Welcome to our community

Be a part of something great, join today!

Register 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 :)

Dragon....

I love my Red but agree completely with Mike. Show after show on the Alexa.. can't give my Epic away for a C camera they bring in a C300 first.
 
mikeburton said:
Kyle, with all due respect, I'm not referring to the individual owner operator here who is making it work in their bedroom or little home office. I'm talking about facilities, DIT's and the like that may not be editing at all and don't care what Premiere can and can't do. Professionals need to work in real time and speeds faster than realtime to accomplish dailies turnarounds that same day or next day at the latest. On big shows that requires a multiple machines with multiple RR's just to do the same thing that they can do with other codecs at a much lesser cost. I'm fully aware of how I can make a short form TVC or industrial / corporate piece quickly on a non-calibrated, simple MacBook Pro system, but that's not what is appropriate for larger infastructure post facilities and DIT's. With the rate of failure on these cards it's a terrible investment yet one that strong arms these companies big and small to own due to the lack of offering a simpler solution which I'm convinced could be done. But where's the money in that? The money is in the treatment not the cure.
Click to expand...


earlier in this thread it was stated that the Dragon will have some GPU debayer support...problem solved...

ps why cant you use an offboard recorder if your shoot at speed and not overcranked? it makes the epic just like the alexa... i dont see why producers would care...the pix240 and other solutions are excellent
 
If you had a project you controlled, one that you held dear, something yours... what would you shoot that on?

You guys know my answer :)

I would love to see some kind of Super Rocket. A top-shelf post hardware tool at a steep price but providing much-better-than-RT 6k processing... Two levels, with the new rocket coming in at the regular price. I dunno though, maybe that just isn't possible for some R&D or component supply reason.
 
gpu debayer will still not be as fast as a dedicated card as a rocket. even if you look at competing f65 uncompressed sony raw, coming down at that resolution at full debayer still requires a huge amount of horsepower. i don't know about you, but lugging around a tower with a pci expander and multiple gpu's just to get that done is damn near impossible unless it's a sound stage or i'm just living on the truck all day. and i consider my place to be on set. so there are always huge bottlenecks, and with 6k r3d, along with gpu's having a tough time with wavelet decompression, all the optimization in the world won't equal to what a dedicated wavelet decompression card can do.

rocket X is supposed to get several times faster than real time results in it's top setup. pci e 3 16x slot. beefy cpu's etc. at NAB I was told they have been seeing 60+fps rendering speed at full 6k. how many gpu's is it gonna take to get those kind of framerates?

what i'm curious about right now is why still no rocket support for monochrome? if it's just wavelet decompression with no debayer, transcodes should be stupid fast. but monochrome footage doesn't even work with a rocket...
 
Yeah, but Tom, 6-12 GPUs cost less than one Rocket/Rocket-X. Don't get me wrong, for mobility's sake it's nice to have that horsepower in one dedicated card, but how often are you in a situation that absolutely needs that kind of one-card+laptop mobility?

Also, when are we getting some dragon stills?
 
Rocket-X with a proper warrantee should make the accelerator card option far more palatable for the bigger shows/post facilities. As Jeff K. suggests, the real commodity hardware answer might just be lurking in the coming wave of hybrid processors...

Cheers - #19
 
Brian .. I agree. I get brought into shoot and offer my Epic with my Pix240 and they still prefer other formats.. The word "RED" scares some. Lack of knowledge and in some cases early bad first experiences. Early Redone days.
 
Mike P. said:
Yeah, but Tom, 6-12 GPUs cost less than one Rocket/Rocket-X. Don't get me wrong, for mobility's sake it's nice to have that horsepower in one dedicated card, but how often are you in a situation that absolutely needs that kind of one-card+laptop mobility?

Also, when are we getting some dragon stills?
Click to expand...


when they start adding more cameras on set. :)

which happens a lot. if i need 3-4 gpu's just to get 60 fps with one camera. i'm down to about real time with 2 cameras. make that 3 or 4 cameras now, and now i need another system with 3-4 gpu's. easier to just hook up a few computers, laptop/tower and have one rocket on one, and subrent one for the laptop. i get a lot of multicam shoots.

don't get me wrong i'd love to be rocket free. my second one just died. and if they can get 60+ fps with just one gpu i'd be set. but i don't think that's ever gonna happen.


also i do a lot of location work too. company moves, have to wrap out of location quick. get on and off the truck quick. cram into tight spaces. i can easily put together my studio setup which is like a battlestar galactica cart. but NG when i need to have a low footprint. a post facility doing dailies is a different story, it makes less sense for them to have a dedicated card to do only one thing for one type of footage. so stacking gpu's makes the most sense. set life... different story. some ppl do it. in ny it's really tough. we don't have a lot of space.
 
