film and scanning it with onelight and a dynamic range of 12 bit @ ~9 fstops contrast.

At this point I want to note that one of the biggest point for me on RED ONE was - beside shooting native RAW with higher DR - to shoot full sensor 35mm, so having the theatrical DOF we all want for telling stories (at least sometimes, when we can pull focus). Right now we can not get that at full resolution 4K at the desired 60p, which would have been really cool for slowmo effects. The compromise - as Greame told me - is the 3K solution.

I do not fully agree with that, and would have loved to see full sensor recording at 60 Hz, regardless of the stored resolution, being it HD, 2K, 3K or 4K or 4+K, just "getting the full sensor at 60 Hz". Many of us will not produce higher than HD 1080p standart for years. As RED is already proceeding, the future will solve this. I just would have loved to see a solution for us "almost now" that simply gives us the 35mm DOF at 60 Hz. It has a big impact in our kind of productions.

Axel

Btw, great thread, but that has been stressed already :)
 
Could someone enlighten me what is the difference between dynamic range of the source and dynamic range of the final product.
Mr. Sontag mentioned that maybe it is latitude, but here is the description of latitude:
http://en.wikipedia.org/wiki/Exposure_(photography)

Latitude is the amount of light in stops you could over or under expose and still recover the highlights or shadows. Most photographic materials have 1/2 stop latitude.

I am repeatedly coming out with more stops of dynamic range after the grading then I have in the source material.

It is almost impossible to squeeze histogram below the source DR and have your frame looking good. Other way around, as soon as you bring all to look good you get extra stop or two.
How do you call the finale dynamic range of the camera? PRODUCTION DYNAMIC RANGE?

Hi Andrew,

I may not be the best person to answer you question, but what you experience is that you shoot "flat" images covering a really huge dynamic range (thus they look flat). Then you choose a section of the dynamic range to map it to the black-to-white range of your display at a rather high contrast, so you increased the contrast of that choosen sub-region. You colors pop up, the image quickly feels more like you are used to from TV etc. You can do this with cropping off under/over colors or by trying to maintain some of the higher dynamic range with a curved LUT (tail/shoulder slope curve etc.)

On 35mm film you had originally like 2.0 densities above D-min on camera negative film. Only a sub range of roughly 1.3 densities above D-min could be printed into the positive, so you needed to decide with the printer light settings if you cut off blacks or white, and where to shift the 1.3 out of 2.0 region. The positive itself comes to screen with about 4-5 densities above D-min if I recall correctly (its been a while I measured blacks on positives manually...).

To many poeple it appears strange that film experts call a flat image a "good image" and a camera that seems to produce so flat images a "good camera".

The point is simple:

When you have the flat image with a very high bit depth, you can extract almost any desired sub-region, and push colors almost wherever you want. Trying the same thing with a typical HD-SDI 10bit signal of a "standart" HD-CAM or Panasonic is almost impossible. You are forced to expose right, right from the beginning, when shooting in e.g. HD. There is the difference.

Hope that is somewhat clear.

Axel
 
Yes, you are right, my clips after grading, viewed on 8 bit Blu-ray are flat. But on wide gamut monitors it is a different story.
I couldn’t put my finger what is going on first since while previewing single frame as a 16 bit TIFF all looks good.
I got myself 12 bit color depth PLAYSTATION3 and started to experiment with 12 bit color depth using AVC codec on it. It is the only inexpensive Blu-ray capable box that can play AVC files without going exactly through Blu-ray standard. Things looks bit better there. Still have to figure out how to map well 16 bit TIF clips to AVC 12 bit.
QT 16 bit with H.264 is not going as nicely through the conversion as AVC or maybe I can`t figure it out how to do it right.

http://mediacoder.sourceforge.net/wiki/index.php/H.264


http://playstation.custhelp.com/cgi...TEmcF9zZWFyY2hfdGV4dD1tMnRz&p_li=&p_topview=1
 
Andrew,

now you confuse me quite a bit :(

You are playing back AVC codec files with 12 bits (???) on a playstation3 on which monitor?

