SCARLET by Red Digital Cinema

[Halsu]

35mm sensor and use interchangeable lenses , come in one unit, and UNDER 3 grand. FAIL.

I'm pretty sure that only a small minority of people thought like that (though it's a very passionate and vocal minority). Red never implied something like what you describe was in the owen at that price, so i personally never expected it to happen - and that probably applies to most people who work in professional video. We know that pro tools come with a hefty price tag, partly because they're manufactured in small series, partly because requirements are much more strict than with consumer gear.

The 2/3" interchangeable lens Scarlet isn't competing with HX-A1 or D5 MKII, it's competing with Sony F23, which costs over 100 000 dollars - and it looks like it's winning.
http://www.sony.co.uk/biz/view/ShowP...site=biz_en_GB

I'm sure red could have released a camera like you describe, if the image quality and reliability (i.e. speed, skew and heat) weren't an issue. The MKII sensor is probably the best that currently can be done at that price point, when mass produced. Unfortunately, it's not enough for a professional video tool.

It's perfectly normal human behavior to always want more. But it also may make people loose the touch with reality ;-)

[Halsu]

Will the low end Scarlet do as well? Sort of. Of course you will have to tweak it in post, but isn't it nice with the A1 to set it up ahead of time and not have to tweak much if it at all?

If Red One is an indication (and i'm pretty sure it is), you should have the best of both worlds.

You can tweak your settings in the camera, and they are saved in the metadata. The proxies are made using the settings you had while shooting. That's the quick workflow - it's more or less the same as working with regular video cameras.

You just have the bonus of being able to go back and tweak those settings afterwards if you wish.

[Braceface]

By professional what do you mean? The camera is aiming specifically at the professional wedding market. Plus, in my opinion the PQ out of the camera rivals what I've seen with the Red, and thats just with people who don't even own it yet. And that is at 3000 dollars. Hmm.... I think also that its either disingenuous, or naive that you are claiming that the 'DSLR killer " isn't aiming at the Nikon or Canon,lol.

[Duke]

You can tweak your settings in the camera, and they are saved in the metadata. The proxies are made using the settings you had while shooting. That's the quick workflow - it's more or less the same as working with regular video cameras.

You just have the bonus of being able to go back and tweak those settings afterwards if you wish.

I think we'll have to see what options actually end up on the camera. My understanding from reading on Scarletuser was that there were a lot less settings on the Scarlet to keep the costs down.

However, rumor runs rampant over there and who knows where that came from and how true it is. I can say the opposite too though, unless you are an insider, what do you know for sure about the Scarlet? Jim did say certain things were left off. Of course that was when it was the pocket camera.

I note your statement that sensors have been built. I should hope there are prototypes galore for testing puroses by now. So why haven't we seen test footage from the Scarlet?

[Halsu]

By professional what do you mean? The camera is aiming specifically at the professional wedding market. Plus, in my opinion the PQ out of the camera rivals what I've seen with the Red

You're correct when you say that MKII image quality is pretty much equal to red one - but that's when it's shooting 12 megapixel raw stills, which it can do only 4 frames a second.

The MKII *video* quality is totally another issue. The only places it shines are low light and shallow dof. It probably skips pixels to be able to maintain 30 fps, which means the real original resolution is low, and it will probably have bad aliasing artifacts with small details, just like D90 does.

I'd rate the video quality (resolution, compression, sharpening artifacts etc.) below HV20, based on the footage i've seen. It would probably be a pretty good "video" camera for wedding photography though, unless it had the five minute record limit before it shuts down for cooling (and only in USA & Canada as it's 30p).

[Halsu]

Jim did say certain things were left off. Of course that was when it was the pocket camera.

Those are mainly options like frame rates, anamorphic squeezing for mnonitors etc., and due to technical limitations. I've heard nothing about dumbing down the basic software features.

I note your statement that sensors have been built. I should hope there are prototypes galore for testing puroses by now. So why haven't we seen test footage from the Scarlet?

Wasn't me. But it seems they had a working proto going at the launch party. I've heard no more details than that, so could just as well be red one sensor bolted to scarlet electronics or something ;-)

I guess we just need a little patience.

