RED Workflow Rant

[David Battistella]

I agree with just about everything you just said. But I would also say that Red had to put more into post production software than anyone else because their system didn't conform to any known industry standards. That has had a very large up side, but the downside has been the need to do the software themselves in order to allow post production to happen in the first place..

Mike,

There really was no other way for them to do it other than just wait to release the camera. They had to go this way because no one else had done it, there was no model. They were doing something completely new and merging Digital Still style into motion cinematography.

I'd still am glad to have taken the growing pains knowing that they are pouring the knowledge into EPIC and making a very solid post tool in RC-X. ( i know it needs to mature) With RCX, Clipfinder and the new SDK you can get anything you need done toward any media finish you can imagine, at a quality that is pretty fantastic.

This is what users asked for. The best advice to anyone now is to put away REDCINE, RED ALERT, RED RUSHES and focus on using RC-X and clipfinder.

David

 

[Mark L. Pederson]

I admire the way they have done that, and the effort they have put into it, but the fact is that it was a practical necessity given the nonstandard nature of the camera, its files, and the lack of finished quality HD video output.

That's the whole point Mike, isn't it?

It's NOT a "video" camera. It's NOT a "film" camera. So okay .. fine ... "nonstandard nature".

As for the "lack of finished quality HD video output" - I get better images out of my 720/59.94 output from the RED ONE that ANY video tap solution for 35mm film cameras - including the newest units which I have tested. And THAT is the point. The video output on the RED ONE is just that ... it's a VIDEO OUTPUT for monitoring - on a "non-video" camera.

So ... necessity ... sure. By design.

Why does everyone try so hard to squeeze RED into an old box made for yesterdays technology?

 

[David Battistella]

And THAT is the point. The video output on the RED ONE is just that ... it's a VIDEO OUTPUT for monitoring - on a "non-video" camera.

Thanks for reitterating this Mark. Many people want thier RED to be an ENG camera and be able to use it that way.

That's not really how it was designed and now that they have enabled the full frame out people can bring along a KiPRO and had their clients a video out that they can edit with when they get back to "the shop". Still need to copy all those files and transfer, etc, so I do not see the grand advantage of this.

David

 

[Stephen Lovett]

First and foremost, let me echo Luca's comments regarding Reducation.

I too attended the very first session and I cannot recommend it highly enough.

The key value is the opportunity to interact with folks like Michael and Steve and clarify and extend your understanding of best practices, and guidance in selection of the best path for your particular final output.

Michael,

If you can find the time to provide an update for your recommended work path in some form, I'm sure many people would benefit, myself included.

(Donning my flame proof suit)

On balance, I'm super happy with the capability of the camera, and the new MX sensor is nothing less than revolutionary.

I wrapped Friday on a two week shoot for a film that was set primarily near magic hour, extending into early evening.

I'm stunned by the capabilities of the new sensor.

RED has delivered on their core business objective better than anyone could have predicted a few short years ago.

Further they have catalyzed the industry, and all of us who seek after relatively affordable tools for small operations like mine that to compete head on with the best in the business owe them a profound level of respect, even when we disagree or encounter difficulty.

Many years ago, I had the unique opportunity to oversee the first couple of years of development of Internet Explorer for the Macintosh. We launched into an all encompassing race with Spyglass/Mozilla/Netscape to provide the best tool for using the web.

At it's inception, it was a balls out race. Those stories of months without a day off and sleeping bags under desks were not hyperbole.What mattered was delivering innovation as quickly as possible.

Over time however, as browsers became central to the way people worked and interacted with others, I was forced to adjust my thinking, and had to adjust the balance of innovation and features vs. stability and security.

IMHO, RED has reached that crossroads as well, and I think there is evidence that they are adjusting accordingly.

I'd bet the team understands that releasing a new sensor without supporting software that has reached final released status in relatively short order isn't optimal.

I had a conversation with Ted last year where he exhibited a clear understanding of the need for maturation of the software cycle before scarlet ships, and I'm confident these concerns are evaluated in the daily triage that is the reality of doing something nobody has ever done before.

Having said that, folks need to realize that they call it the cutting edge for a reason.

Is it frustrating that I've got to go rethink and re-qualify the way I work now that I'm MX'd for example, or based on a new color science release? I admit that it can be.

But the majority of that work is a result of continued innovation and evolution of the entire tool-set, most of which RED has nothing to do with.

