Hi guys
going off the other post - I think it would be amazing if you allowed compressed recording onto the Gemini such as the features the Sound Devices recorder will allow - such as Pro Res 422 HQ and Avid DNx and even the nanoflash 8 bit codecs. Just because there are going to be times when you want to preserve on disk space and you want to just have one recorder, instead of buying the Gemini when you're doing 4:4:4 work and then the Sound Devices or Ki Pro Mini when you want to record 10-bit pro res.

let me know what you guys think and if this has been brought up before.

I had this very conversation with the guys in the booth today. The simple answer as told to me is its not possible because the hardware to do that isn't in the case and that was one of the things that had to be left out to keep it small.

This decision puts that recorder out of reach for me at the moment. :(

I need to be able to do both.

The Gemini works full uncompressed. To record to any other codec would be necessary to add another processor native for that codec.
That would increase the size of the device and the power needs significantly.
rafael

Dear Friends,

While we do not have the ability to record compressed in the Gemini 4:4:4 we offer a solution that will work for many.

After one has recorded the full uncompressed footage, one can attach the SSD to our transfer station, which is included the our Gemini 4:4:4 kit for $5,995.

Then the footage is available as a mounted drive, an SSD with a 6.0 Gigabits per second interface.

The footage can then be encoded to the codec of your choice using Apple Compressor or Adobe Media Encoder or other software.

For 30 minutes of footage, this encoding occurs in about 15 minutes.

Thus, we use the general purpose computing power of your computer to do the encoding.

If we add the ability to record to multiple codecs in our recorder, the power draw will increase by over 10 watts, and the price will also have to increase.

Thus, we offer two solutions.

The nanoFlash, which allows compressed recording for long events (over 10 hours), and the Gemini 4:4:4 for very high quality recording (but this is limited at this time to 2 hours maximum, unless one wants to "Hot Swap" then the uninterrupted recording time is limited by your resources (number of SSD cards).

If you check the power requirements of all other uncompressed + compressed recorders you will find the power draw to be significantly higher. Ours is 8 to 15 watts, with around 12 watts being typical.

Dear Dan,
If I was to purchase a Gemini 4:4:4 I would like to edit in uncompressed what editors would you recommend that would allow the quality to be maintained through the editing process. Final Cut Pro will not do it god knows what Final Cut Pro X will do.

Many thanks

Lance,
Can you explain your statement that Final Cut Pro "editing Uncompressed won't allow maintain the quality through the editing process"?
FC works full Uncompressed (444 time-line) and when rendering required does it at 32b FP.
Can you point to the problem?
rafael

Lance, FCP has since version 5 been able to edit uncompressed HD footage. When you purchase the capture device the drivers come with it. I have been using a BlackMagic Design Multibridge and as you can see in the attached window capture, the drivers and corresponding settings are all there. I currently use FCP 7.0.3. As mentioned elsewhere in these threads, you will also need a wicked-fast RAID array to edit uncompressed whatever.

Dear Friends,

While we do not have the ability to record compressed in the Gemini 4:4:4 we offer a solution that will work for many.

After one has recorded the full uncompressed footage, one can attach the SSD to our transfer station, which is included the our Gemini 4:4:4 kit for $5,995.

Then the footage is available as a mounted drive, an SSD with a 6.0 Gigabits per second interface.

The footage can then be encoded to the codec of your choice using Apple Compressor or Adobe Media Encoder or other software.

For 30 minutes of footage, this encoding occurs in about 15 minutes.

Thus, we use the general purpose computing power of your computer to do the encoding.

If we add the ability to record to multiple codecs in our recorder, the power draw will increase by over 10 watts, and the price will also have to increase.

Thus, we offer two solutions.

The nanoFlash, which allows compressed recording for long events (over 10 hours), and the Gemini 4:4:4 for very high quality recording (but this is limited at this time to 2 hours maximum, unless one wants to "Hot Swap" then the uninterrupted recording time is limited by your resources (number of SSD cards).

