Welcome to our community

Be a part of something great, join today!

Log in
  • 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 :)

RAW-RAW and REDCODE-RAW editing

Andrew M.

Andrew M.

Well-known member
Joined
Jan 18, 2007
Messages
1,885
Reaction score
0
Points
36
Location
Virginia EPIC #2441
RAW-RAW and REDCODE-RAW editing

So far looks like REDCODE-RAW editing has a lot of attention.
What about RAW-RAW editing?
Once RAW-RAW is on the hard drive will REDCINE take it like it does with REDCODE-RAW?

Will I be able to adjust all exactly like in the case of REDCODE-RAW?
What will be the difference/advantage?

One more question, I missed it completely, what is the difference between REDCINE and REDAlert?

Andrew
 
Yes, uncompressed RED RAW versus RAW is what I am talking about.

Maybe we are making mistake getting RED with the RAW port.
Since the material will have long shelve life I want to make sure that it will be stored in the original format as it comes from the sensor.
Who knows what future holds in terms of debayer algorithms and getting more from the data once captured.
I understand that once RAW from the camera is compressed to RED RAW you can not put is back to the original form. RED RAW is visually lossless not mathematically lossless so no reverse path once you capture only compressed RAW from MISTERIUM.
 
The only difference between the two forms is compressed vs. uncompressed. You can still benefit from different future debayer algorithms (for example) with REDCODE RAW.

The uncompressed RAW port is not available yet. Hence there is no post workflow either, but expect it to be brutal. You're looking at 350 - 1000 MB/s (that is megabytes, not bits!). That needs some serious storage hardware to work with.
 
Rob Lohman said:
The only difference between the two forms is compressed vs. uncompressed. You can still benefit from different future debayer algorithms (for example) with REDCODE RAW.

The uncompressed RAW port is not available yet. Hence there is no post workflow either, but expect it to be brutal. You're looking at 350 - 1000 MB/s (that is megabytes, not bits!). That needs some serious storage hardware to work with.
Click to expand...


I don’t think we can still benefit from the different future debayer algorithms.
Once Graeme will come up with some revolutionary new way of converting RAW to REDRAW, the old already recorded material will not be able to benefit from it.
Also, any bugs in the current red raw compression algorithm, like some artifacts or compression outcome that will prevent good green keying in some situation will not be able to benefit from new improvements in the raw compression or debayer treatment.
Once compression is done by current software you can not uncompress to the original data. Maybe Graeme has some trick to do it, and I am sure there will be always a way around in post, but the original data once compressed and not recorded will be lost forever.

Andrew
 
REDcode RAW can benefit from future demosaic and image processing - there is no doubt on that.

Graeme
 
Graeme Nattress said:
REDcode RAW can benefit from future demosaic and image processing - there is no doubt on that.

Graeme
Click to expand...

Wait. Did I read this correctly? Previously shot material can be inproved with a future release of the code?

David
 
Yes because the de-Bayering is not baked into the footage. (Unless you record Redcode RGB, which is different and not the same as Redcode RAW.)

Once compression is done by current software you can not uncompress to the original data. Maybe Graeme has some trick to do it, and I am sure there will be always a way around in post, but the original data once compressed and not recorded will be lost forever.
Click to expand...
That is true. However, it is practical to use a little bit of lossy compression. With the cost savings you can improve quality elsewhere or spend more on archiving it right.

You're also talking about a very, very small hit in quality. In practice, the show will likely be mastered in 2K or lower resolution... and that's likely what you're most interested in archiving. You'd get the biggest future-proofing by first trying to finish and master in 4K (which is no easy feat if you have VFX shots).

At some point you have to make compromises / juggle your resources. Opting for Redcode RAW compression isn't very bad at all considering it's visually lossless and you will almost never see the difference. Whereas something like 4:2:2 "uncompressed" is not visually lossless, yet an industry standard interface for moving material and dubbing tapes.
 
Yes, I've improved the demosaic many times now up to the current version. Old footage can and has benefited from this. That's RAW for you! That's REDCODE RAW for you.

