dustylense
Veteran
Scarlet is not Epic. Maybe the sensor is the same, but the processor is not. Frame rates and resolution. at 24fps, one can watch Epic footage and judge a scarlet. After that.....
Red Scarlet footage anyone?
- Thread starter reem12
- Start date
Rick Burnett
Section Moderator
You keep saying this but how do you know? I asked you before on another thread and you didn't provide an answer. The ASICs inside the Epic/Scarlet can absolutely be the same, because not all ASICs run at the same speed after they have been made. The slower parts can be used in the Scarlet, which makes perfect sense. I'm not saying they are or not, but if you are going to make that claim as if it is fact, I'd like to see where that is stated.
dustylense
Veteran
Rick, it's all over the REDuser forums. The ASIC's were the rejected ones that didn't work in Epic. They can't handle high res high framerates. So they packed them in Scarlet. Like I said, at 24p the images will be identical on Scarlet and Epic. Go above 30p and it's a different story. And as I have said, go to Vimeo and search RED Epic. Every 2 of 3 videos has 120fps or higher in it. Scarlet at 120 fps gives you what? 1.5k res? 60fps gives you 2k? All at very windowed sensor capture. So if you're looking for a medium wide 120fps of a skateboarder shooting a rail on Scarlet, you will be needing a 7mm lens and only get 1.5K res. How is that Epic?
Rick Burnett
Section Moderator
As I said:
http://www.reduser.net/forum/showth...s-not-deemed-Epic-worthy/page3&highlight=ASIC
(Post 26)
Just as I expected, the ASICs are the same, they are just running at a slower speed. This is VERY typical in chip design. Given I worked with the people who worked on these specific ASICs, I can say for certain that I at least worked in the same process node as they did, so I know how the part yields work out generally. (Yes, I've done chip design work for Red).
http://www.reduser.net/forum/showth...s-not-deemed-Epic-worthy/page3&highlight=ASIC
(Post 26)
Just as I expected, the ASICs are the same, they are just running at a slower speed. This is VERY typical in chip design. Given I worked with the people who worked on these specific ASICs, I can say for certain that I at least worked in the same process node as they did, so I know how the part yields work out generally. (Yes, I've done chip design work for Red).
Rick Burnett
Section Moderator
The processor is THE SAME. Just because it runs slower doesn't mean it isn't the same thing. It's digital, so as digital goes, they are IDENTICAL in function. That means, sure, they could do a higher redcode on it, however, the processor speed is running slower so it wouldn't be able to handle the frame rate.
You can say "You cannot run the same level of compression given the processor is clocked slower" and that would be accurate, but that's not what you keep saying. I understand the cropping limitations, but again, it's not because it's a DIFFERENT processor, it's because it is the same processor running SLOWER.
John Caballero
Veteran
Where is the frigging footage?
Rick Burnett
Section Moderator
Exactly. Red says that this is not the case with the sensors in the post I've linked. Speed binning is a standard part of chip design.
Look at it this way, if you buy a 2.2Ghz i7 microprocessor or a 2.3Ghz i7 microprocessor from intel, do they work the same? Yes. Why? Because it's the same chip, just one of them can run faster, and Intel sorted them after being fabbed by running them through test patterns at different speeds to make sure they pass (This is called DFT by the way). ALL performance based chips go through this.
Red did the same thing, and instead of throwing out the parts that are not fast enough for the Epic, they created Scarlet and used them. I would not be surprised if the boards inside were IDENTICAL in both the Scarlet and Epic. Financially, it makes NO sense having multiple processors, and this way, Red doesn't just throw out a huge amount of ASICs because they aren't fast enough for the Epic. Makes GREAT business sense.
Rick Burnett
Section Moderator
I think the big question is, what does the compression at these higher settings do? I think it is a VERY valid point. However, if you can set the EPIC to those settings today (the redcode level) you can test THE EXACT algorithm more than likely TODAY if you want to know. I don't know why nobody has been trying that (unless the Epic can't go that low) 
Edit: I am attempting to find out if people can do this now with the Epic to get an idea of the 4k mode delivery with the Scarlet-X level of redcode.
Edit: I am attempting to find out if people can do this now with the Epic to get an idea of the 4k mode delivery with the Scarlet-X level of redcode.