If you check the power requirements of all other uncompressed + compressed recorders you will find the power draw to be significantly higher. Ours is 8 to 15 watts, with around 12 watts being typical.

I wish on a star that in the future - there could be an added feature- the ability to also bring record compressed off the bat. 22minutes (or whatever precisely it is) of recording is not going to cut it for me when I am doing my work. The Ki Pro Mini gives me about 44 minutes on a 64gb compactflash card and I'm feeling like that's my limit.

It's rather unfortuate because I want to get the Gemini but I just don't know how useful it would be for my doc/tv work.

I wish you guys still make a 10 bit recorder as well as efficient as the nanoflash but with the ability to record to Avid DNXhd and Pro Res. For when you need more than the nanoflash but not as much as the Gemini.

I have faith in you guys!
Ed

Dear Friends,

We try hard to respond to our customers' requests.

We are listening.

I wish on a star that in the future - there could be an added feature- the ability to also bring record compressed off the bat. 22minutes (or whatever precisely it is) of recording is not going to cut it for me when I am doing my work. The Ki Pro Mini gives me about 44 minutes on a 64gb compactflash card and I'm feeling like that's my limit.

It's rather unfortuate because I want to get the Gemini but I just don't know how useful it would be for my doc/tv work.

I wish you guys still make a 10 bit recorder as well as efficient as the nanoflash but with the ability to record to Avid DNXhd and Pro Res. For when you need more than the nanoflash but not as much as the Gemini.

I have faith in you guys!
Ed

This is exactly why I'm going to have to buy the Sound Devices PIX 240 instead of a Convergent Design product.

Uncompressed is a nice idea but my projects won't benefit from it. Nor can I overlook the cost of media using an uncompressed workflow.

I think the Gemini is aimed fair and square at those that want the very best image quality, nothing less. The PIx 240 is a very nice looking device, a bit on the big side for attaching to a camera for handheld and with much higher power consumption and a big fan on the back. It's somewhere between a Cinedeck and a Ninja or Samurai. The Pix will record ProRes or DNxHD in 10 bit, but it can't do 4:4:4, and if you are planning on using S-Log or any other Log or raw type workflow you really want 4:4:4 as your white balance is fixed in camera to the sensors native response, so you want all the RGB data you can get.

I love the convergent design concept of shooting uncompressed and then encoding on transfer. It does mean that you will need a few nice big SSD's, but as someone used to buying SxS cards for my EX cameras the cost per minute of the media is not completely outrageous and will get cheaper over time. With the encode on ingest approach I can choose the codec for each project without having to change the recording device each time. I might have a cinema commercial shoot one day requiring uncompressed 4:4:4 and then the next day a simple corporate where maybe Cineform or ProRes would be better.

I was asked a month or so ago about my opinions regarding the new field recorders coming out. I would like to share my thoughts here since they are directly related to the equipment under discussion.

Understand these are my thoughts on the subject.

============================

I'll put the summary first. In a single sentence - I'm not interested in this or any uncompressed ONLY product.

Do I think it will sell? Absolutely. I think there to be two types of buyers. The first will be those that truly need what uncompressed has to offer and second will be those that want to brag they shoot uncompressed. The bravado of the second group will stop when they figure out that they can't afford an uncompressed workflow (in time, aggravation, or actual $$). If I pick up a project that needs to capture in uncompressed I'll snag a recorder secondhand from group 2.

1. I'm sure its a GREAT piece of gear that will offer the highest quality possible.

2. That level of quality and its demands are not for everyone and not offering a compressed workflow IN THE RECORDER is a serious oversight in my opinion.

3. The true cost of ownership is not obvious until you start buying RAID systems for your edit systems and LOTS of SSDs for your field capture OR adopt a PROXY workflow. Certainly manageable but more complex depending on your editing platform.

The long version:
1. This one is pretty self explanatory. Quality. I know they make a quality product. I have total respect for the Convergent Design team and know they put out an awesome product.

