(Background on thread.  Debate on if digital can archive the quality of film, especially at low light..  The performance of the new Canon 5DmkII was mentioned)

—–

To ad to Julio’s points.

The camera he is talking about is the Canon 5DmkII.

Currently this camera is making a mess of the digital film camera industry. I know Sony, JVC etc are all running scared.
They are literately selling 5x what they expected on this product. It’s success has changed the industry completely.
This camera makes all there 50-100K kit all seem obsolete. Tho , there is more to a camera then a sensor. (But strangely many people cannot see past that. Ie form factor, the package etc.)

Now, Nikon will likely bring out a competitor at the November show for Pro still cameras. And.. the Nikon has even BETTER low light performance then the canon.

In the Stills world, Film is totally been abandoned apart for the niche areas where some aspects of film still produce desired results that digital cannot. However, this is not common, and a very expensive way forward.

This will of course happen in the motion film area too. I am a little surprised it has taken longer then it has, however, as I have said, there is more to taking a shot then the camera, its the camera man too. This is a generation thing more then anything else.

I have noticed that film-tech seems to be an ivory tower with man holding onto the romance of film. Personally I don;t care. I am a computer tech that has drifted into the doing digital side of exhibition. I just observe and based on what I see, make conclusions.

Film is simply out of date.
I like to compare it to DJing (For which I did for 10 years when I was a younger lad) evolved from vinyl records to CD’s or other forms of digital playback..

At the end of the day, the punters on the dance floor didn’t really care if it was vinyl or CD, the story that was the track was shared and they enjoyed it.

Tho the debate between DJ’s was heated and went on for years. Today, the top DJ’s basically never use vinyl any more.
WHY, well as I like to point out, what sounds better is not really the issue here. The main reason they don;t use vinyl is because vinyl has limitations. While digital has Blue sky. The DJ’s, though the use of digital, have created new playing styles and performance possibilities. Impossible with vinyl.

This holds true for Film as well. Film is dead to me as.. it simply cannot evolve anymore. But once we are digital, the sky is the limit. Night shots that are impossible using film. etc.

One of my favorite would be a new standard of 60fps for cinema. Can be easily converted down to any common distribution format in use today, and would make a film screen, using clean digital images, look like a window into the world that the film creates. Something film grain, scratches and gate movement can never do.. (60fps for film is possible, but like 70mm unrealistic anymore. too $$$$$$$)

So, to me. Film, please leave the building, its time to ignite the booster and leave orbit, and you don;t have a ticket.)

To finish of, I would like to also visit the definition of what make one better then the other.
If we want to talk about which can be better.
Many of you say, film looks better…. if your looking at an answer print compared to 2k. But then again, apple with apples, there are 8k digital systems out there. That would even arguably compare with 70mm.
If we talk about the typical image a widely distributed film presents on screen compared to 2K.
A typical 2K system will ALWAYS look better due to the massive lossy process the film has to go through.

So which is better. What is possible or what is typical??
Please take this into consideration.

James

A recent article on Broadcast TV Faces Struggle to Stay Viable Its a great read, I recommend it.

A point to take from this is that larger productions like ER and friends are no longer viable for that industry. In reality we should consider how this will also be effecting the life blood of the cinema industry. The film makers.

Going digital is to the advantage of the distributors. That is what the VPF is all about. However, the world moves forward and new issues are surfacing. The decision to go digital is no longer a matter of the distributors saving/making more money. We need to consider some new developments.

Production viability

In the article I mention above there is an indication that the production of premium content is becoming harder for TV. This indicates that the big and popular blockbusters, the life blood of cinema exhibition, are also likely to be less viable.

Going digital, at some stage, is going to reduce film distribution costs and in the long run let film makers make better films with less money. In effect making them more viable. Feeding us good premium content that patrons will enjoy coming out to see again and again.

Going Green

Every year thousands of used polyester films are dumped into land fill. These films need large amounts of water and electricity to make while also introducing unfriendly chemicals into the environment. If the cinema industry is to move forward with the times it needs to turn over a new GREEN leaf and move to digital. The use of reusable hard drives could be a great thing to do for our planet.

No excuse

By no means do I make an excuse for the distributors to pay less on a VPF. In reality I do not think they have done enough. However, it is no longer time to sit on the fence to see what happens. If the opportunities arrives it is time to move forward and embrace digital.

P.S. This is taken from the blog www.cinetechgeek.com on the film industry.

As a DCI digital cinema expert and past owner of a Film and TV production studio and facility, unfortunately, this comment is a little miss-leading.

Let me explain why.