It is what it is, but over the long term, I bet the market will continue to press for the slider to move more toward the stability end of the scale than it is today.

(I think people vastly underestimate the complexity of the software problems here, as I think some other major vendors are about to discover...)

There are areas however where I think RED is still getting in their own way.

While there have been improvements, unless you are a Reduser junkie, keeping up to date is a pita, as is finding out the relative stability of the current beta etc. And saying "well we told you it's beta" when there isn't a blessed release for MX isn't particularly helpful.

Having to remember to go to page three of a thread under crucial ordinance to find the current beta, and to hunt down threads for a quicktime update isn't in anybody's best interest.

Central location for all software, both beta and release would be a good start.

Further, example workflows that are updated by the RED team as technology evolves would be a big help.

I'm not recommending anything exhaustive, the combination of variables is far too high.

But there seems to be a couple of clear delivery paths that aren't going to be funded at a level to leverage the guidance of a major post house.

That should be the target.

I'd argue that this was essentially delivered in Reducation. It would be great if some general guidance was more broadly available.

I think that would help a great deal.

Stephen

 

[Jeff Kilgroe]

I agree with just about everything you just said. But I would also say that Red had to put more into post production software than anyone else because their system didn't conform to any known industry standards. That has had a very large up side, but the downside has been the need to do the software themselves in order to allow post production to happen in the first place. I admire the way they have done that, and the effort they have put into it, but the fact is that it was a practical necessity given the nonstandard nature of the camera, its files, and the lack of finished quality HD video output.

From the beginning, when I first really started working with RED's footage and tools in late '07, I was able to drop RED imagery right into my production pipeline. I'm not sure if I really agree that it did not fit into any industry standards... Actually I don't agree. Because we had REDCINE and RED Alert and shortly after came RED Rushes. As luck would have it, most of my work at the time was animation and everything was being rendered to 2K or 1080p image sequences and even occasionally 720p still. RED dropped right in, going to 2K DPX, TIFF or EXR.

This really was no different than shooting film. Sure, RED did not fit right into conventional HD "standards", but if REDCINE was viewed as your own personal film lab, it made perfect sense. And I still think of the RED tools that way for a lot of the workflow scenarios I set up. When you shoot film and have it scanned, you have various delivery options. DPX and TIFF are great choices and they are popular choices. These can also be used to compile low-res proxy clips for offline editing. RED's tools enabled this from Day 0.

I don't work in Hollywood LaLa land... I'm located west of Denver, CO, and south of Boulder. If I shoot 35mm film, I have to take it into Boulder or Denver and it takes at least 24 hours to get it developed and scanned, unless I want to pay extra. And never mind the cost of film stock, the processing and scanning is not cheap to begin with. RED has enabled a 35mm cine-style workflow for us and a cost savings unthinkable with 35mm film. All of this from the first day. Right from early 2008 when I received my RED One, I already had a few RED projects under my belt. Small ones, but enough to get me going, and I attended Steve Gibby's LA Red Test in December '07. I put my RED to work right away, the very first project I used it on was a shoot of two commercials. The DP was expecting to shoot on Varicam, but I talked him into shooting RED. We reviewed all the footage on-set, and made multiple copies as it was offloaded. I logged the footage onto a couple systems that evening and the next day had 1080p DPX sequences of all the clips as well as 480p ProRes proxies to edit with. At the time we used a beta version of GlueTools to wrap the DPX sequences with a QT wrapper for finishing. A few days later, we had the DP and client and a few other people over to review everything and they all got to see full 1080p versions on a 71" DLP display. Jaws hit the floor.

The point is, workflows have been possible since the very beginning. Many o those first projects I used RED on never made use of the QT pointer/proxies, never relied on direct support for those proxies or the R3D RAW files within the NLEs, etc.. In one of my posts earlier in this thread, I outlined three workflows I find myself using right now, but they are not set in stone, nor are they the only workflows I use, let alone what I think about. They say there's more than one way to skin a cat... By that analogy, I have truckloads of cat hides around here and they keep piling up.

Thanks for reitterating this Mark. Many people want thier RED to be an ENG camera and be able to use it that way.

