4K FS700 Odyssey 7Q First footage

lol. It's a brick wall, what we've all been waiting for. Stills photographers and lens junkies rejoice :)

Joking aside, if there was original source it would be quite a nice little resolution test though... source files anyone?

cheers
Paul
 
It's something! Focus was on the walls not the fountain. Adam was right that 4K is a whole new focus game.
 
Downloaded a couple video files from CD and man what a wake up call with my current editing system. Can't playback in Vegas at more then 11fps. If I go forward with CD I'll need to upgrade my computer. :/
 
The Pro Res 4K recording is fantastic. I use two 23" 1080P monitors and therefore can only see half the image at once, but the detail is incredible.

It makes me realise what brilliant options it would give for cropping/pan/scanning within the frame for a 1080P output.
 
Very nice, impressive. The footage seems more (lack of a better word) pristine. I can't wait to get rid of that 8 bit AVCHD noise, not that it really bothered me but boy are we in for a treat. Thanks CD, especially Dan Keaton, respectively ;)
 
According to the EXIF info for the 4k Sample Frame, it was grabbed using the Gemini 444. What's that all about?
 
Probably the Odyssey's firmware is derived from the Gemini's, and they haven't bothered to give the Odyssey its own EXIF information yet.
 
I think Dan already said that the O7 could be developed quite fast because it did share a lot of the Geminis firmware (which is a very good thing)

I've scrutinised the footage and i'm very happy. Both with the O7 and the FS700. The 4K is like lifting a veil on the camera, finally i see lines that are smooth and defined. The diagonals on the building look perfect. Vindicating my choice to stick with the 700.

The 2K? Well, back to how i feel about the 1080p from the camera. I can see the downsampling/aliasing on diagonal lines very easily. I've never been 100% on the 1080 output of the camera for this reason. I don't know what scaling happens inside the camera but it's not ideal. There's some very odd artefacting going on around the water which i must assume could be compression related from the camera (the RAW stream is supposed to be compressed). Of course it could be a DNG problem. After Effects CC doesn't open the files with the right colour balance.

The ProRes file isn't log encoded though, so i don't know what is happening with 12 bit linear -> 10 bit.

So i'm a bit 'meh' about the 2K RAW (nothing to do with the O7) but the 4K is another camera.

With some log encoding on that 10bit, i'd be more than happy just to record 4K ProRes. 60fps is a bonus.

Well done CD, you've moved the camera a level up....

cheers
Paul
 
CD have said the 7Q can record 4K raw but are unable to because of limitation set by Sony. I guess to protect sales of the R5. I wonder if Sony will lift that limitation in the future, after sales plateau?
 

I downloaded footage and also watched it on YouTube. The 4K 4:2:2 looks clean for sure. As for the 2K RAW, I'm not sure what is going on as I don't know much about RAW.

Watching the 2K RAW clip on YouTube, the darks are pretty crushed so it looks fine, but you cannot really tell from YouTube as we all know. However, I downloaded the RAW DNG files, it's really, really noisy in the blacks and somewhat noisy in the mids once you start to raise the exposure. Not sure if that's because of the initial exposure of the clips where the filmer choose to crush the blacks, therefore it cannot be raised that much in post; or the RAW is just really noisy. When I first brought in the RAW clip, it's default look was very dark and the black were really crushed. I had to really raise the exposure to see into the blacks, which became really noisy.

I have a question about RAW since I know very little about shooting with it.

1) You still have to / should expose properly right? Just because you can change all the setting in post, does not mean that if you crush the blacks on the initial shot, you can go back later and completely raised the blacks in After Effects or give it an S-Log type look?? Basically, RAW just gives you a little more slack if you mess up initially?

2) Cinema DNG, is that a compressed form of RAW? Kind of like how the RED where one can set the level of compression on RED RAW?

Well, I'm optimistic and can't wait to see more footage, especially of the 2K RAW. After looking at the data rates on the 7Q link, I'm really hoping the HD 4:2:2 10 bit comes soon because of very large 4K and RAW files sizes. You would be going through hard drives like firewood with that much data.

Also, I saw no mention of 4:2:2 10 bit at 120fps. I know Dan K. had mentioned that as a possibility. I really hope they implement that. If you are reading this Dan, is that still on the cards?
 