Matt Ryan said:
http://www.magiclantern.fm/forum/index.php?topic=7139

If RED could accomplish this while maintaining it's RAW state, wow! Instead of an x track with higher shutter speed, what about a lower ISO reading to maintain highlights and or lowlights depending on how you set it up and expose. That would be insane!
Click to expand...

But as I understand it, this along with other dual readout designs (BMCC, Alexa) is essentially a workaround for overcoming a DR limitation of the ADC rather than the photosite array. If your ADC is not the limitation, as I understand RED has attempted with all its camera systems (unclear if they succeeded since we can't really test it as far as I know), then there is nothing to gain by applying varying amplification to the signal, they just set it to the optimal value and nothing that could be done to the amplifiers or ADC or in software will ever really increase the signal range beyond that.

If the sensor is the limitation, then to get better performance, you need a new sensor, hence dragon...
 
Matt Ryan said:
If RED could accomplish this while maintaining it's RAW state, wow! Instead of an x track with higher shutter speed, what about a lower ISO reading to maintain highlights and or lowlights depending on how you set it up and expose. That would be insane!
Click to expand...

It's mostly because the way ISO is presented in digital raw capture on DSLRs is not exactly a "truth". If you capture a "fat" properly exposed REDCODE Raw file you have access to the entire range of latitude (within reason) in that file.

By shooting with a higher shutter speed X track you are actually capturing more "data" that falls outside of the initial maximum captured dynamic range of your base exposure.

This is one of the reasons why most HDR professionals use shutter speed as exposure modifier, rather than ISO (which can produce weird noise issues and you don't really gain anything) or aperture (which can effect the detail captured within an image).
 
Tom.Wong said:
what i'm curious about right now is why still no rocket support for monochrome? if it's just wavelet decompression with no debayer, transcodes should be stupid fast. but monochrome footage doesn't even work with a rocket...
Click to expand...

Perhaps, and this is pure speculation as I have no specific knowledge into how the data is written, but perhaps the Red files are written in sort of "blocks of 4"? The Bayer Pattern is 2 green pixels, 1 blue, and one red, arranged in a square. We all know this. If the color Red files are written like this, and the Monochrome are not (as in, the monochrome files are written 1 for 1, not 4 for 1), and if the Red Rocket is hardwired to decompress in that 4 for 1 pattern, then perhaps the hardware is simply incompatible with the Monochrome files.

And that is one of the poorest written explanations of a concept I can loosely speculate on in my head that I have ever scribed. It boils down to, if the Red Rocket is looking for data written as 1 5 9 13 17 . . . and the Monochrome writes data as 1 2 3 4 5 6 . . . then the formats are simply not the same, and thus the Rocket is useless.

Again, pure speculation, but maybe that's why.
 
Umm can someone from RED please post some actual footage so we can discuss Dragon and not other stuff in this thread. Why do I find myself constantly checking this thread looking for pertinent information and end up being disappointed everyday.
 
@Noah,

The ISO on the Canon sensor can be varied to maximise the S/N below a maximum light reading.

With the MX sensor - there was also a sweetspot 'ISO' that gives maximum S/N.

With the Dragon - there is a broad range of ISO with stellar S/N.

@Matt,

The DR in the Canon sensor is on the low side, 3 stops more DR with mild visual impact is quite a feat.

I think Canon should steal that man before another big company does.


My own speculative thought (if its possible on the RED sensor to read without resetting ) was that there may be mileage in implementing a Canon-like feature - to :
+) RESET -> EXPOSE Sensor for 1/10000 sec
+) READ 'Dark Noise' (Save X track)
+) EXPOSE Sensor for another ~1/48 sec
+) READ 'normal image' (Save normal image track)
+) SUTBRACT the Noise frame from the Normal frame.

This may allow the Red sensor S/N to be bolstered in very low light conditions beyond what a one off Black calibration could achieve.