But then again, RED can be an ENG camera if handled correctly. As Stephen Williams pointed out somewhere here in the last 24 hours -- either this thread or the other Stop Bitching thread, much news was shot on film until the last 25 years or so. Tape based systems pushed film away from the ENG market. Many news productions actually revived much of their film gear in recent years with the transition to HD. Now we are seeing various RED workflows and tools are making file-based systems more accepted, and we have already seen mass acceptance of Panasonic P2 and Sony SxS and XDCam. I'm not really advocating the RED One as an ENG camera, but I definitely see the upcoming DSMC system as having huge potential here. IMO, the 2/3" Scarlet is poised to make a huge splash in the ENG and EFP markets. That is my primary interest in 2/3" Scarlet -- EFP style production.

That's not really how it was designed and now that they have enabled the full frame out people can bring along a KiPRO and had their clients a video out that they can edit with when they get back to "the shop". Still need to copy all those files and transfer, etc, so I do not see the grand advantage of this.

Agreed. But so many people want to fall back onto what they already know. I've encountered many clients or productions shooting RED who are desperately trying to get what they shoot onto HDCAM or into a format they can push into a 10 year old workflow that they have settled into.

 

[M Most]

Is it frustrating that I've got to go rethink and re-qualify the way I work now that I'm MX'd for example, or based on a new color science release? ....
While there have been improvements, unless you are a Reduser junkie, keeping up to date is a pita, as is finding out the relative stability of the current beta etc. And saying "well we told you it's beta" when there isn't a blessed release for MX isn't particularly helpful.

All of what you say is true, but only if you choose to do everything yourself and thus accept the responsibility of needing to know all of these things. I would counter that by saying that in the days of film, a cameraman needed to basically know what a lab and a video post facility did, but he or she did not need to keep up with daily changes in the specific chemistry the lab was using, nor did they have to understand how to set up a telecine, run a DaVinci or Pogle color corrector, or set up a videotape machine. My point is that facilities still exist, and they exist for the same reasons - to perform the post steps in a professional way and produce the necessary products. The fact that it's possible to take on some of that yourself isn't the same as having it become the only alternative. The only people who need to know every gory detail and keep up with the fast paced changes are the ones who have to deal with them. If you choose to deal with it, you take on the responsibility of keeping up with it. Post production has never been as simple as the people now taking on the responsibility of transcoding files seem to think it is. Choosing to deal with it can either be a financial decision, a personal decision, or a practical one. But it is a choice, and there are other choices that are made by a lot of productions.

 

[M Most]

I'm not sure if I really agree that it did not fit into any industry standards... Actually I don't agree. Because we had REDCINE and RED Alert and shortly after came RED Rushes.

Uhhhhh... you seem to be proving my point, which was that Red had to do the software themselves in order to allow use of the camera. The last time I looked, Redcine, Red Alert, and Red Rushes were all Red's own products. Without them, at least for the first two years or so until the SDK was released, there was no way to use the files without using Red's own software. And there was no way to get a finishing quality video output out of the camera, either. For all practical purposes, there still isn't. Industry standards are media that can be used with commonly available equipment, and/or software. R3d files did not fit into that category until the SDK was delivered to various third party vendors more than 2 years (I might be a little off on that, but not by much) after the first release of the camera.

I always praise companies who go above and beyond what they need to do, and Red has fit into that category in many ways. But I have no illusions that the original software development was done out of some kind of altruism, rather, it was done out of necessity.

 

[Stephen Lovett]

All of what you say is true, but only if you choose to do everything yourself and thus accept the responsibility of needing to know all of these things.

Fair enough, I fully agree.

But let me make two points.

1.) A significant portion of the value proposition of RED for the market I'm attempting to serve is the ability to provide the image quality of 35 at a greatly reduced price.

Would I prefer to work in conjunction with a specialist? You Betcha. The fellow who was my gaffer on this last gig is making post decisions for Doggie Boogie, a feature shot here in the bay area. In our discussions, I highly recommended that he go with Light Iron's proposal, even though I have the capability to do the transcodes via my rocket, and LTO backup solutions etc. and I'm local. It really is in the films best interest to have an integrated proven path. But for the short I just completed, I ended up loaning raid storage to cover the last several days of shooting... The budget simply isn't there for me to hand it off.

But yes, this is because of the choice I made regarding the market I'm attempting to establish in the bay area.

2.) Then there is Scarlet.

I posit that for the majority of projects utilizing scarlet for acquisition, a one stop shop, or self finishing is going to be the most common methodology.