Last edited:
How in the world can you watch ANYTHING of value on YouTube? If the intention is to judge the video quality, everything on YouTube is compressed. In order to form any meaningful judgment, one must download the original files and view them either in the Sony RAW Viewer or in an application like Resolve(for the .dng files)
 
How in the world can you watch ANYTHING of value on YouTube? If the intention is to judge the video quality, everything on YouTube is compressed. In order to form any meaningful judgment, one must download the original files and view them either in the Sony RAW Viewer or in an application like Resolve(for the .dng files)

Before you slam me Cuervo, read my post carefully.

I wrote:

"Watching the 2K RAW clip on YouTube, the darks are pretty crushed so it looks fine, but you CANNOT really tell from YouTube as we all know. However, I DOWNLOADED the RAW DNG files."

i then stated my observations based on the DNG files, NOT YouTube.
 
Last edited:
I have a question about RAW since I know very little about shooting with it.
Here's the simple answer: ALL cameras shoot raw. They all do; they all start with a raw scan of the sensor. Then, all raw images need some sort of processing done to them (gamma curves, etc) before they become pleasing to look at. Traditionally cameras have done that with in-camera processing; when recording raw that means that the camera's processing is skipped and you instead have to substitute your own processing in post.

1) You still have to / should expose properly right?
Most definitely. There is no "magic" in raw.

Just because you can change all the setting in post, does not mean that if you crush the blacks on the initial shot, you can go back later and completely raised the blacks in After Effects or give it an S-Log type look??
Correct. If you screw up the shot, the shot is screwed up.

Basically, RAW just gives you a little more slack if you mess up initially?
Not even that. Raw gives you the ability to change the decisions afterwards, as to whether you prioritize the shadows more, or you want a different curve, etc. It doesn't really give you "more"; it's possible that the in-camera processing may be doing the most and the best with the footage and an inexperienced colorist may not even be able to make their raw footage look as good as the camera could have, for example.

Where raw gives you more is in that most cameras use compressed codecs that do throw away information (when subsampling down to 4:2:0, for example, or when quantizing down to 8 bits). The sensor may be 12 bits, and so by recording raw you have access to all those 12 bits. When using in-camera processing on an FS700, all the in-camera processing is (likely) done in 12 bits, but it gets chopped down to 8 for recording (or output on the SDI/HDMI ports). By recording raw you have access to those 12 bits.
 
I downloaded footage and also watched it on YouTube. The 4K 4:2:2 looks clean for sure. As for the 2K RAW, I'm not sure what is going on as I don't know much about RAW.

Watching the 2K RAW clip on YouTube, the darks are pretty crushed so it looks fine, but you cannot really tell from YouTube as we all know. However, I downloaded the RAW DNG files, it's really, really noisy in the blacks and somewhat noisy in the mids once you start to raise the exposure. Not sure if that's because of the initial exposure of the clips where the filmer choose to crush the blacks, therefore it cannot be raised that much in post; or the RAW is just really noisy. When I first brought in the RAW clip, it's default look was very dark and the black were really crushed. I had to really raise the exposure to see into the blacks, which became really noisy.

I have a question about RAW since I know very little about shooting with it.

1) You still have to / should expose properly right? Just because you can change all the setting in post, does not mean that if you crush the blacks on the initial shot, you can go back later and completely raised the blacks in After Effects or give it an S-Log type look?? Basically, RAW just gives you a little more slack if you mess up initially?

2) Cinema DNG, is that a compressed form of RAW? Kind of like how the RED where one can set the level of compression on RED RAW?

I looked at those DNG files, too, and it seems like they may have been quite underexposed. It's a bit hard to tell in Adobe Camera RAW since it clearly isn't quite reading the files right. Just to get a normal, correctly exposed image from the 2K DNGs in Adobe Camera RAW I had to push the red channel gamma curve quite a bit and then push the exposure up by at least 1.5 stops and fix the white balance. Nothing is clipping in the images, not even specular highlights on the water, so its a bit hard to tell how far underexposed the image is especially since its really unclear how Adobe Camera RAW is interpreting the data (except that it is clear it is not correct).

