4k Redcode vs Prores 422 vs 2k Redcode?

[david farland]

sigh.

 

[Shawn Nelson]

I've drawn the same conclusions. And since the best film-out rates from the labs are for 2K Arri, it makes sense to me to shoot 4K, down-res to 2K (or 1080P), and take advantage of the Apple ProRes RT Extreme features and Apple's "Color" app.

What's great is that you always have the 4K files (your neg so to speak) to go back to if your project is successful and you need a higher quality filmout or 4K digital cinema version.

My thoughts exactly. I'll shoot in 4k, downres to 2k or 1080p and work with that, much more affordable and practical.

 

[Damien Molineaux]

Back to the drive issue raised by Shawn.

<cut>

1. The RedDrive is 320gb and I've heard that it will last around 2 hours at 4k. But wait, the RedCode format is supposed to be 27MB. 27MB * 60s = 1660MB per minute. 1660/1024 = 1.582 GB/Min of 4k Redcode. So 3 hours should be only 285 GB, well within the RedDrive limit. So what's going on here?

<cut>

Ever noticed how when you buy a drive indicated as X GB, once formatted you only end up with (X - 6 or 7%) GB. This happens for two reasons. HD vendors label their drives as 1 GB = 1000 Bytes while for your computer, 1 GB = 1024 Bytes. So you've already lost 2.4% right there. Then you have all the overhead of data handling and logging which requires around 4%. So, I don't have a 320 GB drive around to check precisely, but once formatted I would expect a Red drive to have about 297 GB of available space.

The second thing to take into account, is that as a drive gets full, it becomes slower. Mike from HD4indies raised that question a while back, will a Red drive be fast enough to record 4K Red Raw when it is nearly full. Which raises a couple questions, how much of a margin should be left aside ? And if you can, can you choose the amount you ? Trying to fill up a disk completely could kill it, what happens if you're recording and there's no space left ? Will there be a safe guard ?

Cheers,
Damien

 

[Adrian T.]

The second thing to take into account, is that as a drive gets full, it becomes slower.

The speed is only dependent on the following two factors, and not simply on "getting full":

• If the used allocation blocks are fragmented, you have more random accesses as opposed to sequential accesses. And these are significantly slower. But when you start with an empty disk and just record, the disk will not become fragmented.
  
• Access to the inner zone of the disk is somewhat slower than to the outer zone. The speed decreases linearly dependent on the radius. I could imagine a RAID 0 system where two disks are being filled in opposite directions. This would yield a more or less constant throughput.

 

[TimothyD]

I could imagine a RAID 0 system where two disks are being filled in opposite directions. This would yield a more or less constant throughput.

That is a DAMN good idea Lebowski...

I wonder if there is any way Red can force it to write that way? It seems like that would require getting at the innards of the drive? Who knows... Great idea anyway.

What we need is drives that are manufactured as RAID 0 that are already set up that way, where the RAID is built into the drive and it is writing to the platters like you mentioned, one from the inside, one from the outside. Then you wouldn't even need two drives...

Tim

 

[david farland]

I wouldn't worry too much!
If Red can't raid 0, two 2.5" drives with a continual sustain write of 35MB, somethings horribly wrong.
I'd be more worried about tracking your raw color corrections/rec 709 space (Prores). No one's explained that yet!
cheers,

 

[Nick Shaw]

I think somebody (could be Avid?) has already patented the idea of RAIDing a pair of drives which write in opposite directions.

 

[Antoine Baumann]

I'd be more worried about tracking your raw color corrections/rec 709 space (Prores). No one's explained that yet!
cheers,

Well I don't have the answear but Crossing the Line was color corected in rec 709 color space.

antoine