[Duke]

Wasn't me.

My mistake, it was the post above yours. Sorry.

I guess we just need a little patience.

Absolutely.

I think what braceface is getting at, is there is at least a year, if they go to Red users first maybe a bit more, before the Scarlets are available.

The hardware on the 5D is there. I know what you mean by selecting pixels, but the 5D doesn't have the jagged edges of the D90, so I would bet its run through an algorithm instead. It might be some thing as simple as the 5D picking a block of 9 pixels and averaging them because the D90 and 5D clips don't look the same at all.

If the 5D has a firmware change that gives 24p (which would allow a higher bit rate) and manual controls, both of which are possible with the hardware already there, it would be a heck of a camera at a good price point. But that's a big if. Canon might not do it, and hackers might not be able to though they've done so with other DSLRs.

Both cameras have interesting possibilities, and neither is quite there, but the 5D is usable now... even with it's limitations.

I am surprised I didn't get comments on the A1 DOF tests in posts 213-215 above. It's my camera and even I was surprised at how much DOF control I could get. (I'd done it outdoors when I had lots of room before, but this is the first time I tried it confined to the space indoors.) I think that's mostly attributable to the focal length/zoom of a 20x lens.

The fixed scarlet would have less than half at 8x, but a sensor that's twice the size. That's what makes me think the fixed lens Scarlet will have a little less DOF control than the A1, but nearly twice that of the HV20/30. (Yes, I did read what you posted above about it.)

[Halsu]

I think what braceface is getting at, is there is at least a year, if they go to Red users first maybe a bit more, before the Scarlets are available.

So it seems, as far as the fixed lens one goes. The others should be available sooner.

The hardware on the 5D is there. I know what you mean by selecting pixels, but the 5D doesn't have the jagged edges of the D90, so I would bet its run through an algorithm instead. It might be some thing as simple as the 5D picking a block of 9 pixels and averaging them because the D90 and 5D clips don't look the same at all.

Based on the footage i've seen so far, it looks like they're simply skipping sensor elements, just like D90. Averaging (supersampling) looks different, much better. There's quite a lot of sharpening in all shots i've seen, which is another implication that there's not enough resolution to start with. I also recall seeing aliasing moiré in some shots.

To me it looks like D90 footage, if you don't take the D90's odd skipped pixel rows into account.

There's a fix for those, BTW:
http://www.dvxuser.com/V6/showthread...=149663&page=5

Both cameras have interesting possibilities, and neither is quite there, but the 5D is usable now... even with it's limitations.

I thought it's not released yet?

Anyway, when it gets to the hands of regular end users, we will probably start to see some not-as-pretty videos from it... the ones online right now are mainly done by very good photographers (which is almost always much more important than what camera is used).

I am surprised I didn't get comments on the A1 DOF tests in posts 213-215 above. It's my camera and even I was surprised at how much DOF control I could get. (I'd done it outdoors when I had lots of room before, but this is the first time I tried it confined to the space indoors.) I think that's mostly attributable to the focal length/zoom of a 20x lens.

You should use the same framing with both cameras to really judge the difference in the DOF. Zooming more with one than the other makes the comparision a bit meaningless (though i do understand what you were after).

The fixed scarlet would have less than half at 8x, but a sensor that's twice the size. That's what makes me think the fixed lens Scarlet will have a little less DOF control than the A1, but nearly twice that of the HV20/30. (Yes, I did read what you posted above about it.)

It should have about the same amount of blur as your fully zoomed in A1 shots, at the same zoom level as the HV20. If you zoom out A1 to match the framing (or add a tele edapter lens to Scarlet), A1 it will have a bigger DOF than Scarlet (bigger DOF = sharper background).

[Braceface]

We'll see better footage than we see now. Even the worst 5D footage looks better than the best HV20 footage so far.

[Duke]

It should have about the same amount of blur as your fully zoomed in A1 shots, at the same zoom level as the HV20. If you zoom out A1 to match the framing (or add a tele edapter lens to Scarlet), A1 it will have a bigger DOF than Scarlet (bigger DOF = sharper background).

