Never mind HD, is next year going to be the year of RAW?

[Simon Wyndham]

So, first we had the Viper Filmstream, then some DVX guys hacked apart their cameras to come up with the Reelstream, then we started to get the Dalsa, the Kinetta, and now it appears the Red camera.

This to me is far more exciting than just the development of high def.

I've got a feeling that the format wars won't settle for a while. I think that, like many of the better DSLR's, that the option of outputting a raw signal direct from the CCD's will become more and more standard on pro grade cameras. It makes perfect sense that we, the users, should decide what format we want to record our images in, and not be bound by what the manufacturers make us, or being restricted by whatever silly compression schemes and colour sampling are in the compressor.

Just as nobody has to think about what format their DSLR takes pictures in, it is about time video cameras went the same way. No more format arguments, no more HD/SD arguments. We just select the resolution we want to record video in and go and shoot something. The artistry gets put back into moving pictures.

I've got my fingers well and truly crossed that Red actually happens.

[Georg Liigand]

You have a good point. Maybe SD, HD and other formats will really continue being on the market together for many years and appropriate format will be chosen by users depending on what kind of results they need. It could be that there will be no such format which eats out the others (many think that HD will do it).

[Simon Wyndham]

I think that there is a gathering momentum that pro users at least will become less and less tolerant of being restricted by compression schemes and low colour sampling just so that manufacturers can seperate their lines.

As storage space and speeds increase there is less and less reasons why low colour sampling and high compression is needed.

[Glenn Chan]

IMO, compression schemes can be a good thing because it lowers the cost of production. Look at miniDV for example... it's really cheap and really easy to edit. All you need is a $25 firewire card and you can already capture your footage. 5:1 DCT compression and 4:1:1 color sampling (effectively about 10:1 compression??) means you don't need to setup a RAID to edit your video.

I think what worked for miniDV was:
A- Excellent balance of cost versus quality. At normal viewing distances, the 4:1:1 color sampling isn't problematic except in very rare cases (highly saturated object on black background). DCT compression is hard to notice. It seems that miniDV is as cheap as possible while still barely giving excellent image quality.
B- Low cost / economies of scale, since miniDV is also a consumer format.
C- Transfer with native compression, which means no generation loss. It also makes the footage easier to edit because it doesn't take up ridiculous gobs of hard drive space.
This is unlike HDCAM, which you can't transfer with native compression.
D- Widespread adoption means better support. There's a wide range of stable codecs and editing programs to choose from. Since it's a standardized format, it reduces the amount of format chaos around.

What did not work for miniDV was that manufacturers didn't want to cannibalize their more expensive products. So that's partly why you don't see really nice miniDV cameras (i.e. 2/3" CCDs). Also, the perception of miniDV as a consumer format meant low sales of high-end miniDV gear (this may be why Panasonic axed the DVC200?). Economics and market perception also play a role here.

2- What I don't like about uncompressed is that you would have to mess around with a RAID. This detracts from the artistry in my opinion. The lack of compression only opens up artistic possibilities in niche situations, like special effects / keying and extreme use of secondary color correction. In nearly every other situation, I would rather have a cheaper compressed format and put the savings towards other elements of production.

What I'd like to see is something like DVCPRO100 recorded onto hard drive (and in an ideal world, Panasonic would open up that codec to everyone). Plug the camera into the computer and simply copy the files over and you're done "capturing".
Fast, cheap, good quality. If you think in terms of the production triangle, uncompressed offers slightly better quality with big penalties to fast and cheap.

3- In future formats, I would probably like to see better support for multiple channels of sound. i.e. Have 2 analog and 2/4 digital inputs on the camera. Most people may just use the two channels. But for the people who want to record four channels of high quality sync sound, now they can. This way you don't have to mess around with syncing sound in post, saving yourself some time.

If you record on hard drive and use MXF wrappers, I believe this is possible?