Last edited:
klas
Active member
<
Read through this thread.
I think the big question is, what does the compression at these higher settings do? I think it is a VERY valid point. However, if you can set the EPIC to those settings today (the redcode level) you can test THE EXACT algorithm more than likely TODAY if you want to know. I don't know why nobody has been trying that (unless the Epic can't go that low)
Edit: I am attempting to find out if people can do this now with the Epic to get an idea of the 4k mode delivery with the Scarlet-X level of redcode.
The Social Network was shot on red mx, 4k, redcode 42 compression (epic equivalent: 7.5:1). the scarlet 4k compression is 6:1. so i guess it will be enough.
dustylense
Veteran
Rick,
I had read (which I can't find) from Jim saying that they looked at the processors that weren't good enough for Epic laying around and they decided to use them in Scarlet. Jim posted this. They sensors are identical. The processors are not. If they were, Scarlet would be the Epic. Why would they have processors laying around that weren't Epic capable? They were not able to handle the high frame rates that Epic demanded. But they work in Scarlet, up to 30fps. After that, you window the sensor for your 48, 60, and 120 fps frame rates.
Put a Epic and a Scarlet next to each other at 24fps, and you'll get the same image. Do that at 120 fps, better have lots of space behind you or a much wider lens for the Scarlet. Oh yeah, and a 1k monitor.
So what I'm saying here is, IF YOU VIMEO EPIC footage, you will see a plethora of 120fps capture. Do it. Try that now. It's almost as cliche as a 1.2 lens on a 5dmk2. Overused. Scarlet will NOT give you that unless you're satisfied with 1k.
It's the perfect B cam for the Epic at up to 30fps. Which is fine. It will look RED. Just not at high frame rates.
I had read (which I can't find) from Jim saying that they looked at the processors that weren't good enough for Epic laying around and they decided to use them in Scarlet. Jim posted this. They sensors are identical. The processors are not. If they were, Scarlet would be the Epic. Why would they have processors laying around that weren't Epic capable? They were not able to handle the high frame rates that Epic demanded. But they work in Scarlet, up to 30fps. After that, you window the sensor for your 48, 60, and 120 fps frame rates.
Put a Epic and a Scarlet next to each other at 24fps, and you'll get the same image. Do that at 120 fps, better have lots of space behind you or a much wider lens for the Scarlet. Oh yeah, and a 1k monitor.
So what I'm saying here is, IF YOU VIMEO EPIC footage, you will see a plethora of 120fps capture. Do it. Try that now. It's almost as cliche as a 1.2 lens on a 5dmk2. Overused. Scarlet will NOT give you that unless you're satisfied with 1k.
It's the perfect B cam for the Epic at up to 30fps. Which is fine. It will look RED. Just not at high frame rates.
dustylense
Veteran
Here is a thread that maybe mods from RED will jump in and clarify. http://www.reduser.net/forum/showth...-quot-when-a-sensor-is-not-deemed-Epic-worthy
THe last I saw about the ASIC's in Scarlet from Jim's mouth was that they were repurposed from ones that did not meet EPIC's goals.
THe last I saw about the ASIC's in Scarlet from Jim's mouth was that they were repurposed from ones that did not meet EPIC's goals.
Rick Burnett
Section Moderator
Do you shoot all your movies in 120fps? Because I certainly don't. Just because people put up 120fps VIMEO videos really has no relevance to how it is being used for the majority of uses? For many people, they never even shoot outside of 24fps on the Scarlet for the reasons you mention. For those people, the question becomes how the REDCODE works on the Scarlet. That would be my question if I was still thinking about getting one.
But regardless, like I said, the processor is the same. Just like the example of the i7 at 2.2Ghz and 2.3Ghz. One runs faster than the other, but from a functionality standpoint, they can perform the same functions EXACTLY only at a different rate. By telling people the processor is different, that is NOT true. What is true is that it is THE SAME processor, but runs slower because the process variation in these specific ASICs doesn't allow it to run AS FAST.
If I take the Epic processor and slow the clock down to the same speed as the Scarlet, guess what, it works IDENTICAL. Why? Because the processor is THE SAME.
Sean D Brown
Member
Some scarlet footage.
http://vimeo.com/32843511
http://vimeo.com/32843511
Rick Burnett
Section Moderator
That's the same thread I linked!