I think we're saying the same thing in different ways. If the Scarlet did zoom to 10x it would be about the same DOF as the as the A1 but it's only 8x, and with a 2x adaptor it will beat the A1 at 16x with a 2/3" sensor, though you'd then loose some light and clarity.

And you are right I used the fully zoomed A1 to show the maximum it was capable of doing in a limited space. I'd have to back up even farther to have the same shot framing and retain the narrow DOF. All I showed is that I can do a close up of a person indoors without an adaptor and not have the background in focus with the A1. That's a useful thing all by itself.

If the A1 was half zoomed to 10x it would be nearly the same as the HV20, but what would be the point of that? (That's an approximation since the A1 goes to f1.6 at widest and the HV20 goes to f1.8 at widest. The A1 does change during zoom, but not as much as it is an "L" quality lens.) The test was to show maximum DOF control, not to match the HV20.

[Halsu]

If the A1 was half zoomed to 10x it would be nearly the same as the HV20, but what would be the point of that?

That's the only comparison that really matters. What does the DOF look like at any given framing / FOV / zoom level?

Fixed lens Scarlet should have "twice the DOF effect" to A1 if all other things (camera placement, framing) are the same. More, if A1 goes below f1.6 when zooming.

I'd have to back up even farther to have the same shot framing and retain the narrow DOF.

That's the problem :-)

[Duke]

That's the only comparison that really matters. What does the DOF look like at any given framing / FOV / zoom level?

We can agree to disagree on that one. I only care about the ultimate amount of DOF control I have. If I have to move 5' farther back to do it and have the same framing, it won't bother me a bit, which leaves the A1 with a (admittedly very slight) advantage on DOF control as 8x with a sensor twice as big is roughly equivilent to 16x on the A1, whereas the A1 goes to 20x.

Admittedly, you can throw a 2x teleconverter on the Scarlet, and theoretically that should beat the A1, but normal tele add-ons (even HD ones) aren't intended for 3k images. The A1's "L" quality lens is the highest quality Canon makes.

[spideralex90]

Even the worst 5D footage looks better than the best HV20 footage so far.

Even a RED can look bad. It's not really the camera, it's whose behind the camera. And i would hope most everyone would join me in saying that sentence is nowhere near true. Have you seen White Red Panic?

[Halsu]

Even a RED can look bad. It's not really the camera, it's whose behind the camera.

Exactemente.

[Halsu]

We can agree to disagree on that one. I only care about the ultimate amount of DOF control I have. If I have to move 5' farther back to do it and have the same framing, it won't bother me a bit

Just moving the camera is not enough.

You would not only have to move twice as far from your subject and zoom in, but also move the background twice as far from the subject - only then you would get visually roughly the same "amount of DOF". Not so easy in small rooms ;-)

Also, having to always shoot with an extreme tele lens is far from an optimal way to shoot, IMO.

[Halsu]

I took a still from red.com, and cropped it to 3K. This is more or less what we should expect the scarlet output to look like (this shot doesn't have a shallow DOF). My guess is, it will probably look even better, as this image is from an early preproduction version of RED ONE... and of course, these are jpg's..




Full size:
http://eki.pp.fi/digivideo/RED/REDONE_3K.jpg

The same image scaled to HD resolution:
http://eki.pp.fi/digivideo/RED/REDONE_HD.jpg

The original is here:
http://www.red.com/shot_on_red/

[Duke]

Quote by Braceface:

Even the worst 5D footage looks better than the best HV20 footage so far.

I wouldn't agree with that, except in very low light. The HV20 is capable of footage as good as any other in good light. (And I do think the 5Ds potential will be unlocked.)

Have you seen White Red Panic?

What is the White Red Panic?

Just moving the camera is not enough.

You would not only have to move twice as far from your subject and zoom in, but also move the background twice as far from the subject - only then you would get visually roughly the same "amount of DOF". Not so easy in small rooms ;-)

Also, having to always shoot with an extreme tele lens is far from an optimal way to shoot, IMO.

