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 :)

Scratch by Assimilate System

S

sbrookens

Well-known member
Joined
Oct 24, 2007
Messages
55
Reaction score
0
Points
0
The question that has been posed to me from my superiors is, "if money was no object what type of finishing system should we be getting for Red ONE footage?" I am on a quest to figure out the answer to this question.

When we sat down to discuss this and I mentioned Scratch by Assimilate. I have looked at Silverado's website and I have seen two of their options. One for $35,000 and the other for $140,000. What is the difference between these two systems? What are the benefits of the $140,000 solution versus the $35,000 solution? Also, is Scratch the best option or is there a different solution that might be better?

Any thoughts? Thanks in advance!
 
one is scratchcine and one is scratch

you can lease a full scratch suite for under 1000 dollars per month too from silverado


scratch

scaffolds
dpx ingest
qt ingest

scratchcine cant do this things.
both with hardware.

scratch is great...
also is iridas speedgrade
and maybe premieres red plug in will be.
 
sbrookens said:
One for $35,000 and the other for $140,000. What is the difference between these two systems? What are the benefits of the $140,000 solution versus the $35,000 solution?
Click to expand...
Scratch is a colorcorrection system with some basic effects, scratch cine is a playout system with pretty basic colorcorrection.

sbrookens said:
Also, is Scratch the best option or is there a different solution that might be better?
Click to expand...
Scratch is a good system.
In its class, some main competitors are:
- iridas speedgrade and speedgrade XR (which handle redcode in internal versions, and it seems that they will deliver their *amazingly* fast and highquality raw to endusers soon).
- matrix chrome, which also handles redcode raw and has an integrated editing/nle.

Above those class systems systems, more powerful finishing systems do exist, some of then even less expensive than scratch.
one of the most powerful and popular ones is autdesks finishing system: discreet smoke. such systems add lots of features (3D geometry, compositing, complex audio mixdown, complex VFX, background rendering to offload the main station etc etc).
I dont know if discreet will release their native redcode support soon, last time i saw it in action it was still unofficial.
Quantel has also systems in this class - the Q series.

And finally, there are the "herosystems" - different from the systems above, they display and correct 4K in realtime. Some of these would be:
Filmlight baselight, a pretty powerful DI/Colorcorrectionsystem and
DVS clipster.
 
laguun said:
Scratch is a colorcorrection system with some basic effects, scratch cine is a playout system with pretty basic colorcorrection.
Click to expand...

SCRATCH is really not a "color correction system."

SCRATCH is a datacentric workflow system which provides a wide range of capabilities surrounding data workflow. One of those happens to be color correction and color finishing.

But, for instance... we have 9 SCRATCH systems at EFilm in Los Angeles. We are a fairly big part of the conform pipeline at EFilm. But not a single one of those systems does color work. They are data workflow and data conform managers.

At Rainmaker Animation in Vancouver, we have several SCRATCH systems. One of them is a DI system for doing realtime, live stereoscopic color correction. But the other systems are tools for performing either conform or dailies review functions - without any color tools.

It is a common miconception to think of SCRATCH as a "color correction system." It is much, much more. And out of our customer base, a fairly large percentage of them use SCRATCH because it is they believe it is the best overall conform system for datacentric workflow, or the best overall dailies and review system for large-scale VFX workflows, or the best overall stereoscopic workflow and review system.

What I believe separates us from other companies in the industry is this multi-faceted approach to post-production.

If you are looking at SCRATCH in strictly color terms, then your comments are one man's opinion.

But the questions was not "what is the best Color system we should be getting for RED ONE footage." It was "what is the best finishing system we should be getting for RED ONE footage."

And in that category, because of our realtime native R3D support, and because of the multifaceted nature of SCRATCH capabilities, I believe we are by far the best choice for RED ONE finishing workflow.

Best,

Lucas Wilson
--------------
Director, Business Development
ASSIMILATE, inc.
 
Hey luki,

scratch is a good system, but i certainly wouldnt call it a complete finishing system.

Its missing an nonlinear editor, the necessary powerful audio, complex titling & compositing & paint to compete with this such systems - it is a fine DI&workflow system, thats for sure, and i sure hope that scratch evoles in a great finishing system over time - but its not the typical one-stop-solution an online finishing system today is.

regarding the raw support - thats nothing unique to scratch. Most manufacturers have redcode raw support up and running, some slower, some faster than assimilate and didnt release yet mainly due to reds wish, and scratch will have competition from companies as adobe or autodesk - which is a good thing: competition is always good for products & customers.
 