Exactly! They didn't meet the same goals because process variation caused the design to not handle the clock speed. That doesn't mean the chip can't do everything IDENTICAL still, it just can't do it at the same speed. A slower clock has to be used. Or, and this in impractical, it's possible to use special high performance cooling on the chips used in the Scarlet and I am SURE you can get them faster. Temperature is one of the BIGGEST things that slow down ASICs. (Because temperature increases resistance).
One just has to be run slower to prevent fails, that's it.
dustylense
Veteran
That's the same thread I linked!
Exactly! They didn't meet the same goals because process variation caused the design to not handle the clock speed. That doesn't mean the chip can't do everything IDENTICAL still, it just can't do it at the same speed. A slower clock has to be used. Or, and this in impractical, it's possible to use special high performance cooling on the chips used in the Scarlet and I am SURE you can get them faster. Temperature is one of the BIGGEST things that slow down ASICs. (Because temperature increases resistance).
One just has to be run slower to prevent fails, that's it.
Then Scarlet is not and Epic then.
Like I said. Image wise you will get identical results up to 30p (but basing off the 1:30 only of Scarlet footage viewable let's just keep this as theory). Look and browse the Scarlet forums. See how many new guys are on there with 24 posts and such. All talking about what number they are and when will they expect their Scarlet. OK, so you have a Scarlet now. You can claim that you're a 4k RED studio. What IF...you're client hires you because of that claim and the client says to you, "OK, this shot I would like real nice over crank of the models. Let's go 120fps". What is your response going to be to that client? Now go back and look at all the new comers questions about high fps shooting.
No, I don't shoot anything at 120fps. But LOT'S of RED shooters do, as vimeo is evidence that it's the crutch to their work. Look it up! Count them. Just how many videos on there of Epic testing a 300fps of face slapping? It's worse than looking up f/1.2 shots from 5dmk2's.
So once some of these HDSLR shooters get hold of their Scarlets, they will be running around mad looking for solutions to the clients request of that over crank shot, because they advertised themselves as RED 4k studios.
You asked me if I shoot all my movies at 120fps. I'll ask you this. How many times have you felt the need to shoot your movies at 4K? THen down to 3k, then down to 2k, then maybe a 1K shot, all to be delivered at 1080p on Vimeo?
dustylense
Veteran
And still, we have only 1:30 worth of footage to go from. Does anybody other than myself find that puzzling? Again, somebody will post a picture of P Jackson with an Epic. That's like saying if you want to see Canon T2i footage, look at some 7D work.
Rick Burnett
Section Moderator
Actually, I don't think the images will be exactly the same if you use Epic's higher bit rate to your advantage. Clearly there are reasons when using a higher bit rate is important, or, the Epic wouldn't have them in the first place. THAT to me is the question I'd want to evaluate between both cameras. That is the big difference between the two when talking 24p at 4k. Of course, as you also know, the Scarlet isn't technically shooting 4k of resolution either since a bayer pattern cannot resolve the detail at the same resolution as the bayer pattern. So even in that, it's a compromise if you want real 4k footage.
Maybe you are miss interpreting what I was having issue with. The Scarlet does not have the same functionality as the Epic, I absolutely understand that. I understand the windowed modes of the Scarlet, and I am assuming like you, I don't like that limitation. If a studio hires me and I don't know the limitations of my camera, then I am an idiot.
If someone makes claims about what they can offer, and they don't, they'll learn the hard way. But that hardly has to do if the same processor is in both cameras. What it does have to do with is that one of those processors is underclocked compared to the other, and because of that, it limits what functions you can use on it. More correctly, it limits at what framerates they can work at.People are testing 120fps because it's fun. I did the same thing with my FS100 when I got it, and my AF100. (60p) There is a 120fps mode on the FS100 that is REALLY LOW REZ that I've played with as well, but I certainly will not use it in a production! (Way too low rez and limited).
I definitely agree with you on the windowing, I don't like it. The image quality at 1k is going to look really poor compared to 4k brought down. I'm not saying the lower clock speed hasn't made the Scarlet undesirable for me, because I certainly think it has! Again, all I am saying is the processor is the same, it's just running at a slower speed. I *think* what you are saying is don't expect the same performance out of both of them, and that is DEFINITELY correct.