[Ash Greyson]

We are talking video, not still pictures here. The workflow for uncompressed for SD is cumbersome and expensive. For anything complex it requires network rendering. I dont think RAW video will ever be for anyone but huge production companies. I am just tackling a project shot on a Varicam that we are editing in DVCproHD and we had to invest $10K+ just in a fibre raid to keep our workflow reasonable.



ash =o)

[Simon Wyndham]

Ahh, but you've missed the point. I'm talking about having a camera that gives you ALL the choices. A system that lets YOU decide when to shoot raw or in uncompressed SD, or whether to use a form of high compression.

All cameras could have the ability to output a much higher quality signal. But the manufacturers choose to restrict it.

Now, with high def things get much more data intensive. But the point is to develop cameras that give a choice. The Infinity camera for example gives a choice of recording formats. But more than anything I want cameras to have a standardised format that is widely accepted.

At the moment it is all very well for manufacturers to have all the different systems for the cameras. But at the end of the day look at all the decks that would need to be purchased.

Still cameras on the other hand have absolute standards. I want a video camera that has a selection between 4:2:2 50mbits/s SD, 4:2:2 SD uncompressed, 4:2:2 100mbits/s HD, 4:2:2 HD, and 4:4:4 RAW versions of both. Just a simple selection on a menu.

So it isn't a case of the formats being cumbersome. Its a case of letting the person using the camera how they want to use it. I believe much of the restriction to be artificial currently. With various other methods of recording data such as portable hard drives, Bluray etc the restriction of making capable tape recording mechanisms becomes a mute point.

[Dan Euritt]

what you are proposing would require the payment of standards licensing fees and extra encoding/decoding hardware, all of which could be eliminated if the camera generated raw footage, or maybe jpeg2000... and nothing else.

let the end user do the job of encoding it to whatever format they specifically need... that way, people won't have to pay for camera options that they don't need.

[Simon Wyndham]

Licensing to who? Not if an open source format was developed. The Infinity camera is already moving in that direction. (that uses JPEG2000 as well as a few others).

You talk about encoding to whatever format is needed, but that is cumbersome. One person wants a DVCpro master, another wants Digibeta, yet another wants HDCAM. All those different decks! If one open source format was developed that could use a multitude of resolutions there would only be a need for ONE deck. Different broadcasters wouldn't all demand different master formats from one another, and overall costs will be substantially reduced.

[Dan Euritt]

Originally Posted by Simon Wyndham

I want a video camera that has a selection between 4:2:2 50mbits/s SD, 4:2:2 SD uncompressed, 4:2:2 100mbits/s HD, 4:2:2 HD.

what formats were you planning on putting that data in? you'll have to pay panasonic or whoever, if you want the camera output to be compatible with the editing systems that are out there now.

the industry has evolved to the point where the recording deck is typically part of the camera... what you are proposing would be a step backwards, although i can see certain advantages to it.

[Simon Wyndham]

Like I said, an OPEN format. One that could be agreed upon as a standard. NO licenses. The Infinity camera already uses JPEG2000 which is along these lines. There is no problem with doing this, so I'm not sure where all the 'problems' people are talking about are coming from!

The camera won't need a tape deck. ALL the camera manufacturers are gradually moving over to completely tapeless solutions. A tape deck on the camera doesn't even come into it.

I don't know how an open source format camera that can switch between various HD and SD standards could be a step backwards. IMHO it would be a giant leap forwards and get rid of all incompatibility issues straight off the bat.

Right now we have MiniDV, HDV (in at least 3 different flavours), Digibeta, DVCpro25, DVCpro50, DVCproHD, DVCAM, IMX, SX, HDCAM, HDCAMSR and on it goes.

There is simply no need for all of these different systems other than different manufacturers trying to get oneupmanship on one another in the market.

The way I see it video can be broken down into six most widely used types. 4:2:2 50 Mbits/s, 4:2:2 uncompressed, or lossless compressed, HD versions of these, plus 4:4:4 raw.