I hardly believe the playstation will really process anything more than an 8 bit output. But even if it really does 10 or even 12 bits (which technically would be allowed on HDMI), which display can actually show it?

There are clearly displays out there that have e.g. HD-SDI inputs at 10 bit, but on the panel side you enter the problem: Which display panel gives you really the desired color range in internal processing?

I've been reading LCD panel manufacturers catalogs quite deeply and not long ago, so I know that about almost every larger vendor has simply and only 8 bit displays on the market. The only higher bit depth panels I've found had been specific medical imaging grayscale panels with 10 or even 12 bits, but again, purely monochrome.

If you can really encode an H.264 file with 4:4:4 profile: Which software lets you do this?
I would really like to know about it.
Further - assuming the PS3 could really play such high bit depth files and really output those extra bits to the HDMI port, which display do you use to view the material?

I've been investigating the market for years now, looking for true "beyond 8 bit" display technology, and have found almost nothing til now. If there is something new, I need to know about it, as I really have the need for it :)



About your problem:

I am wondering why you say your graded material looks flat when being viewed with 8 bits, while it seems to look better when viewing 16 bit tiff files?
Or an encoded AVC file at higher bit depth?

I suggest you give us some details on the way you process the files, which software, which displays etc. That would be interesting and hopefully helpful to get a clue whats going on.

Axel
 
I believe the highest bit depth panels now are 10-bit (not to be confused with bit depth driving the backlighting, or bit depth of the processing).

In the whole process, it's somewhat helpful to send a 12-bit signal. You'll pick up some rounding error from the conversion between Y'CbCr and R'G'B' (a lot of codes in Y'CbCr describe unreproducible colors and are therefore wasted), some rounding error after signal processing (e.g. to change/compensate for the transfer function of the LCD), etc.
Whether or not you'd see a difference depends on the source material and whether the display is trying to do wide gamut (in a wide gamut system like xvYCC your bit depth is spread out over a wider gamut, so it's helpful to have a higher bit depth). Red footage would likely have noise in it and this will effectively act as dithering so I don't think it's that likely you'd see any banding artifacts. On top of that, it will also depend if any other dithering is going on (some dithering techniques are better than others).

But on wide gamut monitors it is a different story.
It might be that the monitor is intentionally oversaturating the image except for memory colors. This is to make the image look "better". (Obviously this goes against standards.)
 
Most of Sony Bravias have 10 bit wide gamut.
http://www.sonystyle.com/webapp/wcs...10151&langId=-1&productId=8198552921665153951
I am using this one:
XBR5 series features: Full HD 1080p, Motionflow™ 120Hz, BRAVIA Engine™ PRO, 10-bit display panel and processing, x.v.Color™ capability, Deep Color (HDMI v1.3 option), 1080/24p input capable, PhotoTV HD,
HDMI 1.3 supports 12 bits and more.
PLAYSTATION# is outputting 12 bit but AVC codec I am using is 10 bit only.
I have seen some adds about 12 bit wide gamut HDMI 1.3 attached LCD panel somewhere, I will try to dig it out for you. Also some laptops will do 12 bits I think, check Dell Precision 6300 or XPS 1730

I started to do research this subject after you asked all these questions and it looks like when I play AVC from the Blu-ray disk it uses only 8 bits.
However if I put just the file on the fast memory stick USB attached it plays at 10 bits and probably 12 if I can put my hands on 12 bit AVC High 4:4:4 codec I would try it.

8GB USB memory stick I have here can go 300 Mbits/sec on this port so AVC at 4:4:4 will go no more then 300Mbits/sec at 1080p 24

Again, don’t try it from the Blu-ray disk, try it from USB port.
One more thing, PLAYSTATION will not show AVC files on your memory stick unless you press triangle and you ask to show all files on the stick. Then navigate to your avc file and play it.

Anyway I have to make sure that my After Effects actually outputs 10 bits to the file, if not this will explain my problems. It could be also the conversion between Y'CbCr and R'G'B'.
 