2. How good is good enough? Simply - I'm not a measurebater and not interested in over collecting data. What it comes down to is this - You should capture the CORRECT amount of data for your application so your end result has the quality attributes you desire. As example - If you are going to shoot a green screen epic such as "Lord of the Rings" then full uncompressed 4.4.4 makes a LOT of sense because it can improve the quality of your keys in post and give you the most latitude in the image to tweak it mercilessly. That amount of data has a purpose. The downside of this workflow is storage size and bandwidth and cost of capture media. If you don't have a RAID setup that reaches into double digit terabytes and a REAL RAID controller card in your computer that gives lots of bandwidth to those drives this workflow will only frustrate you with dropped frames and piss poor playback performance. You have the option to create compressed proxies for doing your editing but that is rendering time on your machine and time for someone to organize all your proxies in the edit system. Not a pain free workflow but if you have a "Lord of the Rings" budget then this doesn't even come into question. You got the cash to make this a non-issue.

From that level we drop into a much larger population of films where 4.2.2 10bit compressed S-log gamma is perfectly good enough to use (and in many cases maybe overkill). This covers just about everything else that doesn't fit into the paragraph above. Riddle me this - Why do I want to shoot 4.4.4 10bit uncompressed video of a tracking shot of a couple of lovers walking in the park talking to each other? Does that make any sense at all? In my book it doesn't. What does make perfect sense to me is this - If you want to use S-log gamma then get a 10bit recorder that supports reasonable compression and if you don't need something as strong as S-log then drop back to an 8bit recorder or use the cameras internal recording media. The media and computer demands fall substantially thanks to responsible compression and the post workflow is helped by being able to work perfectly well with your average multi-core desktop computer.

3. Cost of ownership and use - Wow $1000+/hr for capture media? Makes Sony's SxS cards look a lot more reasonable at a cost of about $350/hr @ 35mbit. I acknowledge that the cost of SSDs will continue to come down BUT we are not there yet. Here is an example using a real project I personally worked on. We captured about 2hrs of footage per day on an Indie feature I shot last year. That means for each 3 day weekend shoot we would have to have about $6,000 in media available. If we wanted to keep the field capture as a safety backup till the end of the shoot then we would have had more than $30k in media alone. That was more than the entire shooting budget for the film. Add the cost using an editor that can handle uncompressed your available choices are reduced and your costs are increased again. An F3 rental may be in reach of many Indie filmmakers budget wise but an uncompressed workflow isn't.

More isn't always better. Sometimes its just more. More headache, more expensive, more time consuming. Plan ahead and use the right tool for the job. These uncompressed tools will be fantastic for a small portion of productions and a headache for the rest.

Dear Chris,

For now, I just want to address Point 3.

We expect that our Gemini 4:4:4 users will either:

1. Copy the footage off the SSD to one or more Hard Disk Drives on a frequent basis so that the SSD can be reused. We expect that this will be done in the field, and we recommend copying to at least two HDD's.

2. Encode to a compressed codec of their choice, using the SSD as the source, and one of their internal or external HDD's as the destination.

This can be done in the field, or other locations.

Option 1 takes approximately one-third real time, 20 minutes for 60 minutes of footage, for one copy.

Option 2 Takes approximately one-half real time, 30 minutes for 60 minutes of footage, for one codec.
And the codec can be almost any codec, it is your choice.

These times are based on tests that we ran months ago.
We will have updated times as soon as we rerun these tests.

Chris, it is very clear that you want either a compressed recorder, or a full uncompressed recorder that also has an option to recorded compressed, if desired.

We are listening to your requests.

We do not expect that most users, except for major Hollywood studios to have enough SSD's for a weekend of shooting.

Please note that for recording S-Log, we do recommend recording 4:4:4 as opposed to 4:2:2, if possible. This recommendation comes from outside sources as well as one of our team members.

Chris, please feel free to call me if you wish.