 

[Tim Morten]

Uhhhhh... you seem to be proving my point, which was that Red had to do the software themselves in order to allow use of the camera. The last time I looked, Redcine, Red Alert, and Red Rushes were all Red's own products. Without them, at least for the first two years or so until the SDK was released, there was no way to use the files without using Red's own software.

1. Yup, Red provided the software. So that we could use raw. A superior format. Bless them for it. Don't criticize them for it. If you want a lesser format, use a lesser camera.
2. It was not two years. FC support happened VERY early on.

And there was no way to get a finishing quality video output out of the camera, either. For all practical purposes, there still isn't.

You're losing credibility here. There still isn't? Avid. Adobe. Assimilate. DaVinci.

Industry standards are media that can be used with commonly available equipment, and/or software. R3d files did not fit into that category until the SDK was delivered to various third party vendors more than 2 years (I might be a little off on that, but not by much) after the first release of the camera.

Define "not by much"... You are off by a wide margin.

I always praise companies who go above and beyond what they need to do

As they say on Wikipedia: citation needed.

But I have no illusions that the original software development was done out of some kind of altruism, rather, it was done out of necessity.

Necessity would be a simple batch converter. Every app that Red has ever delivered has gone FAR above and beyond this.

I gotta say, I think you are being unnecessarily critical / ungrateful. For a workflow that you are claiming doesn't work, there are a substantial number of us ranging from small shops to major productions that are successfully posting on Red. If you really want to shoot with older formats, use an older camera...

Just my .02.

- Tim

 

[M Most]

1. Yup, Red provided the software. So that we could use raw. A superior format. Bless them for it. Don't criticize them for it. If you want a lesser format, use a lesser camera.

Did you actually read what I wrote? I didn't criticize anyone, anything, or any company. I simply pointed out that other camera companies were making cameras that output standard formats, like single and dual link HD video that could be recorded on existing VTR's, and therefore had no need to develop software to get usable products out of their cameras. Red did not have finished quality HD video coming out of the camera, and did not produce a "standard" file format (such as Quicktime, or MXF wrapped DNxHD or DVCPro) and was therefore in the position - by design - of having to deliver software to make use of their files possible. That isn't criticism. It's a comment on the fact that others here seem to want to assign endless praise for something that was simply a technical necessity.

It was not two years. FC support happened VERY early on.

That didn't - and still doesn't - allow high quality finishing in formats that justify the camera. For that, Red's software was required in order to transcode to high resolution, high quality formats. The Final Cut support allowed editorial to happen. Redcine (Red's product) allowed high quality output to be created.

You're losing credibility here. There still isn't? Avid. Adobe. Assimilate. DaVinci.

I really don't know why you're persisting in treating my comments with hostility as if I'm some kind of enemy, and I don't think I have to establish "credibility" here. The HD-SDI output of the Red One, past and current versions, and with either the old or new chips, is not of the quality required for final finishing in most cases. It is, as Mark pointed out, purely a monitoring feed. THAT is what I was referring to, live output from the camera. And quite frankly, I'm not trying to convince anyone here of anything. I'm simply commenting that Red's software efforts were a necessary part of the program.

I gotta say, I think you are being unnecessarily critical / ungrateful. For a workflow that you are claiming doesn't work, there are a substantial number of us ranging from small shops to major productions that are successfully posting on Red. If you really want to shoot with older formats, use an older camera...

Well, I gotta say that this seems to be another one of those times that I wonder why I post here. I don't know where you get the notion that I said anything doesn't work, nor where you got the notion that I have anything against Red. I happen to be the colorist on the only current network show that is shot entirely on Red ("The Gates"), I graded the pilot for another ("Detroit 187"), and am currently working on a cable series shot primarily on Red that will air later this month ("Covert Affairs"). I've been a supporter of Red from the beginning (ask Ted if you don't believe me), and I work with Red shot material every day, constantly experimenting to make it sing even more. That doesn't mean I can't point out some simple facts every once in a while.

Just MY .02.

 

[Jeff Kilgroe]

Uhhhhh... you seem to be proving my point, which was that Red had to do the software themselves in order to allow use of the camera. The last time I looked, Redcine, Red Alert, and Red Rushes were all Red's own products. Without them, at least for the first two years or so until the SDK was released, there was no way to use the files without using Red's own software. And there was no way to get a finishing quality video output out of the camera, either. For all practical purposes, there still isn't.