I am involved in the installation of DCI projectors within Australia.  As such, I am well informed of technical roadmaps and installation numbers.  Currently the DCI equipment being installed is 2k, general projecting SCOPE as 2048×858 and FLAT (1.85) as 1998×1080.  The DMD, or DLP chips are 2048×1080 in total.

At the current rate, we will not see large penetration of DCI in all cinemas, especially in non-US cinemas, for quite some time to come.  Even in the US the change over is not near 50%.

Current roadmap for 4k DLP support?  Not even being talked about yet.

Life span of a DLP projector.  Expected 10years, however, Cinemas are use to 40 years lifespan on a traditional projector, so expect them to run them into the ground, 15-20 years.

I would expect the world to be fully DCI in about 5-10 years.

If we add all these time frames up, etc.  Don’t expect to see 4K cinema for 20 years. Maybe more.

Now lets look at who can tell the difference between 2k and 4k.  From many people I have spoken to, even trained pros have a hard time telling the difference is at a reasonable distance from the screen.  Sure, if in the front row it is much easier.  In general, however, it is expected that the general cinema goer, 99% would not tell the difference.

As a cinema owner, can you justify spending hundreds of thousands of dollars just to keep 1% a little happier?

What I am trying to point out here is that 2k is pretty much it.  Doing production in more then 2k is a good way to simply spend more money on Post with no real way to get it onto a cinema screen for a very long time to come.

RED1 queasy 4K bayer as opposed to real 4k  (when you say 4k you mean 4k of resolution.  4k baye, DLPr is not 4k resolution.)  the RED1, in a 2k workflow is most likely the best bang for $$$ camera currently available.

RED cry the “Obsolescence Obsolete” tag line, but this is more hype then reality of what is involved in real production.

Cost, destination format, qualified workflow.  This is what you as a producer should be excited about.  Not how many angels can dance on the head of a pin.

SideNOTE:  Over sampling or more then 2k aquisition is very useful for certain post requirements.  Mainly CGI, but in general, it has no bearking to the final result in a 2k workflow.

To gudge this, lets look at the track record.  The RED1.

From talking to people from the prosumer and professional areas, I get two different perspectives.  The Prosumer, wanting the big toys for no money pant RED as the coming of the messiah.  The Pros, who have been using Viper and other uncompressed hi-end systems consider RED a bit of a Joke, but do respect how quickly they have developed an impressive camera.  Still, it has quite a number of problems.

The Pros have some major issues.

1. It over heats under heavy use.  The running joke with the hire companies is it may be cheaper to rent then a uncompressed system, but you need two on set to make sure you have no down time.  Tho I do hear the later revisions  have improved this problem greatly.

2. The colour depth of REDCODE codec is good but uncompressed is better.

3. Its NOT 4K.  Its a good 2K camera.  It uses a bayer filter.  This is not a true pixel for pixel sample.  So please…. stop calling it a 4k camera…  Please goto wikipedia for an explanation.

4. Until recently you could not work natively on redcode-codec files adding a complex addition to the workflow.  I have heard of quite a few major productions having to do neg match to media files as meta data errors and other problems have been occurring on the workflow.

5. Red has not delivered on a number of promised features.  Number of audio track supported.  Support for 4:2:2/4:4:4 single/duel HDSDI out put.  These are major issues to the Pro.

The key to the RED is the sensor.  Who makes the sensor then..  Well that’s a secret.  This is what makes me most nervious of all.

Let me go over some Sensor history.  Many of you may be wondering how RED was able to jump so far ahead in Sensor technology.  Well, this is because it is already a very old and mature industry.  In my career, I have worked for Kodak on the Cineon Project, the first of its kind to bring film into the digital age.  I have been involved in the engineering of a film recorder.  We also planned to make a film scanner, but it’s a bit late for that now.  During this time we went over engineering products suitable for building these tools.  From hi-resolution sensors to line arrays.  Strangely enough these components have been around for quite some time.  There reason for existence.  Robotic eyes.

Very hi-resolution sensors exist, for example to detect imperfections in CHIP manufacture.  These sensors are easily re-purposed into what the RED1 became.  The biggest problem to overcome was that the sensors did not have a data path capable of bringing images of at the speed required for film.  This is the main redesign needed to archive what RED did with RED1.

Sony and other manufactures where starting from scratch as they consider bayer filters as a non-professional way to capture 4:2:2 or 4:4:4 images.  For example, why we have 3 chip cameras, sampling the full colour for every pixel in the sensor.

RED1 4k bayer equates to a little better then 2k at 4:4:4.  So really the RED1 is an amazing 2k camera.  Please don’t call it a 4K camera.  Maybe a “4k-bayer”.

