I got the opportunity to shoot a quick test with one of Sony’s new SR-R1 recorders while in Dubai. The SR-R1 will have a street price in the region of $18k USD and a 256GB SR Memory card will cost around $7,500 USD!!! The big selling points of SR Memory are it’s speed and reliability as each cartridge (bigger and chunkier than most memory cards) contains a solid state raid array. However, while the price of the SR-R1 recorder is not in itself unreasonable, I feel the cost of media is sky high. So with that in mind I wanted to see how it compares overall to the Convergent Design Gemini which is a cheaper solution. The Gemini costs about $6k with a 256Gb SSD running at around $750 USD.
One thing to take into account is the true cost of storage on these two devices. While the Gemini media is cheaper, you do require a lot more of it per minute. If your shooting 444, 1920×1080, 24p then your looking at around 750GB per hour which equates to a cost of around $2,250 per hour. The SR-R1 recording SR-SQ which is the standard quality setting (440Mb/s 444) will fit a little over an hour of material on a single 256GB card, so SR Memory works out at around $7,000 per hour, almost 3 times as much as the Gemini. Of course in both cases you will re use your media many, many times, so in the long run compared to tape the media costs may not be quite so scary.
So what do you get for all this extra money with the SR-R1? Well you do get a faster workflow. The SR Memory cards can operate at upto 5Gb/s, so with more compact SR-SQ material, on paper at least, a transfer speed of around 10x real time should be possible. However in reality transfer speeds will be much lower than this as there just aren’t the data interfaces currently available that can sustain this. Still the transfers will be a lot faster than the Gemini which in a perfect world would also max out at 5Gb/s but again in reality will be somewhat slower, but due to the much larger files size. Even in a perfect world the Gemini, uncompressed, would probably at best reach 2x real time. Being realistic with well setup systems you should be able to get 1.5x real time with the Gemini and 5x realtime with SR-SQ. Now this could be a big time saving on a large project and you have to consider this. You also have to figure out how your going to store the Gemini data long term as don’t forget you probably want at least two copies. Now one of the ways to use the Gemini is to do a first encode as you copy the material of the SSD’s and this can seriously reduce your storage requirements and actually speed up your transfer speeds.
But what about image quality? Well the Gemini is uncompressed, so there is no difference between your recordings and what comes out of the back of the camera. This really is as good as you can get. The SR-R1 recording SR-SQ is compressing the data by about 4.5:1. That’s not that high by todays standards, but it is still a fair amount of lossy data squashing. Comparing Gemini DPX files and SR-SQ files it is very difficult to see any difference. There’s no added noise as occurs when you use ProRes and compression artefacts are hard to find. Look hard enough with a highly detailed image or lots of motion and you can find a bit of mosquito noise, but it’s very, very minor and I’d be more than happy with the material, even for high end productions.
So in summary the Gemini offers a perfect copy of the camera output, for a lower initial outlay and has cheaper media. But the workflow is a bit slower and you must consider how you will archive your data. The SR-R1 is more expensive, as is the media. It produces a near perfect copy of the camera output and has a fast workflow with files that are more compact and thus easier to store and work with.
The Gemini also has the benefit of a great LCD screen, while the SR-R1 is more like a traditional VTR. Gemini uses a lot less power, SR-R1 has a fan and is more power hungry. For me, the Gemini is the winner, the image quality really is as good as it gets. For production companies working on multiple projects or long form films the SR-R1 may be preferable simply because file management may prove to be easier.
Thank you for the quick follow-up on the SR-R1.
With all these options for recording the 4:4:4 output of the F3 coming up, I simply don’t get the following point:
You have these on-camera recorders, recording an enormous amount of data. Data which needs to be downloaded somehow, somewhere. 4:4:4 recordings are needed mostly for medium to big productions in a set. My point is: When you have a set – you have power, when you have a team – you can have a DIT.
So why use these pricy on-camera recorders that go along with drawbacks when you can have a Blackmagic Ultrastudio 3D with a thunderbolt RAID connected to a MBP for half of the price of the Gemini?
Here is my blogpost of this setup:
http://blog.mxr.at/sony-f3-444-rgb-uncompressed-workflow/
Sure you can use any one of the multitude of tethered recording options when working on a set. But you have many weak links in your chain, long cables to the capture device, cables from the capture device to the computer, cables from the computer to the raid array and it all needs mains or generator power. It’s rarely as convenient as simply pressing the on button and pressing record. There will be productions where this is appropriate, but just as many where it is not. A lot of these recorders will be used on TV productions, documentaries, natural history, location drama etc. Many, many high quality productions are done with small crews, sometimes individuals working alone. The high end natural extremes footage that I shoot for the BBC NHU, Discovery and Nat Geo is more often than not done single handed and in some very extreme environments, yet devices like the Gemini mean that I don’t have to compromise on image quality.
We have been able to record with computers for over a decade, yet still VTR’s and external recorders get sold by their thousands simply for convenience and reliability.