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 :)

For those who have used them: How does Red compare to the Genesis and CineAlta?

Andrew,

You are technically correct. But, your results very likely have shadows deep in the noise floor of the RED One. And that is where the dynamic range of RED One becomes an interpretive dance.

How much noise in the blacks is a D.P. interested in aesthetically and technically? Zone II? Zone III? Under what lighting conditions is noise worse? Etc.

While the RED One technically delivers near 11 stops of dynamic range, a D.P. may only wish to use 9 stops in order to keep the resulting blacks cleaner. Thus he/she will lift shadows with fill light while protecting highlights until the scene fits within the desired range. Blacks may then be crushed in post to eliminate the noise altogether. (And vise-versa for film.)

This technique is nothing new and any D.P. worth their credentials is prepared to light a scene to suit the technical and aesthetic constraints of the recording medium, be it film, video, or a RAW camera like the RED One.

Best,

Michael


Yes, I agree with you, but there is the post to manipulate highlights.
Are we talking about 8-9 stops in the source material or we are talking about 10-11 stops in the final product that viewer is looking at?
I recently clipped blue channel in my shot. For the still photography would be almost unrecoverable.

Well, I spent few days on it, blue clipping didn’t want to go away.
Finally I learn how to unclip this damn blue and my sky looked natural again.
Well it saved me 2 stops in blue sky texture or banding, whichever comes first.

Experience from HDR photography came handy though.
 
Regardless of Algorithms, 10-11 stops latitude, propriatorey lab formulas, compression, and cost, the entire issue comes down to the SAG vs AFTRA contract. Money. Residuals. Benefits.

Regarding the new Dalsa camera? I stand corrected, as per my previous post, about the Dalsa. I have not used the newer camera. I think that anything Rob Hummel is involved with is going to be successful. He is a brilliant man. So are the people at Red.
 
How do you get 10-11 stops in the final product if you are starting with 8-9 stops in the source?



I had excellent exchange couple of weeks ago with Mr. Sontag about color temperature of the shot.
I see this DR is learning experience as well, for people experienced with the film and people experienced in the digital RAW, where both groups benefit from each other skills.

Where the stops are coming from? Well after you load the 12 bit wide data from RED we convert it in REDCINE to 16 bit TIFF. Usually I load it (move histogram to the right) bit up so the bits that are responsible for blacks, the first 4 bits are already set to highest value. (1111 or 16 in decimal)
Nicely packed 12 bits wall to wall put in 16 bit space give us the room to manoeuvre in the color correction.

Once you adjusted colors you will see that RGB histograms moved sideways some left and some right giving you extra two bits of data. Now from here you can go further to expand your data width.
 
Nicely packed 12 bits wall to wall put in 16 bit space give us the room to manoeuvre in the color correction.

So basically you are not increasing the inherent DR of the data; you are just increasing the precision of (future) calculations done on that data. Of course once you stuffed the 12 bits in 15 bits (I think AE uses 15 bits in 16 bit modes), then you can interpolate the in-between values to get faux-DR increase.
 
Well, I spent few days on it, blue clipping didn’t want to go away.
Finally I learn how to unclip this damn blue and my sky looked natural again.
Well it saved me 2 stops in blue sky texture or banding, whichever comes first.

Do you think optimising bit resolution by applying your "extra bit manoeuvre" 14 to 16bit tiff conversion would be meanless if we could color correct directly from the 14bit R3D original files?

Put this way : do we always loose DR when transcoding to an other colorspace/codec/bitdepth.

or this way : do all tests done with non R3D supporting software (lustre, davinci, baselight...) not show all the potential of DR due to poor conversion to the intermediate codecs?

Thanks

Pat

EDIT : ooops it's 12 bit raw not 14....
 
I am not the expert in color grading so don’t take it as a final answer, but I did play with few post software packages and I have some RAW data experience from still photography.