Hi Laguun
I am finally leaning the Scratch way.
But do you have any first hand experience with Matrix and its RAW implementation. Or iridas?

I agree, Scratch misses a NLE, and for tape playout there are a few obstacles on a mixed resolution timeline. But all in all it works pretty well.

What systems are you using in house?

Just curious

Gunleik
 
laguun said:
...
regarding the raw support - thats nothing unique to scratch. Most manufacturers have redcode raw support up and running...
Click to expand...

This is untrue.

At this point, RED has seems to have decided against allowing REDCODE RAW to be demosiac'd (debayered) on the GPU -or using any alternative methods.

You may have seen "back room" demos at NAB, but as of today, I don't know of a single software vendor demonstrating high-quality grading and finishing of R3D clips at HD or 2K resolutions besides Scratch. And even then, you must be running on an "extended performance" workstation such as those provided by BOXX or Silverado.

This puts Iridas, Autodesk, Quantel, Adobe, Avid, DVS and pretty much everybody else out of the picture for now. They simply don't have the right type/mix of performance to allow working with the RED camera clips in realtime in HD & 2K.


Gunleik asked a simple question at the beginning of this thread - "if money was no object what type of finishing system should we be getting for Red ONE footage?"

The answer seems pretty straight forward. Scratch is the way to go.
 
BOXXLabs said:
This is untrue.
Click to expand...
Hey Boxxlabs, slow down a little.

BOXXLabs said:
At this point, RED has seems to have decided against allowing REDCODE RAW to be demosiac'd (debayered) on the GPU -or using any alternative methods.
You may have seen "back room" demos at NAB, but as of today, I don't know of a single software vendor demonstrating high-quality grading and finishing of R3D clips at HD or 2K resolutions besides Scratch.
Click to expand...
That you dont know of something doesnt mean its untrue.
In fact, there are several systems decoding redcode raw (CPU as well on GPU).

These are not only running backdoor, but in production. Several users of these systems post here, and respect their NDA.

No one of these manufacturers offers these solutions for sale yet, but programming a jpeg2000 debayer raw is no complex thing. cineform in fact demonstrated that. As example, Autodesk demonstrated lustre and other systems with redcode just recently in Munich - i however wont speak of the other solutions which are in fuill production use, but rest assured, many of their users are reading here.

If you want to see GPU in action with rawdata from many other cameras, just have a look at XR.

BOXXLabs said:
And even then, you must be running on an "extended performance" workstation such as those provided by BOXX or Silverado.
Click to expand...
No.

Silverado and Boxx do nothing else than any standard PC manufacturer or employed engineer: Buy motherboards, CPU, RAM, raid & case, integrate. Or does boxxlabs manufacture any hardware?

BOXXLabs said:
This puts Iridas, Autodesk, Quantel, Adobe, Avid, DVS and pretty much everybody else out of the picture for now. They simply don't have the right type/mix of performance to allow working with the RED camera clips in realtime in HD & 2K.
Click to expand...
That is wrong.
The hardware a boxx system uses (as example: TYAN S2696 & nvidia quadro on the boxx 8300, TYAN S2915 on the boxx 7500, TYAN S2895 on the boxx 7400, iirc?) are exactly the same components in the $$$-$$$$ scale as in any other workstation.

In fact, dedicated 4K realtime hardware as in DVS Clipsters or clusters as in Baselights Filmlight are much more powerful.

The decision to not allow full access to redcode raw has nothing to do with performance offered by standard nvidia/intel/tyan $$$$ combos (which is exactly what boxx or silverado offer), but only with the company strategy and politics of red. one might agree or disagree with those decisions, but its only the go ahead of red which makes the difference, not the hardware.

BOXXLabs said:
Gunleik asked a simple question at the beginning of this thread - "if money was no object what type of finishing system should we be getting for Red ONE footage?"
Click to expand...
wrong - read the post: sbrookens asked that question, not gunleik. furthermore, i wouldnt recommend scratch native redcode for the full 4K - but its fine for the 2K realtime. For 4K with FX etc, there are other and IMHO more powerful workflows available.