An open source format could handle all of those simply and easily. JPEG2000 already does it in fact, and the Infinity camera already uses it. All we need now is for the other manufacturers to follow suit. Where's the drawback in this?

[Keith Wakeham]

The economies of design is an interesting topic. If you include more codecs that means more electronics, chips, larger circuit boards, more power consumpution, etc. If you stick with one then you are forced to alienate current workflows and setups and sacrifice compatibility for a technical advantage.

So, what do you do? Companies are learning to compromise. The HVX and the XL H1 are amazing examples of this compromise. XL H1, a HDV camera, but more people here talk about its HD-SDI that puts it in the class of a HD POV camera that is still have the cost of something like the HDL-40. The HVX has variable framerates and resolutions, interlaced, progressive, and several formats. Most people are still concerned more about the DVCPRO HD @ 24fps then DVCPRO or even the fact that it can do plain jane miniDV. But those things are still their so they can fit current workflows.

HDV is lucky, because Mpeg 2 is DCT and temporal dependant, but the Mpeg 2 processor can do the DCT required for miniDV. But if the formats were so different that different codecs were needed then you would see that as a feature. Just check out XDCAM, their are two, one that does DVCAM and one that does DVCAM and IMX, but the two format one is several thousand more.

So what this all means is that its all a trade off. When they build a camera they have to look at workflow, and even more so with innovative things. Raw can be great if you don't mind doing a lot of work on footage, but sometimes its easier and desired to not have to do things like processing to debayer and colour correction.

[Simon Wyndham]

I can foresee JPEG2000 becoming a widely accepted format in NLE's across the board. So there would be no sacrifice in workflow at all. None.

As I mentioned, the Infinity camera is already approaching this in a similar way to the one I have suggested. It doesn't have a raw output, but it does allow recording to an open source format, JPEG2000.

THe SDI out on the HVX and XLH1 are okay, but I don't want to be carrying around an external recording device. I want the camera to record the footage, not a third party add on.

The impression I am getting from peoples reaction is that they actually prefer awkwardness, the need to purchase loads of extra equipment to gain better quality etc.

If there was really a concern about revolutions in workflow manufacturers wouldn't be introducing ever newer formats for NLE's to impliment.

Let me put it this way, there was far more risk for Sony to introduce XDCAM and for Panasonic to introduce P2 than there is by making a camera that records to an open source format that has industry wide NLE support.

[Dan Euritt]

an "OPEN format" that you have not outlined any sort of logistical solutions for... you are talking pie-in-the-sky theory, but you haven't outlined any sort of interface connection or spec between the head end of the camera and a non-existant recording device... or even how an unspecified signal should be output to an unspecified recording device that uses an unspecified codec.

talking about editing with a vaporware format like jpeg2000 might be a step in the right direction, however :-) because at least we now have a format with a name... the next step would be to convince all the hardware and software manufacturers to include support for it.

[Simon Wyndham]

I think you'll find that JPEG2000 is not vapourware. There is a camera that already uses it. It is called the Infinity camera and is made by Thomson Grass Valley, and is currently being field tested by people as we speak. It records to REVpro media. So that camera for one isn't pie in the sky. Its a reality.

But all of this is not an issue anyway because the whole jist of this thread was precisely about idealism. Did I say I was going to design a camera right here? No. I suggested an idea and that is all. I don't need to justify an idea by outlining every single technical in and out for technology that might not have even been invented yet!

Given all the vapour speak that goes on around here about various cameras it is pretty funny to see people shooting down a mere idea. Anyone would think I had suggested that cars should have square wheels from now on!

[Ivan Hurtado]

Dan, i follow Simon in this one and he´s completely right in all points.

And don´t worry about the hardware and software companies... Can you imagine what would might happen to Vegas if they don´t to put DVCpro between their codecs? What do you think Canopus or Final Cut or AVID will do if ALL the industry start to speak about JPEG2000 and/or the RED codec? Will they risk their customers by not admitting to edit them?

Industry will follow us, too.