Mike Thomann

With all of the other incredible Scarlet features aside such as frame rate and RAW code that make the Scarlet incredible, I am interested in understanding the viability of the 3K format for the Scarlet.

As the primary feature of the Scarlet is it's inspiring 3K format, it's worth considering whether the camera is capable of that format or if that format is just an extra overhead for the camera's actual resolved resolution as measured on the charts. Being a fixed-lens camera, this consideration becomes more of a concern.

If a 3K format camera doesn't resolve more resolution than other lessor format cameras, then a 3K format would only be an unnecessary overhead for the camera. If a 1920x1080 format camera can resolve 1000 lines on the charts, a 3K format camera should far exceed that otherwise the added overhead would be of no benefit over 1080.

Of course we don't yet know what the lens may be capable of, but it's difficult for me to conceive that a 2/3" lens at this price point could resolve so much more than 1000 lines to make use of 3K. If, in fact the lens is a match for 3K, I would be in awe. Unfortunately it will be quite some time before we can see the charts.

I'm unsure of the actual resolution of 3k, but since it's 16:9 maybe it's around 3000x1688?

Bill Koehler

I would only point out that there is no shortage of cameras built around 1/3" sensors shooting 1080 HDV with no one complaining about the output.

A great example at the $3000+ USD price point is the Canon XH-A1.

So if you more than double the sensor size, I see no problem getting significantly more lines of resolution out of the machine.

What does concern me is again, the megapixel game is getting played.
The pixel image sites are getting teeny-tiny. What is the low light performance like?

Of course, it gets some real glass hanging off the front to help with that.

Brian Drysdale

For a 3k Bayer sensor think a resolution of just over 2k. The resolution of the 4K RED One is around the 3K mark (3.2K on a RED test - I believe shot with a Cooke S4).

Mike Thomann

I'm more interested into getting to the bottom of actual results of resolvable detail in today's crave for more and more resolution recording formats. I love the concept of a camera whose resolution format matches it's resolution capability. The JVC HD100 is the perfect example, having more resolved resolution at a low 720 format than most other 1080 format cameras. The added benefit of that optimization is better color and low light sensitivity. Plus you don't have to work with unnecessary resolution overhead, which means more record time and easier workflow.


Good example, even though it's at the upper end of resolved resolutions for 1080 format cameras, the HD100 resolves more in it's smaller 720 format which is precisely my point.


So maybe larger sensors have the potential to pull more detail through equivelent-quality glass than smaller sensors?


Me too. It's important to not want resolution formats that are higher than cameras are capable of resolving because color, low light, recording time, and ease of workflow are compromised.

Mike Thomann

Ah I see. So with a debayed resolution of 2K, the final resolved resolution pretty much amounts to the lens ability.

Bill Koehler

I think trying to extrapolate performance of the RED Scarlett from the RED One is difficult due to the vastly different sensor sizes.
2/3" vs 35mm? That's a huge, huge jump.

Adam Palomer

Who would make the glass for the Scarlet lens? Canon grows its own crystals and then cuts them into lenses. There's a whole field of expertise behind that. What can we expect from RED? Are they going to source out the lens? Build it on their own? What kind of quality glass can we expect?

Steven White

The joy of 3K RAW is that you would expect a Scarlet to get the equivalent of 4:4:4 10-bit 1920x1080p imagery on down sampling conversions.

Again, that's 4:4:4 10-bit 1920x1080p from a 2/3" sensor. If that sounds familiar, it's what the Sony F950 does. The main difference one would expect is the quality and versatility of the lens.

Mike Thomann

The F950? So if the Scarlet were made with an interchangeable lens for another grand, the demand would be more than ten-fold and the world would be running around with the Scarlet version of the F950. Heck, Red could even offer it without a lens for the same price and it would cost much less to manufacture and the demand would still be insane.

But that's not likely to happen, so for comparison to some other cameras that it's expected to outperform, the most conservative and repeatable testing method (MTF50) for the Canon XH-A1 yields 810 Lines Horizontal and 660 Lines Vertical and the HV20 yields 784 Lines Horizontal and 705 Lines Vertical. So the Scarlet should easily resolve over 900 Lines Horizontal and 800 Lines Vertical when tested with MTF50, and I'd be very pleased with that.

Regarding MTF50; when looking at the charts, as the black resolution lines converge there's always that large region beginning at which blur begins to reduce contrast and then extending to the point of no contrast. It's fairly easy to identify three points: 1) The beginning of the blur, 2) the mid-section of the blur at 50% contrast (where perfect line distinction just begins to become difficult to identify), and 3) the point at which blur is in full force. In simplest terms MTF50 is the measurement taken at point #2.

As far as resolution, measured MTF50 is all it comes down to.

Odd Nydren

Well spoken. I agree totally.

Brian Drysdale

I suspect there might be a few compromises in the Scarlet to keep the price down and possibly it mightn't handle highlights and other issues as well as the more expensive cameras. Rather like a car that goes fast in a straight line, but doesn't go around corners that well.

It's very possible that third party manufacturers will modify the Scarlet with say a PL mount, so that you can use Super 16 lenses. However, I suspect the cost could be as much as buying the camera body (perhaps slightly more). That's no big deal to professionals looking for say a "B" camera, or a more versatile portable "A" camera for TV documentaries or lower budget drama.

Although, that would start entering into the SI 2K market and I can see them having to upgrade to a 3k chip. However, in that regard, much will depend on the images that the Scarlet actually produces.

Steven White

Obviously the lens options on something like Scarlet limit the full potential of the sensor. However, if the lens is well designed, it is probable that image quality of Scarlet would exceed that of the F950 with a similar lens.

It's easy to do the math and come up with the theoretical maximum resolution attainable by a 3K sensor. If you look at the Red One, with a good lens, it approaches the theoretical maximum resolution attainable by a 4K sensor, and I expect the same performance of Scarlet at 3K.

At its price point, the lens will be the limiting factor on what Scarlet is capable of.

Brian Drysdale

There's more to image quality than the resolution, it's just one factor. There's the sampling and other processes. You could have something that looks like a high resolution cellphone camera. Although, the actual resolution is closer to 2k and that would match up with having a 2k or a 1080p output. The Red One is around 3.2k according to RED.

This is all speculation, it's up to what RED makes of the Scarlet that will be the proof: at the moment it's just a shell. The RED One currently has it's own issues - subject to seeing how its build 16 turns out. So this is a process that takes some time.

Steven White

If you're an engineer at heart, you don't wait for someone else to solve the problem. Engineering *should* always start with a list of target specifications and a price, which is what Red has provided for Scarlet.

We have a lot of information with regards to the performance of the first generation Mysterium chip. We have a lot of information with regards to how the Red One was engineered. As such, very reasonable predictions can be made.

Brian Drysdale

Also, in business you don't want to trample on your own products. If the lens is the only limiting factor, RED could lose RED One business to hot rodded Scarlets modified for interchangeable lenses by 3rd parties. Most productions don't actually need 35MM DOF, so there's no reason for these people not to buy a Scarlet fitted out with a PL mount for say $6k total cost and fit it with Super 16 lenses.

The RED One currently tends to be noisy in tungsten lighting and has some other issues, so you can't pass on predictions.