E2E Production workflow diagram

[david farland]

http://www.reduser.net/forum/uploaded/48_1176397848.jpg

Hi all,

Thought I’d publish a basic workflow to see what post tools are available, how much they are, and begin to define workflows options.

I know I’ve a missed lots of stuff like sound, 3d tracking, colour calibration, publishing etc but I’m hoping others can correct me here.

A couple of things….the data rates will change but figured I’d put down something.

The initial ingest is obviously via Redcine or any offline that can read QT wrapped redcode I guess.
The offline edit (low rez) is dependant on the online system...i.e. what edl/xml, etc standards does it ingest but it also can be same system.

For onlining, Redcine via RPL or a native online system that produces the highest rez files your system can handle.

On the non RED post workflow, BMD, AJA, Bluefish etc HD-SDI boards are for the uncompressed rgb/yuv i/o & the uncompressed 10Gb fibre raid is to be announced. Also RED will need something to parse the uncompressed raw.


3D Modeling – don’t ask me…! Comparisons are here: http://wiki.cgsociety.org/index.php/Comparison_of_3d_tools (http://wiki.cgsociety.org/index.php/Comparison_of_3d_tools)

Effects /Compositing - most of the companies said their file ingests are resolution independent and 32bit float…umm.
I did my best on the I/O figures but some of them are wrong.

Online Grading – speed, file size, ease of use is everything!

Just hearing how the online companies designed their online systems was a treat. Really was fasinating! Hopefully others who really know these systems can help de-mystify these workflows/architectures.

Anything I can do to help, I will.

Hope you can help and provide more information/corrections!

Dave,

[Chris Gearhart]

That's a nice chart, David. I appreciate the software list with bit depths and resolutions.

[Gianny Trutmann]

Thanks man

[Alexis Hanawalt]

I think the 3D programs are about 1/2 those prices now - Lightwave 3D is around $1k

[GlennChan]

With some online systems, the conform process may vary from system to system and you may need to do some time-consuming things to conform a project correctly.

A- Only a few types of effects in the offline will carry over. With anything else, you need to manually re-create it and/or pay careful attention.

With Final Touch (old versions at least), you need to make sure everything is cuts and dissolves. It won't handle speed changes. So you need to scrub your project of everything but cuts and dissolves, grade in FT, and then bring that back into FCP where you re-apply all your effects.

B- Some systems work via an EDL + ingest via SDI. (Unfortunately Red doesn't seem like it has a deck that reads Recode natively and outputs SDI.)

EDLs support cuts, dissolves, and speed changes (speed ramps need to be manually recreated).

When you capture off tape, sometimes things will come in a frame or more off. You need to manually compare the ingested footage with an offline cutting copy, and manually slip everything into place. It's not that time-consuming once you get fast at it, but it's still a PITA.

This is not quite as painful as Final Touch's data workflow from what I remember of FT (because FT has data incompatibilities and lots of bugs; and media mangler if you want the EDL route, which has advantages). In my cynical opinion, Final Touch is why you should be cynical about data workflows. Theoretically XML data workflow is a good idea... but in practice it hasn't worked that well.

2- Ok, on the other hand these problems shouldn't be hard to solve with a little creativity. One of the problems with FT I think is that people were left on their own to figure out a workflow that works for them... FT didn't exactly have very good documentation.

In the worst case scenario, suppose that the online system only works well with SDI ingest. If Redcine can ingest an EDL and spit out two movie files, then you could load that up into some sort of virtual VTR program (I believe Aja and Blackmagic offer such utilities?). By emulating a VTR, you would slide into existing workflows and you also have the advantage of no tape shuttling + reel changes. Tape shuttling + reel changes do suck up time.

3- To add some uninformed speculation (gotta love that), perhaps Redcode or Red raw will have a workflow similar to Cineform RAW?

You can edit your Red footage directly at low quality. When you get picture lock, you change the codec settings to give you a high quality debayer (and perhaps even do simple grading on each shot in Redcine). And from there you just render your whole project, without having to conform your project and re-create your effects.

The only downside with this is that FCP and other desktop NLEs are slow for online work.

4-
I guess there will be more info at NAB.