Graeme
 
Graeme Nattress said:
Yes, I've improved the demosaic many times now up to the current version. Old footage can and has benefited from this. That's RAW for you! That's REDCODE RAW for you.

Graeme
Click to expand...

So the codec is as modular as the rest of the camera, etc. You rock! You roll! You rumble!

I am just getting my head around the fact that the footage shot today can actually look better a year from now. What other format at can offer that?

Thanks Graeme...

David
 
Film

Film

As scanners improve and as print stocks improve, the original film keeps getting better, as in The Wizard of Oz and Gone With the Wind. Isn't it great to know that the Red Camera offers us a similar future proofing of our work?
 
I thought the original film from Gone with the Wind was gone (with the wind). Wasn't it originally 4:3, then cropped for widescreen and the original lost?

Graeme
 
So are we wasting money by buying RAW port?
I agree that for 2K we probably do, but for 4K, we need 8M unique R and G and B pixels out of 12M of bayer pixels of Misterium.
Maybe I don’t know all what is going on when compressing RAW to REDRAW but if Graeme is saying that we can apply different debayer algorithms later in the future then this REDRAW is really RAW.
So how RAW is REDRAW?
It is single channel
It preserves the unique value of every pixel?
It preserves dynamic range of the value red from the pixel.
It deals with S/N ratio unique for every pixel.
It deals with uniformity of the values from adjacent pixels?

I put the question mark where I couldn’t find the answer to the question on this forum.
I know that right now 1TByte is a lot, but in 3 years will be nothing.
Well, by then we probably will have RED II but we will be stuck with RAW from RED I.

Andrew
 
REDCODE RAW is raw. It's a record of the sensor in terms of the light hitting each pixel. It's also compressed 12:1 or thereabouts. We do that in such a way that the end image looks "visually lossless" ie, if you eyeball compressed and uncompressed, you can't really tell which is which. Dynamic range is un-altered by the compression. The precise values of each and every pixel are altered, but in such a way that many will be the same or near enough, and those that differ do so in a way that's not too visible.

Graeme
 
Looks like I will reconsider the $6.5K for RAW port.
Any future improvements/revolution in resolution or noise canceling will probably come from inter frame micro movements of picture content or camera and as long as REDRAW preserves each and every pixel value with 10 bits resolution we should be OK.

Andrew
 
Andrew said:
Looks like I will reconsider the $6.5K for RAW port.
Any future improvements/revolution in resolution or noise canceling will probably come from inter frame micro movements of picture content or camera and as long as REDRAW preserves each and every pixel value with 10 bits resolution we should be OK.

Andrew
Click to expand...
The RAW port is going to require drives that can sustain a write speed of about 300+ megabytes per second. Currently that means any one of only a couple of Raid systems. THis means portability is an issue and that you will basicly be doing the equivilant of shooting to tape. Maybe tha RAW could work in a studio set-up, etc.

I do not understand the raw port for anything but stills. The amount of data that needs to be read, written and archived is 12times the amount you would need for REDCODE RAW.

David
 
I need some clarification, from you tech master. If an algorithm is used to compress something, can you not unwind (for lack of a better term) the footage with another algorithm back to raw.

If you can do that, would it be possible some day to. Say take a 4k Maya sequence and compress it into redcode. Do a redcode offline edit, and then decompress the finished timeline. Then do your conforming and print to film for distribution.

I mean If Captain Picard can get La Forge, to recombine the anti-mater, through the Dilithium matrix to restore the impulse engines, so they can escape the gravitational pull of a twin pulsar. All before the end of the third act.

Anything is possible.
 
adobeone said:
If an algorithm is used to compress something, can you not unwind (for lack of a better term) the footage with another algorithm back to raw.
Click to expand...

This is the nature of a compressor / decompresser. As long as the algorithm is not demosaicing the material, then its still RAW, just compressed RAW. What determines how close you can get depends upon if the process is lossless or lossy. The journey through both paths of the process can have zero impact if its a mathematically lossless process. The issue is trading off the footprint (size) and the quality. That is where the real art is.
 
You must log in or register to reply here.
Back
Top