Support Contacts | Convergent Design (http://www.convergent-design.com/Support/SupportContacts/tabid/1767/Default.aspx)

Thanks Dan,

Please don't take my comments as negative on the product you make. Nothing is further from the truth. You guys make AWESOME stuff. It was a pleasure to meet the team at NAB this year on the floor and during the DVInfo get together.

My concern with jumping to uncompressed in general is in using a bigger tool for the job than is necessary. Specifically in this case thinking that more is always better or that more will benefit every production. That just is not true or should I say that it may be true from a pure technical sense but far from true in a practical sense. I only advocate that people consider their application and understand the total costs/benefits before making the jump.

I am VERY interested in knowing the specifics as to why 4.4.4 be used with S-Log vs 4.2.2. The 10bit thing is obvious but why 4.2.2 colorspace is a problem is not so obvious.

Dear Chris,

Thank you for your kind words.

We take no offense whatsoever. Your point of view is one that we certainly respect.

Background:

Today there are some amazing cameras.

The Sony F3 and The ARRI Alexa are two examples.

To get the most out of the Sony F3, one can consider using CineGamma's, or S-Log.

With S-Log, to get the most, one should use a very capable recorder.

We started designing the Gemini 4:4:4 prior to learning of the "to be announced" F3.

It was wonderful how the two were almost made for each other.


And I want to mention, as I have before, if one is not going to use S-Log, the nanoFlash has been proven to produce just stunning images with the F3.


The Gemini 4:4:4 also has a place with the ARRI Alexa.

Both before and after we come out with an extra cost option to support ARRIRAW.

I am VERY interested in knowing the specifics as to why 4.4.4 be used with S-Log vs 4.2.2. The 10bit thing is obvious but why 4.2.2 colorspace is a problem is not so obvious.

Hi Chris, I've been doing some recording with a Gemini prototype and can add my two cents...

Since the F3 S-Log mode doesn't allow electronic white balancing or color correcting beyond either the 3200K or 5600K CC choices, you will always be doing a grade on the S-Log imagery, both for contrast and for color balance. 444 should give you the best possible chance of doing these wider balance adjustments.

Secondly, the F3 might be the first camera with a 444 option, but it won't be the last. While there is debate on how valid the ultimate 444 of the F3 is, since it starts from relatively lower pixel count bayer source, no question each future camera will build on this until 444 is a pretty standard and required method of recording. My nanoFlash from Convergent has kept my Sony EX1 "happy" and very current for the almost 4 years I've had it... longer than ANY other video camera I've owned. I'm certain the Gemini will keep the F3 as current as the nanoFlash has kept the EX1 for as many years, and be ready for any other future 444 cameras.

Finally, uncompressed is NOT the way I'll edit. But, and this is the important part, the original .dpx files are a way to put off any decision making ABOUT post until you're IN post. You can delay any decision to compromise the imagery the camera can make until you are ready to edit, or your client has decided how it will be posted.

Regards, Jim Arthurs

Secondly, the F3 might be the first camera with a 444 option

The Sony F900 and the Sony 1500R with the CineAlta option have been 444 capable for quite a while... (Both are 3 chip, and hence no bayer filter).

Billy

Correct, and other high end cameras (Viper, etc.)...

...I meant to say "Prosumer", as I still think the F3 qualifies as, based on build quality and form factor... I wonder what next, and from what manufacturer will go this route at equal or lesser price point...

Regards,

Jim Arthurs

There are quite a few reasons why you want to record 4:4:4 from the F3 in particular.

As has been said when using S-Log you can only choose 3200k or 5600k so colour correction will be the norm, so getting every last drop of colour data you can squeeze out of the camera is desirable.