When we do grading we have to do it in some colorspace. After Efect allows you to choose even 32 bit per channel color space.
RED RAW arrives as a 12 bit Camera RAW colorspace (whatever it means)
When you will work directly on the RAW data in After Effects or in SCRATCH (supported now) you pick the color space and usually it will be 16 or 32 bpc
All adjustments are manipulated with 16 or 32 bit resolution depends what you pick.
You are acting on the data that is 12 bit only but you have 16 or 32 bit range.
So you have 12 bit of original data range regardless if you convert to the 16 bit TIFF or you work on the original data. Actually all your adjustments are not touching the original data at all until the time you decide to transcode it.
When adjusting the original 12 bit data you can increase the level of one of the colors or suppress some blacks. At the end you can make blacks blacker and filter the noise. You can recover the highlights by looking which color was clipped and then adjusting the others to show up the details and then bring it back up again to the level where two other colors didn’t clip but the one that did clip was not there.
Well you loos the clipped colors details but the whole texture is there in two other colors.
Best example is clipped blue on the sky. If you didn’t clip red and green you will be able to recover details of the sky anyway since information is there in two other color channels. Sure you have to bring the blue back, otherwise the sky will be discoloured.
Think about it as coloring the black and white movies. The texture is there, just color is missing.
Recovering blacks is the whole different ball game. You have to be good in the noise and surface textures manipulation. The best what works for me is recovering texture on the black clothing or irregular shapes. Also smooth black surfaces are easy to recover.

If you lit the scene to 11 stops and you shoot wall to wall on all 3 color channels you get 10 stops.
Don’t light the scene above 11 stops because you will be too busy watching everything but focus:-)
 
We do not need more stops as such. we need people with the skill to work within the stops they are given.

That is why some people can get nice shots from RED and they are happy and some are not hapy and they get crappy shots:-)

805_1210629456.jpg
 
I am a newcomer here, drawn by the comparisons of Red to budgetary black holes like Genesis.
Seems to me that in a controlled situation (DP, lighting) a reduced number of stops would have less importance than in a verite situation. Your crew has already put you in the ballpark to make sure the camera is not operating on the edge of its efficiency.
 
We do not need more stops as such. we need people with the skill to work within the stops they are given.

Maybe you don't need more dynamic range, but I'd sure find it useful... There are certain types of lighting effects, nice ones, that rely on a wide dynamic range, like that hot overexposed Robert Richardson spotlit effect where you expose for the ambient bounceback into the shadows from the hot backlighting. Having less dynamic range to work with becomes creatively restricting sometimes.
 
I am a newcomer here, drawn by the comparisons of Red to budgetary black holes like Genesis.
Seems to me that in a controlled situation (DP, lighting) a reduced number of stops would have less importance than in a verite situation. Your crew has already put you in the ballpark to make sure the camera is not operating on the edge of its efficiency.

I would not pay to much attention to the FAN BOYS in this thread. Listen more to what Mr Mullen has to say (below). I'm very sure that the DR is high on RED's list and in the future we will get there, but not to have a little more of it would be very very sad.


Fredrik Callinggard


Maybe you don't need more dynamic range, but I'd sure find it useful... There are certain types of lighting effects, nice ones, that rely on a wide dynamic range, like that hot overexposed Robert Richardson spotlit effect where you expose for the ambient bounceback into the shadows from the hot backlighting. Having less dynamic range to work with becomes creatively restricting sometimes.
 
In regards to the "more difficult to film out" argument I'd have to say I agree. Film stock LUTs do weird things to any color especially when you're trying to match LAD aims with a soup that changes so much. Bayer pattern sensors are a bit cludgier, but really its nothing that will prevent someone from making great imagery. It really depends on your DI software and the skill of your colorist.

Reds biggest connundrum is marketing to a very critical professional community, while encouraging indy films to make the jump from a prosumer to a professional camera. I could rip red a new one all day long technically, I can point out artifacts that will dispease anyone on everything from 35 to mini dv, but flat out I want red to do better and having a quible over 8-9 stops or 10-11 isn't going to do that.