Perhaps I may be proving your point. And perhaps I'm not entirely sure what your point may be. Actually I do see your point now, and this helped clarify:

I simply pointed out that other camera companies were making cameras that output standard formats, like single and dual link HD video that could be recorded on existing VTR's, and therefore had no need to develop software to get usable products out of their cameras. Red did not have finished quality HD video coming out of the camera, and did not produce a "standard" file format (such as Quicktime, or MXF wrapped DNxHD or DVCPro) and was therefore in the position - by design - of having to deliver software to make use of their files possible. That isn't criticism. It's a comment on the fact that others here seem to want to assign endless praise for something that was simply a technical necessity.

So then that leads us to what the RED One means to myself and, I'm sure man others, who bought it. I did not buy the RED One with the intent of connecting it to any conventional HD deck, VTR, or capture system. Never, not once, have I ever been tempted to record a single frame being sent out over the HD-SDI or HDMI output on my RED One. I bought the RED One to shoot and edit 4K RAW from day one. If I wanted another HD camera, I would have bought another HD camera. The HD outputs on the RED One are for one purpose only -- monitoring. Do you record the feed from a video tap on an Arri 235? OK, I'm sure some probably do because they can, but really...

Other camera systems have been introduced and have required new hardware and/or software out of technical necessity in order to utilize their new formats. Take a look at HDCAM SR as an example. For the first several YEARS, it required the purchase of expensive decks/VTRs or Sony-only licensed software solutions. Not everyone who bought an F900R just shrugged and went, "meh... I'll just record to my old HDCAM deck". Many of them bought it right out of the gate knowing SR was new and unproven and would cost as much as a very nice car or two in order to set up at their facility.

And media... RED uses commodity media, about as commodity as you can get in the form of CF cards. Of course, we're limited to RED branded media with the exception of the LEXAR 8GB cards, but I don't see how this differs from most any other professional camera systems out there. P2, XDCam, DVCPRO, HDCAM, etc.. all initially used media that was only available through their respective manufacturer-owners. Even now, you can buy commodity HDCAM tape, but a portion of what you pay still goes back to Sony for the labeling.

I think there are several different mindsets on the forums here. I'm comparing the RED One to cinema cameras -- Arri, Aaton, etc... Or other "transitional" cameras like the Viper and Genesis. Just look at the workflows available for the Viper for the first three years of its life... Look at what is available now and compare to what we have with RED. Sorry, but RED wins hands down every time I think about shooting with the Viper.

Others are trying to compare RED to HD cameras or even prosumer HD camcorders and video-capable DSLRs. Well good for them... If they like "good enough" solutions and the easy workflows that come with them, then more power to them. Let them eat cake and shoot the festivities with line-skipping, long-GOP, aliased glory in 4:2:0 color space. I don't mean to bash these cameras for what they are -- I own a 7D myself and have rented a 5Dm2 on occasion and have shot tons and tons of home videos and even some pro stuff -- mostly shots to cut in with RED where I physically could not get the RED One into a tight space. That and I think the 5Dm2 and 7D are both worth their price alone as a stills camera, for which I do use them professionally. The video ability is a nice bonus, but not the ability I want for day to day professional video work.

 

[Mike Gross]

Uhhhhh... you seem to be proving my point, which was that Red had to do the software themselves in order to allow use of the camera. The last time I looked, Redcine, Red Alert, and Red Rushes were all Red's own products. Without them, at least for the first two years or so until the SDK was released, there was no way to use the files without using Red's own software.

Sony makes SR decks so you can play back what was recorded in camera. Not sure what the fuss is about Red needing to make software?

 

[M Most]

Sony makes SR decks so you can play back what was recorded in camera. Not sure what the fuss is about Red needing to make software?

There isn't any. Someone made the statement here that nobody has put the effort into post support software for their camera that Red has. I pointed out that one reason for that is because they had to. That's it. At that point, the conversation was taken completely out of context and into areas never intended.

 

[Frank Weeks]

Relax Mark. Most of us realize that you are one of the more knowledgeable ahd supportive people on this forum. We know the Red workflow path has been a bit slow in developing but for those that have worked in film. It's still a huge leap forward. Being in a small market it would take a couple days just to find out if I had usable footage. Red eliminated the wait and expense, It also forced me to learn the art of color correction. I have loved the transition from quality video and film professional to digital cinema artist. I have learned a great deal on Reduser. And with Epic and Divinci on the horizon the future is bright. Red has changed the way I do business. Let the good times roll.