The other issue is that to get 4:2:2 S-Log out of the F3 requires that you use the Monitor Sdi output to feed the recorder. With S-Log enabled you only get 4:4:4 out of the dual link ports, there is no 4:2:2 S-Log from these connectors. While that in itself is not an issue, the quality of the 4:2:2 on the monitor SDI appears to be full quality... the issue is that you then can't use the cameras built in Look Up Tables as these would be applied to the SDI monitor out. So on-set you have no way of seeing the full recording range unless you use an external device that can apply a LUT and this complicates things, especially on location.

Chris mentioned using a proxy workflow as a way to help deal with the large files that the Gemini will create. This is one of the great things about the F3 4:4:4 workflow. You can record the uncompressed 4:4:4 S-Log (or standard gamma) on the Gemini while recording 35Mb/s 4:2:0 proxies in the camera. The proxies can have a Look Up Table applied while you shoot so that the off-line can be done with images that approximate the final look of the programme. Any time coded preview footage can be copied from the in-camera proxies that look half decent, unlike the S-Log which will often put producers and those that don't understand the workflow into a panic believing the shots are incorrectly exposed.

I think Chris has some valid points. The Gemini won't suit everyone, that's why CD will continue to sell the NanoFlash, which is still a class leading device thanks to it's use of the broadcast industry standard XDCAM HD codec. This means you can dump files from the NanoFlash directly onto Optical Disc for long term storage or archive.

I've used uncompressed, ProRes, CineForm, DNxHD over the years. They all perform well. Yes there is a storage overhead for uncompressed, but one thing that does get overlooked is that the processing requirements for uncompressed are lower than for a compressed codec. Provided your data pipe is fat enough, working with uncompressed is actually quite easy and fast unless your working with multiple layers in which case data throughput can become an issue. I'm still trying to figure out what my "real" SSD requirements will be. Yes I will need to spend some money on storage, but just 4 years ago I was spending close to $8000 USD for 25 mins of storage. The original 8Gb SxS cards cost almost that when the EX1 was launched. Today I own enough SxS cards to shoot non stop for 10 hours. I see the same happening with SSD's.
I'm very excited about being able to capture the pristine output from the F3 without added artefacts or encoder noise.

There are quite a few reasons why you want to record 4:4:4 from the F3 in particular.


The other issue is that to get 4:2:2 S-Log out of the F3 requires that you use the Monitor Sdi output to feed the recorder. With S-Log enabled you only get 4:4:4 out of the dual link ports, there is no 4:2:2 S-Log from these connectors. While that in itself is not an issue, the quality of the 4:2:2 on the monitor SDI appears to be full quality... the issue is that you then can't use the cameras built in Look Up Tables as these would be applied to the SDI monitor out. So on-set you have no way of seeing the full recording range unless you use an external device that can apply a LUT and this complicates things, especially on location.



Alister, I would like to offer some info regarding the signal possibilities of the F3.

4:2:2 is available on the A/B SDI ports of the F3 with S-Log enabled. You can either set the output to 4:2:2 @ 1.5G + S-Log in the menu OR if you set the camera to 4:4:4 @ 1.5G + S-Log each SDI port outputs a 4:2:2 signal which when combined equal the 4:4:4. {Edit - I will be checking to see if one half of the 4:4:4 is a valid and recordable 4:2:2 signal. According to the specs in the manual it is but this will be my first opportunity to test it.}

With this you have the flexibility to record 4:4:4 S-Log OR 4:2:2 S-Log to your recorder and still use the internal media + LUT and MONITOR + LUT.

If you select 4:4:4 @ 3G + S-Log you will only have a signal on SDI (A) and no option for 4:2:2 except on the MONITOR port.

I'll be testing out one of the PIX240 recorders this weekend. I'll be happy to demo a Gemini as well when they are available. ;)

...

As has been said when using S-Log you can only choose 3200k or 5600k so colour correction will be the norm, so getting every last drop of colour data you can squeeze out of the camera is desirable.
...

Is this the case w/ all Sony high-end cameras? As it can be quite difficult to accurately dial-in WB in post, esp. when dealing w/ mixed lighting sources.