This explains how RED had a perceived jump on the incumbents.  It was not so much a jump as looking at the problem from outside the square.

So really, RED has aligned itself with one of these sensor makers and focused them on making sensors which are more suitable for film acquisition.

The real question here is.  Can RED get this secret company who specialise in sensor development to make sensors to the specifications announced in the recent dream fest?

My personal opinion is that what they have put on paper is possible.  But the time frame arguable.  And now canon and Sony may start playing with the same rules. RED may have some real competition.

Consider this.  If Canon/Sony, with its years of engineering know how, brings out a camera even close to the RED products. Who are you likely to purchase.  RED who are learning as they go along, or Canon/Sony, who really know about building pro kit?

This is simply a reality check for those wraped up in the RED hype. If the RED1 does it for you great.  Canon looks like having kit coming to market soon that will be equivalent in some ways to RED1 and better then the original promis of scarlet..  RED has promised us some vapourware we may see in the future.   But really.  If I cannot touch it, its not worth my or your time.  If your in production, its what you can do today, not what you may be able to do in the future.  Focus on what can be done and your more likely to actually get a prodcution finished.

To help with examining these raw files, I have both cropped a sequence and isolated an image to demonstrate some of the observed issues with the raw files produced by the 5D..

The above SWF file shows a section of the sky from the “bridge shot”.  I have extracted frames from this shot, converted them to uncompressed PNG and placed them in a SWF file.  You’re looking at the raw, uncompressed footage playing back in the Flash window.  Total of 25 frames.  Note the noise and visible macro blocks.

5D example

I took the above section from a frame of a predominantly bright image.  As to be expected, a well lit image looks clean and devoid of “blocking”. But on closer inspection we can see something passing in front of the lens creating the darker to lighter gradient.  As you can easily see, noticeable “blocks” are evident, especially in the blue portion.

Otherwise following is my brief outline and reasoning for this being a revolution in Film acquisition.

As a Still camera, the 5D is a resonable upgrade to compete with Nikon’s new range. However, the Nikon range does still appear to have an advantage in over all still quality.

In the Video area, the 5D looks better then the RED-scarlet (3k imager with bayer filter = less quality then 5D)  RED has withdrawn the current Scarlet specs.  Becuse of the 5D or not, the general still camera industry has surpased the scarlet product and currently I would expect a RED-1 type feature set in the Scarlet body if they plan to still “blow us away” as it is put.

The 5D, from my understanding, can function as a video capture device while at the same time capturing 22mega pixel stills.  This indicates that it is doing a real time sample from 22mega pixel for 35mm sized imager, to 1920×1080.  As such we are geting a large over sampling of the captured image down to 1920×1080 HD resolutions, we will be getting an amazingly clean image.  The 5D also has a new sensor that is more sensitive to light giving you images that look amazing in low light.  This also lets the film maker get away with less lighting (Saves costs) and also get a look usualy only available in much more expensive kit.

The big question is the video codec used and effective quality obtained.  36.5mbit H.264, at double the effective quality as MPEG2, will no dough give us amazing images.  H.264 has also been extended to be able to do 4:2:2, 4:4:4 10bit, 12bit colour and up to 4K image sizes. Will the camera be able to take advantage of this?  Does it now?   Strangely I have not bee able to find any details on the Internet as yet. Even is the  camera does not support these standards yet.. future versions are likely to.

The limitation to 30p (30 frames per second) is most likely more of a marketing issue then a technical one.  After all, this camera is not ideal for a film maker.  The form factor is for a still photographer, not a film/video maker.  I personally expect to see a new Canon XL-video camera in the future that has 5D core technology in the centre with the film makers form factor around it.  Add all the different frame rates and other film/video centric options. (And double the costs of the 5D most likely) and we have the future of Canon Video product line.

If nothing else, this camera indicates that a 35mm sized imager should now be standard for any PRO video camera.  Raw 4k imagers are also old hat.  We now expect over sized images giving us a over sampled result down to 2 or 4k sizes.  These are ideas RED is even struggling with.

RED may have started from scratch with nothing to loose by making a film/video camera taking full advantage of current technology.  However, once Canon and Sony also take this position, canabalising the current product ranges, RED is unlikely to match the experience and knowledge of either of these two companies.

Canon are at those cross roads now.  Sony are likely to get their last as they are the leaders in the top end.

This will be an amazingly popular camera as armature film makers can really justify its costs as it is inexpensive and can also be used as a very good stills camera, which is also an item any Film/video maker should have in his bad.