I'll half way agree with that. In those I got the DOF narrowed down to less than 2'. In the bottom one some DOF effect is shown in less than 1'.

a) Its not usually necessary to have the DOF that narrow.

b) Not all shots need or use a narrow DOF. IMHO a narrow DOF is to shift the focus of the audience or convey emotions more directly.

Let's say I'm shooting a romance with the couple sitting at a bistro table. I might have a wide angle establishing shot when they start talking with no DOF control necessary.

Then the couple starts talking about their love for one another and I go to close ups with a narrower DOF so the other patrons are out of focus and the audience pays attention to what the couple is saying.

Since the people are sideways to the camera if the DOF was 1' parts of the couple would be in focus and part out of focus. I'd probably want a 3' wide DOF so the talent, including their hands is in focus, but the people around them aren't.

In that instance maybe 16x zoom would do it. I could do that in a garage sized studio. The Scarlet should be able to do that too.

Duke

[Braceface]

Even a RED can look bad. It's not really the camera, it's whose behind the camera. And i would hope most everyone would join me in saying that sentence is nowhere near true. Have you seen White Red Panic?

White/Red is a good little short, but it has a terrible picture. I've gotten much better pictures from an HV20 than that. The color grading is interesting, but look at how noisy the picture is. You wont get that with the 5D.
The 2 cams are apples and oranges (hv20 and 5d) The scarlet though is a mess of a system. But thats my opinion.

[Halsu]

White/Red is a good little short, but it has a terrible picture. I've gotten much better pictures from an HV20 than that. The color grading is interesting, but look at how noisy the picture is. You wont get that with the 5D.

Have you seen color corrected D5 footage yet? ...i didn't think so ;-)

Anyway - i had a look at your highway film. Pretty nice work, some good images there. But the technical quality is less than what one should ecxpect from HV20...

I'm wondering, did you leave the camera to full auto? The exposure pumps a lot, and there's many places where the footage is badly overexposed. Also, it seems you're getting a lot of vignetting with the adapter and it's very blurry overall, as if nothing was in focus - was this intentional?

The 2 cams are apples and oranges (hv20 and 5d) The scarlet though is a mess of a system. But thats my opinion.

You have a right to your opinion, naturally, i tend to disagree ;-)

[Halsu]

a) Its not usually necessary to have the DOF that narrow.

I agree totally - i was talking purely from technical standpoint. HV20 DOF is good enough for me in most situations... sometimes though i'd like just that little bit of extra ;-)

***********************

I guess i shouldn't do this, but here we go anyway... i uploaded some shots from paid jobs i've done with HV20. A "demoreel" of the HV20 stuff i happened to have on my machine right now, if you may.

There's some footage from low budget commercials, plus footage from a feature film's DVD extras. I don't really want this to be publicly available, so i will remove this link in a day or so. After that, if someone's interested, i can give the link privately (PM me).

http://eki.pp.fi/HV20_demo_v001.mpg
(720*408p mpeg1, 4 minutes, about 90 MB, no sound)

Most of the stuff is without adapter (sometimes using the variable ND trick from my sig.) There's one commercial i shot using my DIY 35mm adapter - extremely shallow DOF worked with that one...

[Ian-T]

White/Red is a good little short, but it has a terrible picture. I've gotten much better pictures from an HV20 than that. The color grading is interesting, but look at how noisy the picture is.

No..no..no...I think that short looked good for what it is. The picture was intentionally "dirtied" up. If you read what he was trying to accomplish with the picture you'd see where he said he was trying to dirty up the picture somewhat. My assumption on why he chose to do that is for a few reasons. ... The first...was...because he chose TV mode and on top of that he chose a 1/60 shutter instead of the conventional 1/48 shutter. Those two reasons alone brought it to look more like video...but whatever he did to muddy it up a little in post helped alot IMO. There were moments however that it looked like video to me...but overall...that short looked great.

You wont get that with the 5D. The 2 cams are apples and oranges (hv20 and 5d) The scarlet though is a mess of a system. But thats my opinion.

What makes you think that you won't? Like Halsu mentione...you have not seen any color grading with the MarkII as yet. I think I know why he stated that..and I believe it has to do with the compression used.