Peter,

Shooting a grey card will help with that. I have one attached to the back of the slate so when I roll camera I shoot both sides of the slate. The slate also has some color chips along the top. I don't take them as gospel for accuracy but it does help to have something you know is constant to reference in every shot.

This was my good friend shooting film, it's now my good friend with the F3/S-Log combo... roll a few frames on this for every lighting setup change as an industry standard reference...

KODAK: Gray Cards (http://motion.kodak.com/motion/Products/Lab_And_Post_Production/Gray_Card/index.htm#graycardplus)

Regards,

Jim Arthurs

4:2:2 is available on the A/B SDI ports of the F3 with S-Log enabled. You can either set the output to 4:2:2 @ 1.5G + S-Log in the menu OR if you set the camera to 4:4:4 @ 1.5G + S-Log each SDI port outputs a 4:2:2 signal which when combined equal the 4:4:4. {Edit - I will be checking to see if one half of the 4:4:4 is a valid and recordable 4:2:2 signal. According to the specs in the manual it is but this will be my first opportunity to test it.}

Sadly it's not that simple. One half of 4:4:4 RGB is not 4:2:2 YCrCb. You can't just take one leg of a dual link 4:4:4 RGB output and use it as 4:2:2, you will be missing half your colours.

The S-Log output options are:
1.5G RGB 444 & S-Log -- AKA dual link, both A and B used for RGB
3G RGB444 & S-Log -- AKA Single link RGB

All the other options are not S-Log, so the only way to get 4:2:2 YCbCr S-Log out of the camera is via the monitor HDSDi.

The output options are interesting as if you select 1.5G + Video and set the camera to 50i or 60i the output on the A/B connectors is 50/60P.

I called Sound Devices today and they confirmed that they don't do RGB so the PIX240 is not compatible with S-Log from the F3.

That is a disappointment.

RGB only works over dual link or 3G HDSDi so if you want to use the A/B connectors you will need a DualLink or 3G recorder and there are not many choices. For a portable unit there is the The Gemini or Cinedeck, for a luggable device there is the Weisscam DM-2, Codex, Sony R1 or R4 and the DigiDisk (there may be others). All of these are more expensive than the Gemin, all are much larger, all are heavier, some ridiculously so.
If your not in the field then a Mac or PC with a DualLink Blackmagic or AJA card is also an option.

Of course you can still get S-Log 4:2:2 YCbCr out of the Monitor HDSDI output, but as I said you then loose the ability to record proxies with a LUT applied and monitoring with a LUT direct from the camera. Adding a Blackmagic HD-Link would sort out the monitoring LUT's but it's an extra piece of kit to lug around and power etc.

Based on the current state of recorders I've decided to continue to use cinegamas and hold off on purchasing anything till things settle out more.

No doubt with more cameras coming out with these capabilities, recorders will continue to evolve.

Sadly it's not that simple. One half of 4:4:4 RGB is not 4:2:2 YCrCb. You can't just take one leg of a dual link 4:4:4 RGB output and use it as 4:2:2, you will be missing half your colours.

The S-Log output options are:
1.5G RGB 444 & S-Log -- AKA dual link, both A and B used for RGB
3G RGB444 & S-Log -- AKA Single link RGB

All the other options are not S-Log, so the only way to get 4:2:2 YCbCr S-Log out of the camera is via the monitor HDSDi.

The output options are interesting as if you select 1.5G + Video and set the camera to 50i or 60i the output on the A/B connectors is 50/60P.

The exact makeup of the RGB flavor of the interface wasn't easy to find. I finally found a good writeup on it in a whitepaper on the Sony HDCamHR (go figure).

I know this may be slightly OT but I wanted to share what I read and now understand about what the F3 outputs when in RGB mode.

When in RGB Dual Link 4:4:4 mode SDI (A) consists of a full rez "G" and even numbered scan lines of "R" and "B". This is in effect "A" 4:2:2 signal but not a usable one. SDI (B) transmits a 0:2:2 signal that consists of NO chroma and the ODD lines of "R" and "B".