 

[Peter Moretti]

Can some of the posters here PLEASE lay-off Mike? Holy cow, Red isn't your grandma. Someone should be allowed to say something that can be interpreted (rightly or wrongly) as crictical w/o a verbal call to blows.

It's a camera workflow, chill out.

Anyway, yes, Red's workflows are new. As was necessitated by the fact that the camera's format is new in every possible way: compression, frame size, color representation, etc.

So the Red workflow simply does not fit as easily into an existing post production chain as something like the F35 does.

But most Red users posting in this thread don't work in an industry entreched post production house w/ a lot of heavy iron equipment, so they don't see the difference.

So is the hassle worth the benefit? It sure looks that way. But to make it sound like there IS NO issue, makes no sense. The very existence of this thread shows that there are issues and confusion surrounding Red's workflows. And you don't need to go to SonySchool to lean how to post F23 footage. (But you will need a deck that costs a lot more than a RR card.)

The multitude of Red movie and tv projects in production shows that workflow issues are considered more than counterbalanced by image quality and the favorable ratio of value to price which Red offers.

So Red wins, and just don't take any of this personally.

 

[Adam Nielson]

CS5 is the way to go

CS5 is the way to go

Give Adobe CS5 a try .

Might be a bit late but CS5 is a great solution for the workflow problem.

 

[Petri Teittinen]

Might be a bit late but CS5 is a great solution for the workflow problem.

Mine stopped working today. Can't open any R3D files any more; crashes on import. D'oh.

 

[Blair S. Paulsen]

Virtual "Lab"

Virtual "Lab"

The mindset that I believe is most productive when dealing with the RedOne is the recognition of a virtual telecine step as part of the workflow. If you are coming from a 35mm film centric base, like Jeff Kilgroe, then eliminating the trip to the photochemical lab, and physical transportation of film cans, leaves you with most (let's save the DR, color fidelity arguments for another thread) of the benefits of 35mm film with less cost and easier logistics. You still need to "develop" the RAW into RGB (or YUV) images at some point. When you do that, and how many times (one light, offline, beauty pass, final, trim pass for other deliverables, etc) is up to you to choose based on your specific project.

As an early Red adopter I decided to go the Scratch route as they had a "digital process" solution that allowed me to do a one light color correction and transcode into one of the well supported formats Mike Most refers to (DPX, QT, etc). It was worth it to me to have Assimilate's engineering on my side rather than counting on RED's free tools which at various points in their development (no offense to Rob, Deanan and the rest of the Red Team) have been buggy. Its also worth noting that pre RedLine and pre ClipFinder there was little in the way of batch processing and trimming tools which are pretty essential to create more automated portions of a complete workflow. Scratch offers a plethora of excellent tools for media management, conform, multiple pre-built output nodes, etc that I still consider the best partner to Red post activities.

In any case, not every free tool from Red is going to seamlessly integrate into every post facility, NLE, etc on the planet - particularly older systems. If your point is that RED has a responsibility to proffer free post tools that can satisfy the entire motion media community from desktop hobbyist to top post houses then I give them props for getting as close as they did to that impossible threshold.

The good news is that the Madman and the RedTeam have managed to sell enough R1s that the mainstream post community is motivated to support r3d files better every day.

Note: The delay in releasing the SDK was not ideal, but there had to be a legitimate concern about putting out the SDK before they had locked down some basic file structure issues, etc. Should they have released an updated SDK every 90 days instead? I'm not an app developer or finishing box builder but that seems like it would have its own set of problems. Just sayin'

Cheers - #19

 

[Jeff Kilgroe]

Note: The delay in releasing the SDK was not ideal, but there had to be a legitimate concern about putting out the SDK before they had locked down some basic file structure issues, etc. Should they have released an updated SDK every 90 days instead? I'm not an app developer or finishing box builder but that seems like it would have its own set of problems. Just sayin'

The only thing worse than not releasing the SDK would have been releasing it early. It has already undergone changes in the way of adding support for RED Rocket and the new Color Science. And it's not always as easy as simply recompiling your software with the new SDK. The Rocket is a whole new piece of hardware with its own abilities to explore... The new color science has additional controls that the software can interface with (like FLUT).

If RED would have released the SDK early and kept adding new features every few weeks while undoubtedly making other changes, no one would have wanted to develop for the R3D format. You would've had a number of frustrated developers right away.