AJ
 
hmm

hmm

yea i'm probably a little too optimistic on decent gpu debayer support for dragon i admit...but its all relative. not everyone needs fast on set transcoding, and like i said if they can get away with an offboard recorder...its a heck of a lot cheaper then a rocket too... i can bring a red, a pix240 (or equivalent) and a simple laptop and just cycle cards while real time getting dnxhd or prores. its quite amazing how much it can do without a bulky workstation actually!


the only problem being shooting off speed doesn't work well with this solution. for that a rocket is still the way to go. i swear by the rockets, but they aren't exactly inexpensive. almost all the work i do the productions supply the gear so i get by with long transcodes and edit raw at bad playback quality on my system when it comes to r3ds since i just don't need it... i cant even get people to buy checksum programs, much less rockets when it comes to budgets!

one day i'd like to own one...but theres no financial incentive for me to do so at this time... i've got a dragon to slay
 
Phil Holland said:
It's mostly because the way ISO is presented in digital raw capture on DSLRs is not exactly a "truth". If you capture a "fat" properly exposed REDCODE Raw file you have access to the entire range of latitude (within reason) in that file.

By shooting with a higher shutter speed X track you are actually capturing more "data" that falls outside of the initial maximum captured dynamic range of your base exposure.

This is one of the reasons why most HDR professionals use shutter speed as exposure modifier, rather than ISO (which can produce weird noise issues and you don't really gain anything) or aperture (which can effect the detail captured within an image).
Click to expand...

After talking to Phil and brushing up on the history of film-stocks, emulsions, formats etc... What the Dragon can do no other camera comes close to achieving.

- E
 
Ryan Sauve said:
I get more noise in my audio when I go from Sennheiser g3 -> Epic vs Sennheiser g3 -> h4n -> Epic. It’s ugly high pitch too. Quiet but definitely there. Maybe if I can find some time over the coming the week, I'll share a test to back this up.
Click to expand...

Ryan,

Try another microphone or the input from a mixer and see if that high pitch noise is still there. If it is still there I would open a support ticket with Red as you should be getting clean audio.
 
Paul Ellington said:
After talking to Phil and brushing up on the history of film-stocks, emulsions, formats etc... What the Dragon can do no other camera comes close to achieving.
Click to expand...

Oh you remembered all that mumbo jumbo?

Dragon may indeed have the secret sauce to exceed film relative to format and sensor size in nearly every way. I need to stress, this really hasn't been done before.

Some will always prefer the look of film and that's fine. I have bruises that I both love and loathe from working with various film formats over my career. There's technical and aesthetic merit to it. For me it's about where do we go from here.

In regards to resolution with Mysterium-X that's already been tackled really, Dragon's 6K basically takes that up to a new level. Dynamic Range however is the key to Dragon's mystery and actually surpasses film in terms of performance. It's 16+ stops of latitude and that's an easy 16 with more to spare from what I saw outside on that histogram. Light sensitivity is also well past film's optimal range and that's also been true with MX. Image noise is greatly improved and we're looking at perhaps around 3.3 stops better performance on that front. That's going to be a lot of freedom in post. A lot.

This new technology effects everything. Texture, feel, tonality, color. It also effect what you, me, and we can do with the material after capture. REDCODE is flexible and you can sharpen into the file a good amount if you want that "fresh lettuce" look, which many cameras don't handle well by the way. Go for a simple down sample and maybe add a grain pass then you'll land on a more organic feel. Done properly you can pretty much match any film stock if that's what you prefer. Pushing the tones around is what I'm most excited by from a grading standpoint, and this is really going to be where the digital cinema cameras show their differences.

At the end of the day I'm betting that Dragon at 6K will look as good or better than a properly shot and scanned VistaVision negative. Perhaps landing or encroaching on 65mm 5-perf in terms of image quality. There's some talk of that being exceeded even. Super35 in terms of it's technical marks will be surpassed by Dragon. I don't really have any doubt about that.

This is all rather good news because these new 4K displays and projectors demand a higher standard of visual quality. You'll be able to see and feel that difference on a 60 foot theatrical screen or even on screens as small as 8.5 inches diagonal with a tight pixel pitch.
 
3.3 sounds so very plentiful Phil! :) It will mean wonders in the highlights I bet.

With an even distribution that's 1.5 up top for roll off and you could underexpose 1.5 to pull that back up... which you'll never ever have to do unless you were waaaaaay off on exposure haha. I can't wait to see this!
 
You must log in or register to reply here.
Back
Top