[Gopher77]

The last ime I checked FCP is 8bit for 4:2:2, it is only 10 bit for uncompressed 4:4:4 Also I thought, haven't checked this that Nitris is 14bit.

[donatello b]

FCP- i thought it was 8 bit for all RGB ...and 10 bit for YUV ??

[David Newman]

Regarding the CineForm entry. Prospect 2K will work at 4K for source data, for oversampling, and re-framing, then master out 2K. For 4K master we enable that also, but through After Effects not Premiere. Complete you edit in Premiere Pro real-time at 2K, with mixed source resolution (4K,3K [D20], 2K, 1080p, 720p, etc.) sources, then import the Premiere project into AfterEffects Pro for 4K DPX output (if you really need that.)

[Gopher77]

FCP- i thought it was 8 bit for all RGB ...and 10 bit for YUV ??

It maybe 10git YUV, What I posted was from thier web site I had read when I was considering moving to mac and FCP.

David,
I was wondering about CS3, have you guys been working with it yet? How is it looking as a possible editor for red footage?

Gopher

[David Newman]

CS3 is working well. We will be showing CS3 with 4K footage run real-time on standard hardware (Intel dual proc core duo) in our Booth. CS3 should will work fine with Red footage, but even better (I predict) running a CineForm Prospect 2K engine (although maybe will supprise us.)

[Anders Holck]

FCP- i thought it was 8 bit for all RGB ...and 10 bit for YUV ??

FCP rendering engine is max. 32 bit float for YUV, and 8 bit for RGB.

[Don Woods]

Good work very nice thank you..

[david farland]

Thanks guys......very much appreciate the corrections/feedback
For predominantly timeline edit systems I left out image file ingest rates. The other rate I could have put is internal process rate but that was afterwards and I thought some of them changed depending on type of file but wasn't sure how and to find out that with a reasonable degree of accuracy for most of them meant no family time!
so for FCP max rates are:
Video i/o = 10bit 4:2:2 HD and 8bit 4:4:4 (RGB)HD
Tape i/o = 10bit 4:2:2 HD

David, saw you had 4K up your sleave for special customers and price?
I think I get it....using your codec in pp2 2k project doesn't destroy 4K source footage and when you import into AE it'll render 2K & 4K footage at 4K. sorry I don't know enough about cineform. In this scenario is the footage that gets rendered to say 10bit dpx 4K, cineform versions of the original 2K & 4K footage, I can hear the sign....

I'll redo all the 3D pricing in the next couple of days.
Alexis, if your up on 3d package have you seen this list...use Babelfish
http://www.cgtalk.ru/~lynx/graphics.html
It's in russian which I'm in the process of cleaning up. forget the saint that did it...but will duly credit


thanks again,

[JD Holloway]

Very nice work!
Now I have to figure out what it all means.
I have to find a path of least resistance....

[tj williams]

The B4 input is thru an adapter to a windowed sensor which is almost S16 size. As far as I know there is only one size window. So windowed output is of only one type?

Wonderful information on the digital workflow details, bit depths etc.etc... Thanks

[Gavin Greenwalt]

Your diagram is a little bit off. The 3D Renders go into the compositors, not the online graders. Even a fully CG movie will first pass through a comp workstation before hitting the online edit and conform. Also I'm not entirely certain your use of the word "Conform" is accurate in the diagram. It is more akin to "Printing" or "Telecine" than conforming. Conforming is more of an editorial task.

I would love it if REDCine could actually conform, or at least get really close, but in this case all it'll do is take the EDL and export an online debayered copy to be conformed by another application.

And yes except for the FFI machines they all support 32bit-float images.

[david farland]

Glenn,
Thanks for that info. I’m not an editor thou I’ve done a little avid/ppro stuff so thanks for the reality check on final touch/fcp.
I suppose a thing to look at is online ingest and related conform problems on the various online systems. Though I’m thinking the guys at RED working on their RPL should have experienced all this pain with different export files. I know what you mean regards checking the offline cut with the online cut. I forget who I was talking too, one of the main online player and made a comment they had no video i/o…few of them don’t, and he mentioned they were going to install it just for that purpose.As I say I’m no editor. Look at me as the office temp for this stuff!

With regards to just doing offline basics and ‘NO RENDER’, it’d be interesting to hear what people who edit all the time think. Thou I expect most editors get into a routine and accept necessary evils.