Andrew,

thats a pretty interesting thread now. We've started some research over here. A collegue found a freeware codec / encoder called "X.264" which may also include the 4:4:4 profile (He'll be showing it to me tomorrow). It seems to be for Apple right now, maybe there is a PC version too, can't say.

The point is that 10 and 12 bit display tech will allow us for better grading. In SpeedGradeDI we have e.g. dithering, so we can gain extra 1 or 2 bits in color display, but its rarely used. If the noise floor is high, you may not need it. But native 10 or 12 bit displays would be an experience anyway.

Axel
 
Ive seen x264 for about a year now in the pirating world. They have been ripping Blu-ray discs to x264 and xvid, After Linux runs from the Hard Drive on a 360, I'll be able to watch them somewhere other than my computer.

I had 'Night at the Museum' over a year ago, and the size was 525mb, and the quality was twice as good as a 700mb Axxo xvid equivalent.
 
Woow.... 4:4:4 I have to have it, where and how much?
I was playing today with Adobe Encore, looks like it has build in transcode from AE.
I didn't finalized the AE project, just left it there after color edit and went to Encore that supports now linking back to AE projects. Nice shortcut. It also opens the Encore to do much better and faster transcode.

I have noticed that AVC codec that comes with AE is bit better in color but much worse in resolution compared to MPEG2 - 40Mbits/sec
Different story in the Ecore, here you have a choice of MPEG2 and H.264. that is AVC
Well, clear winner is H.264 AVC better colors and uses 10MB/sec less speed. (30-35)

As to the resolution maybe it was bit suggestive but looked bit better, maybe because colors were better. But almost no difference.
At the end I am disappointed with the Encore codec set and even more with the AE.

Maybe they will update codecs since AE can process end to end in 16 colors so limiting codecs to 10 bits is too restrictive.

Have to ask around about SCRATCH and H.264 QT codec that has choice of 8 and 16 bit.

Do you know if QT H.264 choice in SCRATCH is actually 4:4:4 AVC wrap?
How the heck they can say 16 bit QT codec choice if H.264 supports maximum 14 bits as far as AVC is concerned. What QT is wrapping around here?
 
I was playing today with Adobe Encore, looks like it has build in transcode from AE.
You may get different results, but all of my internal Encore encodes look like crap. You either have to use MPEG2 (which is much less efficient than H.264 and ends up looking far more compressed), or their version of H.264, which always ends up giving me a half-vertical resolution output of my file. I dunno what the deal is... I've been searching on various forums and it seems this problem is very common.

Encoding in After Effects is an extra step than should be necessary, but I have been very happy with all of the files outputted from there. I crank up the bitrates to the max and get absolutely beautiful Blu-ray files from my RED footage. I'd like to hear about your Encore experiences and if you find the H.264 component worthy.
 
Häakon, I was so impressed with your 2Kclip.mov of crying boy that I was working very hard trying to get the same quality.
I did try SCRATCH and AE.

The best results so far I am getting in AE when I render to AVI uncompressed from 16 bit TIFF.
Well 4.2GB for 30 sec.

Next is to use Encore by importing this AVI to it and outputting to H.264 blu-ray. (35MBits/sec maximum)
Remember to enable color management option in the project setup in AE and switch to wide gamut RGB there.
Then link it from Encore and make sure that floating 32bpc is enabled.

Remember, bpc is not bpp

24 bpp (bits per pixel) is 8 bit per color
32 bpp is 3X8 bit per color plus padding or alpha channel.
32 bpc (bits per channel) is 32 bits per color

I have to check your observation of half vertical, didn’t see it when referring/linking the AE project from Encore.
Project that I was checking all these was in AE 4K 2:1 16 bit TIF.

However the whole this experiment ends up with some bending visible, since I suspect that even uncompressed AVI file is 8 bit 4:2:2 maximum.

I even question output from SCRATCH or REDCINE to 16 bit QT H.264. First the H.264 standard at its highest level is supporting only the maximum of 14 bits per color.
 
I even question output from SCRATCH or REDCINE to 16 bit QT H.264. First the H.264 standard at its highest level is supporting only the maximum of 14 bits per color.