This is a description of the video payload itself and there are lots of other things going down the pipe too such as timing packets and the audio streams.

When in 3G mode the entire signal is sent down one cable exactly as a 4:2:2 signal would be but with the extra addressing in the stream for the extra chroma info.

Anyway, Thanks for the patience guys. I think I'm up to speed on it now.

... and HDCamSR is 12 bit with the extra 2 bit's in a separate stream on "B" AFAIK. The F3 is 10 bit only. The issue is that the 4:2:2 is RGB not the more normal YCbCr. As there is no standard for HDSDI 4:2:2 RGB it is unusable with current devices.

Alister and Dan,

Thank you.

You convinced me to go Gemini. And also the S Log out from the F3 as well did. The workflow won't be too complicated. I will have a DIT on set during Gemini shoots. Who will get a drive (256gb ssd at this point) - and on set convert it to PRO RES 444.

The more and more I shoot with S Log - the more I need to color correct the images better and also this way - I am assuming this is correct - I can use the Gemini to get 60 FPS 1080p out of the Sony F3? Please tell that's true. Because I miss overcranking.

Also I can start to use the Gemini on Red Epic and Arri Alexa shoots as a system to get well for the Red Epic - footage that's already 1080p and good enough for web work ? Correct? For the Alexa, it would give me longer recording times than going to the 32gb card. Again how does that sound?

I love my Nanoflashes - I own 3. They are the most reliable recorders I've ever had and look forward to the Gemini.

Until I did some tests with the Gemini on Sunday, I had assumed the best way to work with F3 S-Log was to simply get used to the way the flat S-Log looks under different conditions and do a mental adjustment. I think it's easy enough for us shooter types to figure that out. Of course, add a client, or art director, or anyone else into the mix and you suddenly need correction.

So, doing some run and gun work, I decided to turn on the F3 LUT (rec709 800%) and see what it would be like to use the LUT in combination with zebras at 100% as the main basis for making exposure decisions for S-Log. I then would glance at the Gemini LCD panel for a "confidence" view of the flat S-Log whenever I was concerned. This proved a great combination.

I found that using the LUT allows a greater deal of predictable headroom, as you make the choice on highlight exposures, assured that you have a good deal of recoverable headroom beyond where your viewfinder/zebra is telling you there is clipping. My analogy is with the painted lines on a highway; You might not need the lines in order to drive down the road, but the lines make it easier to stay in the proper lane, and keep your tires out of the ditch!

Here are some split screen examples, taken on Sunday. Very high contrast, shadows and shade. All at 3200K. Upper left image is the on-board 35Mb/sec with the cameras burned in LUT. Upper right is the un-doctored S-Log. Bottom left is a grade to show what picture information is still available in the image high lights. And Bottom right is a more or less "normal" grade...

http://ftp.datausa.com/imageshoppe/outgoing/S-LOG/PEA_QUAD_VIEW.jpg
http://ftp.datausa.com/imageshoppe/outgoing/S-LOG/SHEEP_QUAD_VIEW.jpg
http://ftp.datausa.com/imageshoppe/outgoing/S-LOG/STABLE_QUAD_VIEW.jpg

And here is a zip of the Gemini's .dpx files from each if you want to play with S-Log...

http://ftp.datausa.com/imageshoppe/outgoing/S-LOG/GEMINI_DPX_EXAMPLES.zip

The Gemini will let you take this one step more... add video village to the mix. You will be able to set up a completely different LUT for the Gemini's HD-SDI or HDMI output, giving you a three point viewing method;

-On camera LUT tailored to exposure
-Flat S-Log "confidence" on the Gemini monitor
-A conventional "beauty" viewing LUT for Video Village.

All of this from a single package the size of a SmallHD DP6, powered by the camera's Swit-style battery with a D-Tap that records 4:4:4.