The other thing I noticed is those big online systems are full of stuff and have had a lot of R&D investment gone into them.
I liked Baselight kit where you have a bunch of quad core processors with their own raids, each handling a slice of the frame and then recombining in some serious GPU. So in the Baselight 8 system for 4K each unit handled a 500x2150pixel array. Could also be done on the Baselight one but much slower.

A few of them did 2K corrections and rendered out via xml to 4K farm. Didn’t think there was anything really wrong with this although must be problems that I just wouldn’t see i.e. importing 4K composites etc. Most people said for the bigger systems you were paying for ease and lack of work arounds. To me I’d rather buy a lighting truck than a 4K realtime online system. Or both!!
I expect those online systems will radically come down over the next two years.
Maybe the cheapo solution will be 2K proxy out, GPU based, 16 core monster with a cheap panel.

I’ll tidy up the diag, add NAB announcements and see. Be also nice to have a list of camera accessories but for now the workflow solutions should keep me busy. thanks Gavin, I'll correct that..can you tell me more? and yes TJ , I was hopping not to add another asterisk but I will!

Thanks again

[laguun]

hello david, nice project you are started, however there are ugly mistakes in it.

some correction after a very quick look.

combustion,
should be together with its autodesk sisters & broters.
tape 2K RGB 12bit
video 2K RGB 12bit.
and quite important, file i/o is not only 16 bit, but floating point precision.

also i would add toxik.

we will suport you by gathering some more information, if you would like and we find a little time between the projects.

[david farland]

....however there are ugly mistakes in it.


Thanks Laguun,

I suspect you're referring to my avatar here ....I'll change it.

Thanks for the tech info.

Yes I don't appreciate the difference between 16 bit interger and 32 bit float and couldn't get precise specs on all so I dumbed it down so as to not say a product was higher spec'd than it was. Will change.

Forget the docs I was reading on Combustion but one Autodesk doc here (http://usa.autodesk.com/adsk/servlet/index?siteID=123112&id=5562401) said :


- DV capture and output through OHCI .rewire.
- Capture frames or sequences directly with select QuickTime-compatible
capture devices (including 10-bit QuickTime).



So you're saying Combustion does 12bit RGB out a hardware tape device...yeah?


Will add Toxic...I think it's about $9k.


Thanks for your support

...Cheers

[laguun]

Forget the docs I was reading on Combustion but one Autodesk doc here (http://usa.autodesk.com/adsk/servlet/index?siteID=123112&id=5562401) said :


- DV capture and output through OHCI .rewire.
- Capture frames or sequences directly with select QuickTime-compatible
capture devices (including 10-bit QuickTime).

this is true for the "naked" combustion.
the specs i have mentioned are specs combustion archieves with 3hrd party hardwareboards, without them the techspecs from the website are valid.




So you're saying Combustion does 12bit RGB out a hardware tape device...yeah?

its the aja, bluefish and blackmagic products who provide the i/o for combustion and bitdepth, some require external software to be launched.
edl processing inside of combustion is basicly non-existing, so you usually add a 3hrd part software/hardware.

the boards we are using here only do 10bit, at ibc however i spoke with aja employees and they said that they could do 12bit and publish that information on their website as well.

remains to be checked if they are referring to analog/digital 12bit only, because i wouldn´t understand how they do 12 bit i/o via duallinked hd-sdi, except they would use a propietary standard.




Will add Toxic...I think it's about $9k.

9k? i am not sure, could be that 9k was the 3 user license.

finally, one could add to the list
- hash animation master (3d animation secret tip since a long time)
- premiere w/o cineform, decklink/aja might be better suited for red
- sony vegas might be interesting as NLE as well

p.s. sorry for the "ugly mistake" comment, its a pretty common german expression which maybe sounds a bit to tough when used in english - its just used to differentiate a tolerable small glitch from something which needs correction in its originating language.

[edgar01]

Dear David

1- what is the basic Red digital media option I need in order to obtain the best quality (S35mm 2540p RAW Frame Rate) whit my red one camera?

2- Explain me why. Im so beginner in those matters

3- I can Buy cheapper digital media option whit others party or I have to buy Red One (Red One be expecial configuration so no others can sell you those spec?