Actually, I have seen someone's attempt at color correcting...and it looked like crap. matter of fact I posted it in this forum some weeks back (can't remember where). There is no way that this cam could compare with the "Hypothetical" Scarlet...especially when Scarlet comes to fruition. Yes...it has a big sensor...but you are very very limited in how you can bring out its potential. Not with a Red cam however.

[Braceface]

the movie had a few scenes where I used auto settings. Those were without the adaptor, in a couple of the roof-mounted boom scenes. I'm sure you can guess why I used automatic for those....
Other than those, there is no "pumping".
OFCOURSE, a siht load of the scenes had blown out back-grounds, and I HOPE you can guess why.....
I didn't put it in my rules of engagement plan to only shoot when the sun was "just right" for a road movie with no budget. It wouldn't be practical with my budget, and the audience never mentions that as a problem, only a few people from this forum.
Highway 27 was in a different difficulty level than White/Red, purely for the amount of time, and organizing it took to get it done, and it being (the majority of time) away from electricity for controling light. Oh, I could shoot it a lot better, with the dough.... but for what I had to work with NOBODY has really done it to that degree and posted it on this forum. It was very extreme. Perfect, hell no.

[Braceface]

No..no..no...I think that short looked good for what it is. The picture was intentionally "dirtied" up. If you read what he was trying to accomplish with the picture you'd see where he said he was trying to dirty up the picture somewhat. My assumption on why he chose to do that is for a few reasons. ... The first...was...because he chose TV mode and on top of that he chose a 1/60 shutter instead of the conventional 1/48 shutter. Those two reasons alone brought it to look more like video...but whatever he did to muddy it up a little in post helped alot IMO. There were moments however that it looked like video to me...but overall...that short looked great.





What makes you think that you won't? Like Halsu mentione...you have not seen any color grading with the MarkII as yet. I think I know why he stated that..and I believe it has to do with the compression used.

Actually, I have seen someone's attempt at color correcting...and it looked like crap. matter of fact I posted it in this forum some weeks back (can't remember where). There is no way that this cam could compare with the "Hypothetical" Scarlet...especially when Scarlet comes to fruition. Yes...it has a big sensor...but you are very very limited in how you can bring out its potential. Not with a Red cam however.

....ok

[booggerg]

Actually, I have seen someone's attempt at color correcting...and it looked like crap. matter of fact I posted it in this forum some weeks back (can't remember where). There is no way that this cam could compare with the "Hypothetical" Scarlet...especially when Scarlet comes to fruition. Yes...it has a big sensor...but you are very very limited in how you can bring out its potential. Not with a Red cam however.

Taking a page from my experience of working with RAW files from small P&S cameras. The quality of the sensor overwhelmingly determines the latitude of the resulting image. My Ricoh GRD has a RAW mode and the RAW images from it can not stand up to JPEGs from my Canon 5D when editing in photoshop. Not even close.

I suspect it to be the same case when it comes to Scarlet 2/3 vs 5D2..

[Braceface]

Yeah, the 5D will be way easier to grade without noise, regardless of some dumb example on Vimeo. You can destroy anything if you grade it without skill, or taste.
I've seen REALLY bad HV20 grades, obviously too right here at our forum.
Anyhow, everybody should use the tools they want. The Scarlet has the 3K to 28K thing going for it obviously, and depending on what you pay, but that isn't the most important thing to me. A 2/3 sensor is not one of my interests, even with interchangeables. I'm looking for full frame 35, with 35 mm lenses. I like the 35mm look better than the 16mm look.....
For my needs, the 5D is PLENTY adequate for the next couple of years for me, and I'm not getting rid of the HV20, so if one does better for one thing over the other, I'll use the right one. In a couple of years I have a feeling that I'll be getting another camera besides the Mark II, since technology is exploding in digital imaging. But I'm not going to go the route of Scarlet/Epic. I have a gut feeling that I'd be stuck with something very tacky, and outdated, FOR THE PRICE. ( not the quality, which I'm not trying to degrade). I'd be interested in how many people are like me in that they just want to buy a really good friggen camera, in ONE piece, even if it isn't something upgradable to becoming a 30,000 dollar beast.