Regards,

Jim Arthurs

Sadly it's not that simple. One half of 4:4:4 RGB is not 4:2:2 YCrCb. You can't just take one leg of a dual link 4:4:4 RGB output and use it as 4:2:2, you will be missing half your colours.

The S-Log output options are:
1.5G RGB 444 & S-Log -- AKA dual link, both A and B used for RGB
3G RGB444 & S-Log -- AKA Single link RGB

All the other options are not S-Log, so the only way to get 4:2:2 YCbCr S-Log out of the camera is via the monitor HDSDi.

The output options are interesting as if you select 1.5G + Video and set the camera to 50i or 60i the output on the A/B connectors is 50/60P.

I don't have time now to read the manual, so I don't know if there's an RGB<->YCbCr conversion built in as well, but for converting 3G (or dual-link) to/from "1G/1.5G" you can check out this offering from AJA. I also didn't check out how much it costs...

3GM Product Page (http://www.aja.com/products/converters/converters-hd-3gm.php)
3GM Product Manual (pdf) (http://www.aja.com/pdf/support/manuals_conv/AJA_manual_3GM.pdf)

Billy

The AJA box is certainly an option for taking the 3G or dual link RGB and converting it to single link 4:2:2 YCbCr. I wonder how good the colour space conversion is?
It's another box to attach somewhere, more cables and power. I'd rather take the non-LUT 4:2:2 out from the monitor out direct from the camera to the recorder. Then feed the Dual Link or 3G into a Blackmagic HDLinkPro box and apply the viewing LUT in the HDLink. That way I have the bare minimum of devices, cables and connectors between camera and recorder yet still have a monitor out with LUT, plus I can feed both an HDSDI and HDMI or DVI monitor from the HDLink.

Either way it's still a lot more complex and fiddly than a dedicated 4:4:4 recorder.

To state the obvious, a new CD machineis wanted : one that can compress 4:2:2 or 4:4:4 and write to an SSD. Hopefully, CD is already working on this.

The flash solution becomes expensive quickly as top quality flash that can record 60 minutes of 280 Mbps I-Frame is about $1000/128 GB (Sandisk Extreme Pro). I recently shot a sporting event with EX1 = Nano and covered it in 50 Mbps MPEG-2 LongGOP because of the flash cost issue. I'd gladly pay $500-1000 more for a Nanoflash with SSD storage. There would be some hesitation in buying a device that doesn't handle 1080p60 or whatever the next iteration of HD will include. 720p120?

Dear Gints,

1. Stay Tuned!

2. The Axtremex 128 GB 600x CompactFlash cards are available for $489.95 at nanoFlash.net (http://www.nanoFlash.net).

These have been very thoroughly tested in the nanoFlash.

Dear Dan,

Can you give an indication of when a radio signal can be expected for the ones staying tuned?

I fully agree with Gints, 1080p60 is what I am looking for, too.

Dan, can you give information about this issue I asked on the FS100 forum: does the FS100 output 1080p60 for in case external recorders arrive that can deal with it?

Dear Cees,

We are aiming for next week.

I am not being coy or attempting to hype this; doing this right requires permission and licensing from other companies. We cannot announce until everything is in place. Then we will announce our plans.

From reading the Sony FS100 manual it seems like 1080p50/1080p60 is available from the HDMI output.
But, we have not tested this, so this is a great question for Sony.

Since this is thread is a request for a compressed codec, please note that recording 1080p60 in any of the commercially available codecs is a major task.

This is one of the things that one does not really know if it is possible until one actually does it.

.

dear Dan, Thanks for your clear reply.

2. The Axtremex 128 GB 600x CompactFlash cards are available for $489.95 at nanoFlash.net (http://www.nanoFlash.net).


Thanks, Dan. I may spring for those, but your site states the 128GB version tops out at 220 Mbps.

Dear Gints,

The Axtremex 128 GB CompactFlash cards are about to support 280 Mpbs in the nanoFlash.

Please allow about two weeks for this to come true.