Regards

Edgar

[Ace]

Isnt maya 2k?

[Antoine Baumann]

Isnt maya 2k?

No not Maya Unlimited, only Maya complete (or something like that), which if I remember well does not have Maya Live, Fur, Cloth, and maybe Liquid simulation or other restriction... go Blender it comes free with all the options :-))

antoine.

[Gavin Greenwalt]

...go Blender it comes free with all the options :-))

Yeah well that's one area where you definitely get what you pay for.

P.S. I don't believe Maya Unlimited has liquid sim either. It has "fluid sim" which is a different beast altogether.

[Rocket]

Maya will render 4K no problem, so will all the others, unless I am misunderstanding you. You could render at 64K if you had the processing power and time, the output is entirely resolution independent.

Hmm, have to edit my post now. I'm an idiot. I just realised you were talking about price.

[Antoine Baumann]

P.S. I don't believe Maya Unlimited has liquid sim either. It has "fluid sim" which is a different beast altogether.

Thanks for the correction, yes I was talking about fluid sim, not liquid.

antoine.

[M Most]

Your pricing numbers are very misleading. All desktop programs require a computer to host them, whereas systems products are sold turnkey. So while the cost of, say, Avid DS, really is about $100K, that includes the software, the host computer, monitoring, interfacing, and storage. Final Cut, on the other hand, is $1.3K for software only. You still need the host computer, monitoring, interfacing, and storage, making the true cost of a system competitive with your $100K DS more like $15-20K, not $1.3. Granted, that's cheaper than the Avid product, but it's not $1300. Same with Shake, Digital Fusion, Nuke, Combustion, and all the other desktop programs you're quoted pricing on. You also quote Assimilate Scratch based on the cost of a turnkey system, which would be fair if you quoted everything else the same way - and by the way, Scratch is a software product and is often purchased that way. If you want to do a fair cost comparison, add in the cost of the system to run the software only products on, and quote those systems with the same equipment the turnkey systems (i.e., all Avid products other than Media Composer Software Only, Autodesk systems products like Inferno, Smoke, Flame, and Lustre, all DaVinci products, and Baselight) have. The only area that has pricing quoted that reflects reality is the 3D software area, since none of these products are sold in turnkey systems. However, all of them do require host computers with proper graphics capabilities and monitoring, as well as significant storage in order to perform well.

If you want to make a pricing chart that's accurate, quote prices on all of these products that include the necessary hardware to run them on.

[Lucas Wilson]

Hi David,

First of all... I'll be at SMPTE Australia this week. If you live in Sydney - would be great to meet you! Contact me offlist at lucas@assimilateinc.com so we can work it out.

I applaud your chart and the time you took to make it. Good stuff. I want to point out a few things that may help:

1) I'm not exactly sure what you're trying to share with the "Image I/O" thing. SCRATCH can support 8, 10, 12, and 16-bit files, and processes everything internally at 32-bit float.

2) As Mike also pointed out - the 90K number can be correct for a turnkey system, but SCRATCH is frequently sold as software only. And as modular software, a customer can purchase only those modules that are necessary for the task at hand. So... list/retail prices for SCRATCH can run from $5K at the minimum to $65K at the maximum. But also keep in mind that the "maximum" is something that very few people buy. It's sort of like buying a car with ALL the options. You almost never need all the options...

3) In your workflow chart, this is correct... if you're using any online/color system except for SCRATCH. With SCRATCH - REDCODE is supported natively, so the Online Ingest/Online Grading step is combined into one box. Load an EDL... conform to the original REDCODE files... and start grading. You do notate that with a "*" designation at the bottom of the chart... just wanted to highlight it as it is a huge workflow difference - especially with longform projects.

Cheers,

Lucas
-----
ASSIMILATE, Inc.
LA, CA, USA

[Rocket]

I was also going to say that figure for Scratch was very configuration dependent, as well as location dependent. Our turnkey Scratch configured with 8TB storage and data subsystem for two streams of uncompressed 2K comes in at just over $200,000 here in South Africa but that includes a Panasonic 2K projector (which we are now swapping out for a Christie CP2000, and a 50" Plasma). We have now decided to go for BrightDrive 4K Fibre Channel to give us 4K in real time through Scratch, and I'm still waiting for that quote.