1)
As for the differences in RAW, it really depends on the workflow. What do you mean "expose properly"? Of course you should always expose properly, but the proper method may change depending on what you are trying to get out of your images and how the camera works and how your post workflow works. For example I expose quite differently shooting RAW on RED than most other cameras. On RED, the ISO or apparent exposure level is entirely metadata that can be manipulated in camera, so you can always look "properly" exposed no matter what by just adjusting your ISO. If you want to keep the windows from burning out, you can close down so the windows aren't clipping and then increase your ISO until your midtones are where you want them. You can adjust your over/under range and gamma curves manually in camera to suit your scene or your look. If you shoot at ISO1250 you can get around 6 stops of highlight range and even more at higher ISOs. Other cameras do not necessarily work that way.

With most other cameras, when shooting RAW you still monitor with a relatively fixed gamma curve (unless you have a lut box or something) that has a fixed midtone and fixed highlight compression/range, so if your highlights are blowing out and you close down or ND up to protect your highlights, then your midtones may start looking too dark. In that case, what is "proper"? That depends on what you need out of your images. Do you protect your highlights because the colorist can always bring up the midtones in post, or do you set your midtones right where you want them even if it means your highlights will clip because you are monitoring your RAW with the a gamma curve that only has 5 stops of highlight range. I always prefer to shoot images that will work well even if no colorist is ever going to touch the images, which means nailing the midtones and especially skintones right where I want them, but that is tough to do even with RAW if the camera's RAW workflow still depends on displaying RAW video with a limited set of gamma curves that only allow you to have a limited amount of highlight range. A lot of the time, you don't have the control to be able to expose "properly" both for your midtones and your highlights when the two have a fixed relationship on the camera's monitoring or metadata gamma curve. In this way RED's RAW workflow is quite powerful because you have the power to basically grade your images in camera and always shoot optimally without being tied down to one fixed gamma curve. Of course the tradeoff is that RED's approach doesn't do the camera any favors in low light. Workflow is everything, and it remains to be seen how well CD's CinemaDNG files play with various workflows. It will be a real problem if you can't be sure the editor/colorist will be seeing the same image you saw on set, or if people are forced to color correct all of their material because gamma/wb adjustments made in camera are not read properly by post software.

2) Cinema DNG can be compressed, but CD's version is not. The 2k DNG frame is 3.3MB which is exactly the size an uncompressed 12bit 2k bayer file should be (2048pix * 1080pix * 12bpp / 8bitsperbyte = 3.318MB; actual file is a little larger to make room for headers and metadata and such). The 4k Prores may look big at 89MBps, but can you imagine they originally wanted to only do 4k DPX (~950MBps)?
 
IMHO,
People can level all the nit-picky criticism they want against 2k. From what I see, 2k, 12 bit, 422 is such a cut above 1080, 8 bit, 422 that I'm totally awed. Sure 4k is very nice. But, at what cost of storage space and processing pipeline? I'm VERY pleased with 2k! 4K has zero distribution capability with out downrezzing. And I'll wait to see if consumers jump in the 4k display....just like i waited on the 3d displays...which ended up being a flash in the pan fad.

I just saw an ad on the tele in which they placed a 4k display in an office, such that it looked like a window. Then they displayed a meteor striking the earth on the "window". Envision an office full of people panicking...LOL....holey moley!
 
IMHO,
People can level all the nit-picky criticism they want against 2k. From what I see, 2k, 12 bit, 422 is such a cut above 1080, 8 bit, 422 that I'm totally awed. Sure 4k is very nice. But, at what cost of storage space and processing pipeline? I'm VERY pleased with 2k! 4K has zero distribution capability with out downrezzing. And I'll wait to see if consumers jump in the 4k display....just like i waited on the 3d displays...which ended up being a flash in the pan fad.

I just saw an ad on the tele in which they placed a 4k display in an office, such that it looked like a window. Then they displayed a meteor striking the earth on the "window". Envision an office full of people panicking...LOL....holey moley!

As an experiment try downsampling (with a good scaling algorithm, not view at 50%) that 4K to 2K and then comparing 2K to 2K. Look carefully at motion in fine line areas and noise overall. I don't see how anyone can possible say that the 4K isn't a huge improvement over the 2K. Just capturing in 4K doesn't mean you have to work or distribute in it.

So long as we can get log encoding in that 10bit 4K ProRes file i think with the light data rates prores offers and downsampling in post you'd get exceptional 2K imagery.

cheers
Paul
 
Back
Top