H.264 was not designed to be compatible with Scratch or RedCine. Even if the standard says 14 bits, I think the assumption (much like JPEG) is that the are representing non-linear data (unlike Red's data). 14 bits representing non-linear data is a good range.
 
Hi Andrew,

we did some research today.

According to our own search results and the current entries on

http://en.wikipedia.org/wiki/H264

we came to the conclusion that there is currently NO software that supports 11-14 bit high profile 4:4:4 H.264 encoding at all. See the lower part table in the wikipedia.

Even 10 bit seems to be almost out of reach.

I really wonder which software you are using there, that gives you these options...

The X.264 codec allows us to select many options that would yield towards the 14 bit capable profile, but the important setting is kept in baseline, so ending up in 8 bit output too.

We found two major issues with the H.264 encoding as of now:

1. The slight gamma shift, which is driving us nuts over here, as the very same file is having at least 3 different looks in playback on the very same incarnation of Quicktime, depending on how you play it etc.

2. The artefacts on lower quality settings. The 75% setting we used for COLD STORAGE teaser was almost the lowest end we can go, without further sacrificing the look with artefacts.

We did today encode a music video at 75% with H.264 and it was really bad experience, the background was a crappy ocean of banding. All the subtle gradients from the shooting where dying there. We'll try again tomorrow with higher quality settings to see how or if we can safe it.

What is your experience with brightness / gamma shifts in H.264 or QT in general?

Axel
 
we came to the conclusion that there is currently NO software that supports 11-14 bit high profile 4:4:4 H.264 encoding at all. See the lower part table in the wikipedia.

I think this has to do with licensing issue. Only the Baseline profile of H.264 is (relatively) "free" of licensing issues as major companies, such as Microsoft and others, which have had some participation in the development, have announced.

Other profile comes with a licensing / royalty cost and software developers have to be concerned about that.

Even 10 bit seems to be almost out of reach.


We found two major issues with the H.264 encoding as of now:

1. The slight gamma shift, which is driving us nuts over here, as the very same file is having at least 3 different looks in playback on the very same incarnation of Quicktime, depending on how you play it etc.

In the worst case, since x264 is open source, one can re-write the appropriate sections of the codec.

2. The artefacts on lower quality settings. The 75% setting we used for COLD STORAGE teaser was almost the lowest end we can go, without further sacrificing the look with artefacts.

We did today encode a music video at 75% with H.264 and it was really bad experience, the background was a crappy ocean of banding. All the subtle gradients from the shooting where dying there. We'll try again tomorrow with higher quality settings to see how or if we can safe it.

You have to understand that operating H.264-type compression on linear light data (I assume you tried to compress Red's data directly with H.264) is not very conducive as the transforms used in such compression are not expecting such "noisy" data.
 
Additionally, if I remember correctly MPEG-4 allows the use of Wavelet transform also in addition to DCTs, so you might like to see if you can find a software that lets you do full-frame encoding in MPEG-4 and compare that with H.264. For DCT-based MPEG-4, H.264 wins easily.

H.264, is also BTW part of MPEG-4, just not to be confused with "regular" MPEG-4. H.264 is typically designated as MPEG-4 Part 10, where as the "regular" MPEG-4 is MPEG Part 2 (if I remember correctly).
 
That´s correct. Usually the work of a DIT or a nice DP is to measure the contrast of the scene and therefore choose the best "ISO setting" of the camera.

Imagine you have a scene with only 3 stops of contrast (ie: a brick wall in a cloudy day), you´d like to set the camera at a very low ISO, lets say 250.
Now if you have a high contrast scene (sunny day with clouds and shadows) you might want to set your mid grey higher in the zone system (usually in the V) and therefore you do this by increasing the ISO setting.
Basically this is old information, since this is similar to the way slide films should be exposed using the zone system (exposing to the highlights).

allright i'm confused already. i learned the zone system through B&W still photography and from what i know 18% grey is 18% grey, that is to say if you want to realistically render a scene, a grey card is in zone V by definition, regardless of ISO.
 
Back
Top