It's going to leave Baselight8 in the dust, and even with the 4K drive subsystem it will cost a whole lot less money.

I'll stop now, I'm get a bit excited when it comes to Scratch, almost as excited as with Red. The two of them together... phew, I'll stop now. I need to calm down a bit.

[Bruce Allen]

David

This is a cool graph but also a little misleading. It kinda encourages you to buy into the Red 4K hype. As you know I am a fan of acquiring 4K but finishing 2K or HD (and only doing a full 4K finish if your films gets bought by a major studio). In 3 years' time things may have changed, but for now 4K VFX work, etc is just stupid.

Firstly:

1. If you are integrating the 3D with your film and you're doing any kind of moving of the camera besides simple panning, you need to include a 3D tracker into your budget.

2. The whole budget thing is kinda crazy anyway. The major cost is ALWAYS human talent. For that reason, you should definitely look at whatever solution allows for "good enough" shots to be done at a reasonable. For me that means a good, responsive system, and being sensible and doing things at 2K or HD resolution.

I have a bunch of compositor / vfx producer friends in LA. Between them, they have worked on all of the major VFX blockbusters that came out this season. Every single one of them would burst out laughing if you told them to do the majority of their work at 4K. And they are working at big Hollywood effects houses with stratospheric budgets.

Personally, I think there is way too much 4K measurbating going on at Reduser and a lot of people trying to ride this stupid hype. The idea that a system (Flame, etc) is "inferior" because it can't render out at 4K and high bit depth (which is implied in your graph) is a total joke.

The key to all of these is workflow. I use After Effects with Magic Duck to do color correction all of the time, but I wouldn't seriously consider using it as my sole grading tool for a feature because of its inability to conform changes, lack of responsivity, etc. Unless every shot was VFX. Even then you'd want to pay a scripting programmer to set up a workflow that was similar to what current grading systems already give you. And you'd still need a bunch of $500+ per day working in AE to keep everything consistent from shot to shot.

4K Hub, let's be honest here, the fact that you're setting up a Scratch system in SA is really cool - last time I checked, we definitely needed more competition to keep the Sasani conglomerate on their toes - but the whole 4K thing is really a marketing ploy.

I'm sure you'll be able go and use it to sell idiot producers on using your system because 4K is the next stupid buzzword. But if you want to work with someone who knows what they're doing, you should instead sell yourselves on having talented colorists / finishers, a strong track record and a good price and workflow. Of course, Scratch has an excellent Red workflow - I would hype that up instead of the whole 4K thing if you want to be taken seriously.

Bruce Allen
www.boacinema.com

[Rocket]

4K Hub, let's be honest here, the fact that you're setting up a Scratch system in SA is really cool - last time I checked, we definitely needed more competition to keep the Sasani conglomerate on their toes - but the whole 4K thing is really a marketing ploy.

I'm sure you'll be able go and use it to sell idiot producers on using your system because 4K is the next stupid buzzword. But if you want to work with someone who knows what they're doing, you should instead sell yourselves on having talented colorists / finishers, a strong track record and a good price and workflow. Of course, Scratch has an excellent Red workflow - I would hype that up instead of the whole 4K thing if you want to be taken seriously.

Bruce, you are entitled to your opinion, and I also acknowledge your experience and the experiences of your industry contacts. I suppose some of us here assume that it goes without saying that talent in the end is what is going to make a great feature, commercial or whatever. However, this is also a technical forum and so it's no surprise that it's the tools that get the focus.

Sasani is big, they have Film Lab, Video Lab, MCC, ZSE... practically every big name in post in the country. We are a small no-name that is coming up behind them with a purely digital cinema focus, for 2K and up, that's specifically targeting productions that would normally be shot 35mm, be scanned to 2K for DI and either printed back to film or output to SD and HD for broadcast. We will be offering them a no-compromise 2K, 4K or even 4.5K solution from lens to digital cinema server, to HD or whatever format they need it in, and in S.A. we can do it cheaper than almost anywhere else, L.A. or London. What's stupid about that?

