Redcine 576p?

Yes, 576p is in there also. Basically, no real size limits, withing reason. That document was drawn up rather quickly in the run up to IBC, so was a touch incomplete.

Graeme
 
576p

576p

Graeme_Nattress said:
Yes, 576p is in there also. Basically, no real size limits, withing reason. That document was drawn up rather quickly in the run up to IBC, so was a touch incomplete.

Graeme
Click to expand...

Graeme, nice to have an ally for those of us dragging our feet in PAL country:)
One question: Having seen first hand that downconversion is no simple feat and often best delegated to serious hardware, I'm just wondering if you'll be including special trickery algorithms of your own in the REDCINE software and.... oh dear.. dare I ask: Possibly GPU assisted ones at that ?

Steen Dongo

PICTUREWISE FILM
DENMARK
#221 (I think)
 
I predict that in 8 -12 months these forums will be filled with questions like: "Why does my Red footage look so bad when I downconvert? Why does it moire and shimmer? Why is it so soft? Why is it so sharp?" In most cases it will because of the downconversion procedure used - and poor Red will have to deal with the bad buzz (rumor-mill buzz that is, not video buzz...) Given that 90+++% of all initial footage will be viewed and scrutinized as some form of downconverted image, I think Red Company would do well to educate their clients on the ins and outs of downconversion. In particular, I think that REDCINE software should include idiot-proof terminology like "low quality" filtering for (fast) offline editing, ftp posting, etc and "high quality" filtering for (slow) online, vfx, etc.

On discreet Inferno, for example, a sample factor of 16 or higher should be used to downconvert 2K footage to NTSC, with a softness of 0.8 or so to avoid the blurries. Its also good to customize your settings based on the specific footage - a hi-con logo on a black background should be treated differently than soft footage of a girl's face. The problems compound exponentially when dealing with interlaced footage, video vari-speeds, etc. Our clients are always asking why their HD spots look so bad when they are downconverted to NTSC - again this is usually because the downconversion guy just used the basic-default settings. This happens ALL THE TIME by some very pro-folks who should know better .

On a related topic, I don't believe that down-conversion will always result in a sharper image. People think that blowing up an image will ALWAYS make it softer (correct) and shrinking an image will ALWAYS make it sharper but this is not always so. Because high filtering/sampling uses sub-pixel interpolation, when an image is shrunk down by 1, 2, 5, or even 10% you may loose a tiny bit of sharpness in extreme detail. I say this because many folks will be mastering their RED footage at 2K and then downconverting to HD, which is a very small ratio. It may turn out that cropping to HD will provide a superior image, although only Graeme would know. I actually suspect that downconverting from 2K to HD will produce a nicer, more filmic image as Red's problem WILL NOT be sharpness - but we'll have to wait and see I guess.
 
You raise good points. One thing we can do though, as we know the camera and the nature of it, we can get the downconvert parameters "right", and if you don't like them for one reason or another, you can always do it yourself.

Graeme
 
Sounds smart - but just be preppared for people to call REDCINE software slow - because "right" generally involves high sampling / filtering / anti-aliasing, etc. But this is as much a marketing issue as anything - if RED delivers great images out of the box, and REDCINE delivers great downconverts with its default settings, then you'll probaby preempt the negative internet chatter (whoa, sounded like Rummy there for a minute - scary!) Anyway people can always change the settings if they want a downconvert real fast - but then they know who to blame!

Graeme I have a quick question for you - gotta write it down. Be back.
 
Last edited:
Okay, this is a REDCODE question. Obviously everyone is doing backflips over your 10x compression ratio - I for one can't wait. However... I also do a lot of green-screen shooting and am a bit worried that 10x compression will be a bit much for pulling good keys. And uncompressed RAW @ 323 MB/s is too much. For me 3-6X compression sounds about right. For those rare occasions - will REDCODE have options (i dunno like Sorenson 3), or is it just the one setting?
 
REDCODE can do lossless all the way down to yuck. 10:1 looks sweet. The quality might even improve from what we're doing today too. There are hardware limits and drive limits to compression, and I'm not sure what they'll be exactly, but, I think you'll have no problem pulling a great key.

Graeme
 
Graeme_Nattress said:
REDCODE can do lossless all the way down to yuck.
Graeme
Click to expand...

Awesome - does that mean the user can dial in the compression parameters until they hit the data rate bottleneck?
 
BTW - anyone who's ever tried to pull a key off of HDcam will know what I'm talking about. Unless your subject is SpongeBob SquarePants or IceMan from Top Gun, you're going to get some blocky keys
 
dear: for your JIBBER JABBER MIND

dear: for your JIBBER JABBER MIND

boothba said:
Sounds smart - but just be preppared for people to call REDCINE software slow - because "right" generally involves high sampling / filtering / anti-aliasing, etc. But this is as much a marketing issue as anything - if RED delivers great images out of the box, and REDCINE delivers great downconverts with its default settings, then you'll probaby preempt the negative internet chatter (whoa, sounded like Rummy there for a minute - scary!) Anyway people can always change the settings if they want a downconvert real fast - but then they know who to blame!
Click to expand...
jibberjabber
 
Last edited:
On the subject of REDCINE software being potentially 'slow', I was just wondering if there are possibilities for harnessing the hardware downconversion, LUTs etc built into the camera. You could use the REDCINE interface on a computer, but let the camera do what it does 'on the fly' when recording REDCODE 1080p for example.
 
Nick_Shaw said:
On the subject of REDCINE software being potentially 'slow', I was just wondering if there are possibilities for harnessing the hardware downconversion, LUTs etc built into the camera. You could use the REDCINE interface on a computer, but let the camera do what it does 'on the fly' when recording REDCODE 1080p for example.
Click to expand...

funny, i was thinkin the same thing
 
You must log in or register to reply here.
Back
Top