Present the choice to a DP if you're a vendor. Or if you're a DP or a Director make the choice if the camera is what what you want aesthetically, and for God's sake TEST IT!
 
we own and use both, red and hdcam, and both systems are excellent.

both systems have their advantages and disadvantages, and which one should be preferred is really depending on the project.

to give some examples:

red + : s35/s16/23 DOF, hdcam - , only 2/3
hdcam + : 1080P 10bit 4:2:2 and 4:4:4 HD-SDI out, red - only 720/exotics
red + 120p, hdcam - only 60p
hdcam + instant on, red - boots ~90secs
red + resolution 4K CMOS, hdcam - 3*2.2MP CCD
hdcam + perfect readout & ECS, red - rolling shutter and no ecs
red + 12bit, hdcam - 8&10bit
hdcam + much lower lowlight noise, red - much higher lowlight noise
red + inexpensive, hdcam - expensive
hdcam + excellent (worldwide) service, red - rental house and fedex
red + dedicated team with cool attitude, sony - even national managers in the company often are shackled
hdcam + very good realtime in-camera colorcorrection (->battlestar galactica), red - only "preview"
red + almost no moving parts, hdcam - drum hours cost
hdcam + final delivery media as result, red - slow transcoding (and buggy) transcoding for delivery media
red + very high quality sensor-raw colorcorrection, hdcam -no raw
hdcam + fully functional & compatible realtime workflow in all formats (YUV, RGB, tape, disc) , red - rendering an incompatible with the huge mayority of the postworld without slow & unnecessary transcoding
red + users are pretty connected and have team attitude, hdcam - tough competition
hdcam + completly developped and wide post industry support, red - camera (slightly) beta, software very beta and/or buggy and most industry post needs transcodes etc.
red + good prices on accessories, hdcam - many extremly overpriced accessories
...
...
the list is sooo much longer, and i am really happy that now finally real competition for the digital cinematography market does exists.

i think it is widely overseen, that both companies and formats, sony/hdcam and red, help each another quite a bit - we see a massive increase in demand for hdcam tape recorders, and the sony cinealta 4K projectors are certainly the best reason to shoot with red in 4K (and vice versa).

On the other hand, red offered many of the missing tools sony oversaw for sooo long (as 35mm DOF, higher res, 120p etc) which made many productions still avoid digital cinematography and also finally brought the cost advantages of high-end digital cinema to the markets.

I suppose that once reds Achilles' heel, the incompability of its native format with the huge mayority of the existing postproduction world, is finally resolved, these advantages will really change many games.
 
Just out of curiosity Laguun are you referring to HDCAM or SR? I wouldn't even attempt to put HDCAM up against red. 1440x1080 with a 7:1 mjpegA type compression at 8 bit makes me cringe.
 
Just out of curiosity Laguun are you referring to HDCAM or SR? I wouldn't even attempt to put HDCAM up against red. 1440x1080 with a 7:1 mjpegA type compression at 8 bit makes me cringe.

Hi Jeff,

i refer to hdcam and hdcam sr, both of them. hdcam looks excellent - remember, its >18 Mbyte/sec for 1080p vs 28/36 Mbyte/sec for 4K on the red and has in fact less compression. For VFX-intense stuff Hdcam cameras can do uncompressed 10bit to disc.

Hdcam has almost all of the advantages listed above: its realtime to final output, has the colorcorrection, can be delivered, has 1080p realtime input/output, can be recorded and played out in realtime on discreet, quantel, avid, adobe, iridas, etc, is instant-on, has a worldwide service network etc.

HDCAM SR adds more quality than HDCAM at a much higher datarate than red (880 Mbit) in realtime or up to uncompressed 4K with the SRW-5800/HKSR-5804.

As i said: Its not better or worse, different scenarios make hdcam / sr / red the best choice. I would prefer HDCAM to show the olympics or having Steve McNutt as the DoP on Battlestar Galatica, and i would prefer red for many many fullfeatures. red to hdcam is not comparing apples to oranges, but blondes to brunettes.
 
Back
Top