I am surprised that you don't assign more importance to the 4K workflow in discussion because more and more cinemas in North America and Europe are jumping on the 4K projection bandwagon. What would be stupid is to DI and finish a film at 2K to send out to cinemas for projection at 4K.

[Fergus Meiklejohn]

If I film with RED I want to finish in 4K, no question. I understand that VFX is not ready. That's a problem that will be solved...
I know one of the London FX houses has started to do everything 4K, they're showing off, but it has started to happen.

[Bruce Allen]

Ah man, I'm just trying to call things as I see them. I think it's wonderful that new pro-Red, digital finishing companies are starting up, and I don't mind if they use the "we are better, we do 4K" line to counter the FUD that the traditional finishing companies are going to throw at them to try to scare their clients away. I totally support you guys and think it's great that you're doing this. That said...


We will be offering them a no-compromise 2K, 4K or even 4.5K solution from lens to digital cinema server, to HD or whatever format they need it in, and in S.A. we can do it cheaper than almost anywhere else, L.A. or London.

Your solution may offer no compromise on the technical side, but that isn't what's important. Let's be honest here - the quality of the finishing solution is ultimately dependent on human talent and machine responsiveness. You have said a lot about what resolution you are outputting at but very little about who your colorists, finishing experts, etc, are.

To me, the no-compromise solution is to work with an experienced, talented and communicative finishing team, even if they are working at 2K. The human talent should be the bulk of your costs. SA has some great colorists but very few people with experience finishing 4K, for example. If you are doing it a lot cheaper than LA or London, it is because you are cheaping out by not paying to bring in some experienced talent. I'm sure your team is great and can't wait to knock the stuffing out of the finishing houses around the globe but until they have stunned the world in a manner similar to Weta, you can't sell yourselves as no-compromise. And remember that Weta went to great pains to hire the best from around the world to get themselves going.


What's stupid about that?

It's stupid to trade 4x the rendering, processing and storage requirements for a small return. I'd rather put the budget into something that needs it more.

Look, if the big boys of VFX are not doing everything at 4K right now it is because they are not quite there yet cost / benefit wise. They'd rather spend their render cycles / drive space / etc on more important things - better crowd simulations, better fluids, more render passes, more versions of a shot, etc.. Currently, people do the majority of their work at 2K and will only go to 4K for extreme problem shots.



I am surprised that you don't assign more importance to the 4K workflow in discussion because more and more cinemas in North America and Europe are jumping on the 4K projection bandwagon. What would be stupid is to DI and finish a film at 2K to send out to cinemas for projection at 4K.

1. Wow, that'd be almost as stupid as showing a 35mm feature at an IMAX. Oh wait, they do that!

2. Personally I think a good 2K film finish such as Master and Commander will still look like it deserves its cinematography Oscar when projected at 4K. 2K digital projection often sucked for reasons besides resolving power - horrible screen door effect, bad color, etc. A good up-res of a 2K film to 4K will look smooth and good on a 4K projector.

3. Also, what kind of film? Studio? Indie? Do you have a distribution deal? How much is VFX and how much is just straight color? That is what should determine HD vs 2K vs 4K. I have seen 4K projection and prefer it. But there are a lot of cases where it just doesn't make sense yet.

Please show me one person on this forum advocating end-to-end 4K finishes for everyone that has actually DONE an end-to-end 4K finish.

I have done some stuff at 4K (recently, the teaser for Ocean's 13). And I (and my VFX friends) say it is probably not worth it yet.

Again, I'm really glad that you guys are doing this, just don't like the 4K emphasis over all of the other great things you should be pushing - great Scratch Red workflow, etc. Anyway, after this post I think I totally owe you guys many beers. I'll be in SA in September - will see if I can track you guys down.

Bruce Allen
www.boacinema.com

[Rocket]

Again, I'm really glad that you guys are doing this, just don't like the 4K emphasis over all of the other great things you should be pushing - great Scratch Red workflow, etc. Anyway, after this post I think I totally owe you guys many beers. I'll be in SA in September - will see if I can track you guys down.

Beer always sounds good!

Experienced colourists and vfx artists are hard to come by in S.A., it's going to involve a lot of head hunting and wooing some top guys away from other facilities, not only in S.A. but elsewhere also. It's not going to be easy that's for sure, and it's a long road ahead. But it's not only capital that is needed, it's also at least 12 months running costs up front, and that includes those salaries.