BOXXLabs said:
The answer seems pretty straight forward. Scratch is the way to go.
Click to expand...
while scratch is a good choice, there are other powerful options and its worth to speak en detail with manufacturers as chrome, autodesk/discreet, iridas, dvs or yes, adobe.

Not everyone prefers a native scratch over a discreet smoke or a filmlight baselight - just look at the sales figures.

another important recommendation to anyone: wait until IBC before buying.

Your mileage might vary, however.
 
Hi Laguun.

THAT is always true. (the milage thingy...)
there's a couple of catches with the "wait untill the next big trade show" attitude - even though I CAN wait for IBC :)

@ some point you have waited beyond your customers patience.

Scratch is way cool. Either the software or the hardware is (too) expensive when you put on those comparative glasses. The problem is that so far: Scratch is the only system I have really seen working. If I am to buy something else, I need to be educatedin a believable and very practical way. So far only assimilate has done that to me, and in a pretty (not very...) convincing way - I have to say.

The hardware of the Scratch systems I've seen so far hasn't been too impressive really, I agree. It's a bit of mumbo-jumbo. And it is annoying listening to... BUT it has been working. I like that part :)

So what about this post?
Scratch is cool!
I wait for a relevant quote.
If it is good, I think I may take it. After IBC. If nothing else shows up there...

Sorta...
 
laguun said:
... Not everyone prefers a native scratch over a discreet smoke or a filmlight baselight - just look at the sales figures.
Click to expand...

Laguun, Your opinion is clear - you think there are other workflows that are just as good and valid. And that's fine - your tools and your workflow obviously work for you.

But facts are facts.

SCRATCH is the only high-end native R3D workflow right now for current, B16 cameras and firmware. Anything else is just smoke and mirrors. End of discussion.

And... you have no idea what our sales figures are. ;)

Lucas
 
Sorry, Lukas, but that is sorta misleading, too. (Ok, as a potential customer I shouldn't dig into this) but there is quite a few circumstances where Scratch isn't a R3D workflow, but more like a "transcode to dpx and do stuff like with filmscans" workflow. Which is not bad, but definitely not native RED, if I haven't misunderstood a lot of things (which I would actually have preffered to have...)
 
gunleik said:
Sorry, Lukas, but that is sorta misleading, too. (Ok, as a potential customer I shouldn't dig into this) but there is quite a few circumstances where Scratch isn't a R3D workflow, but more like a "transcode to dpx and do stuff like with filmscans" workflow. Which is not bad, but definitely not native RED, if I haven't misunderstood a lot of things (which I would actually have preffered to have...)
Click to expand...

Can you explain further? We have a ton of customers doing native R3D workflow for dailies, conform, and finishing. Are you referring to a mixed-media workflow?

Lucas
 
digitalfx said:
both Quantel and Discreet can play real 2K in real time.
Click to expand...

Neither can debayer redcode at 2k in reatime. All systems mentioned can play back 2k in realtime, but that's from 10bit DPX files.

And I have yet to see a convincing argument for debayering prior to grading. IMHO, it is always a mistake. Until systems are able to play back 16 bit files in realtime it will continue to not make sense to build the image prior to grading.

cheers,

JT
 
gunleik said:
Sorry, Lukas, but that is sorta misleading, too. (Ok, as a potential customer I shouldn't dig into this) but there is quite a few circumstances where Scratch isn't a R3D workflow, but more like a "transcode to dpx and do stuff like with filmscans" workflow. Which is not bad, but definitely not native RED, if I haven't misunderstood a lot of things (which I would actually have preffered to have...)
Click to expand...

I assume you're refering to tracking & audio? Those are high on the developers priority list, and I'm sure will come down the pipe asap. One thing Assimilate is not shy about, that's releasing updates.


cheers,

JT
 
For the record (@ 2K), a simple (ie, low-quality debayer) is trivial to do, and easily achievable in "real-time" speeds. The decoding of JPEG2000 data is inherently time-consuming. "real-time" decoding in software is available, but the codecs are few and far between (and expensive). As with everything, things will change over the next 2-4 years, and soon JPEG2k will be trivial for the average processor. Doing a GPU-based "average-quality" debayer (for pre-grade finishing) is certainly within the real of today's machines.
 
You must log in or register to reply here.
Back
Top