A 4K E2E workflow is not only possible, it doesn't have to be that expensive either and it will always be capable of handling lower than 4K media all the way down the line. I think we have a big advantage in that we are building and fitting out a facility from scratch. We are not paying off existing equipment that we need to suck a few more years of paying jobs through.

Our biggest issue is political and social instability, this is the one thing that can still prevent us from setting up here in S.A. at all.

It's always good to have a dose of reality injected into a discussion like this, there's nothing wrong with that. I just think there is a lot of interest brewing under the surface for an entire 4K post solution that doesn't break the bank, and I take my hat off to companies like Red and Assimilate that place thier products in a truely accessible price range.

It's early days for sure, but the point at which that cost/benifit ratio plus increased demand will swing the balance in our favour (and others investing in 4K infrastructure now) will come.

There is currently a window of opportunity, that's all, and it won't last very long once things really start to gain momentum.

In the end, when the doors swing open and the tools are there to be used, it's talent and marketing that will make or break us. That's business.

[Fergus Meiklejohn]

Absolutely right Bruce, even if Apple's Color supported RED 4K and I could run the thing in real time on my Mac Pro, I'd still want the best colourist in the world to be working it

[donatello b]

"I'd still want the best colourist in the world to be working it"

the word above is "want" - we all WANT this/that but in the end it will come down to $$$
many times the "BEST" work at the houses that have a extreme hr rate ..
so in the end you go with the BEST that you can afford/budget ...

[david farland]

Edgar,
Depends on your quality /cost requirements.
Do an offline cut using any basis NLE, then when you’re ready to conform you have a hundred choices from completely raw to very compressed…..from using your offline package costing $1K for onlining, to online grading systems pushing $700K, and an array of other 3D/tracking/compositing products you may use along the way.

One reason I started the diagram was because I figured I could map out what was commonly available and included cost because a lot of people had no idea of pricing or even what software were available. The cost gave an ‘indicator’ of what speed/quality level each package was capable of. The costs of larger systems can vary by $100K’s, much more than the lower end systems of which more people are familiar. I should have included the asterisk meant hardware included.

The offline/online/grading software packages are containers of modules/codecs that are bouncing video footage around at different quality levels & speed. The quality element is 'probably' determined by what are the ingest codecs /transformations, native processing formats/resolutions/maths which goes on inside the app. The speed element is determined by fast & distributed hardware, optimised/clever software or firmware. In regards to mapping speed & quality the only real way would be to benchmark them at different codecs/resolutions because lots of companies would say their package could do this at higher resolutions, but you knew their system would be barely alive for any ‘real time’ functions or take weeks for renders. I started to map out what packages used what codecs/bit resolutions but stopped because I had no idea what native formats /transformations were used in each package. As mentioned, to do a proper comparison you’d need to benchmark each system using test files and defined tasks at various quality levels. Most of the bigger packages had a SDI tape interface. Image I/O means sequence image files i.e. Tiff, DPX.

Thanks for your feedback. Happy to consider rewrites.

Cheers,

Dave,


Luki, can’t make SMPTE, but will be at Ted’s ACS demo on Friday night.

Mike, you’re quite correct, I could add at $2-20K statement with kit breakdown for lower end. Still need the breakdown for higher end kit upgrades, i.e. Stone, Unity. Would also need 3D additional hardware/software options.

Bruce, sorry if it came across fan boy and naïve, but I started the diagram so others could discuss the costly end of town products… I’m hoping in a couple of years there’ll be applications indies can buy, which will be optimised to use 8 CPU’s, multiple GPU’s, multiple disk controller cards for screaming raids, and fibre comms along with cheaper grading panels, projectors etc, that are doing what today’s higher end systems are doing now.

[Lucas Wilson]

Luki, can’t make SMPTE, but will be at Ted’s ACS demo on Friday night.



Hi David (and anybody else from Oz paying attention here... :) )

I am on the Future Reality stand at SMPTE, and will also be at the ACS presentation on Friday evening - so we can meet there! Ted Schilowitz is also here at the Apple stand with a camera...

Lucas
-----
ASSIMILATE, Inc.
LA, CA, USA