See Gizmoto review here.

The thing that makes the HG20 special is that it supports a more realistic bitrate for the AVCHD codec.  Most of the recent cameras from Sony and other brand have only supported up to 16mbit.  The Canon unit will support AVCHD up to 24 mbit. This is a huge difference as, the picture quality is greatly improved for shooting more demanding video.  This is more then a typical consumer quality camera.

For example, the new Sony EX1 and EX3, $10,000 (On average) cameras are the hottest prosumer cameras around right now.  And for good reason.  The build and tactile interface is fantastic for one.  Plus the quality of the 35mbit MPEG2 is stunning.  Build quality means nothing unless you can capture a great picture.  MPEG2 at 35mbit does just that.

Now, looking at AVCHD (H.264) at 24mbit.  Yes, it is not as high as 35mbit, but it is well known that the newer technology of H.264 is up to 2 times as efficent.  2×24mbit(H.264) = 48mbit(MPEG2) in quality.

I personally do not thing it is twice as good, but one could expect it to match 35mbit easily.

So we now have $1500 consumer cameras matching the quality of $10,000 prosumer cameras.  Thats big news to anyone who shoots video on a day to day basis.

The other huge advantage is that storing/archiving your content is easier as at 24mbit over, 35mbit, is a major advantage.  This may not sound like a big issue, but video files pile up fast, and it is not unusual to have 10-100 of terrabytes of video.

Standard are very important

Finally, the long term advantage. AVC is a SMPTE specifications which takes it into the future.  For example. support for up to 4k resolutions, 4:2:2, 4:4:4 and 10, 12 bit colour.  No other SMPTE standard codec supports these characteristics.  People talk about the RED all the time using the RED codec.  I find it a sign of ignorance that they do not understand the advantages of supporting standards based codecs.  H.264 at High 4:4:4 12bit 4k is positioned to become tiff/tga of video. A file format anything can read.

An example of this is that DCI (Digital Cinema Initiative) have standardised that a Digital Film Master has to be supplied in 16bit 2K or 4K TIFF files with uncompressed 48kHz 24 bit stems (audio file for each channel, usually 6 (5.1)).  This standard is based on open file formats and not proprietary RedCode or ProRES for example.

The Bad

There is one major issue with AVCHD.  Its not easy to edit.  As it is a complex codec with long GOP (Group of pictures) When a editing system requests a certain image while playing a time line or rendering a timeline, it needs to go to the nearest I-Frame or full frame before the frame it wants, then calculate down to the exact frame.  Very inefficent.  Faster computers with more memory will make this less of an issue in the future, but in real terms, it is usually best to transcode the content to a more editable format whitch also keeps all the quality.  For example, ProRES for Final Cut Pro.  This will, of cause, mean the video content will take up many, many ties the amount of disk space compared to the original.  This is common in editing work flows. For example EXCAM (Sony EX1) and RedCode (RED cameras).

AVC also have other iterations of its codec to make it easy to edit.  An example of this AVC-Intra used by Panasonic in their t hi-enf ENG range of cameras.  This is similar to the old MPEG2-I-frame codecs.  Its just a H.264 I-frame codec.  As it is I-frame only, it runs at much higher bitrates such as 50 and 100mbit.

The Road ahead

AVCHD still has a road ahead.  H.264 is said to have up to 2 times the efficiency then MPEG2.  In my experience I do not find this to be the case.  But like DV and MPEG2 based codec cameras when they where first released, the real time codec implementation was always very version 1.0 and over time improved As such, I expect AVC to also improve in real time camera capturing in the future.

I also wanted to give vimeo a go.  I heard it was a good alternative to YouTube.  And I must admit.  Its very good.  I found it very easy to use and gave great feed back with the status of my video upload and encode.  The final quality is also quite good. They also support HD. My only problem is that I uploaded a MPEG2 8mbit file which I used for the DVD.  As such it has some interlace content.  vimeo did not deinterlace the content.  Doh!

digitAll Show Reel July 2008 from James Gardiner on Vimeo.

They are:

• Verismo Box Plugs Web into Your TV: this is a very good example of the trend towards a open STB with no direct connection to the larger companies.  Ie Microsoft, Apple.  This is intended to be a cheap STB that can utilise as many services as they can get API’s from/for.
• Panasonic Pipes YouTube Directly to TVs: Another example of a large consumer electronics maker making an open STB for general free content on the web.
• Sony To Bring YouTube To HDTV Sets: Sony is getting into the act too.
• R.I.P. SanDisk’s TakeTV, Fanfare Shut Down: A filed quirky attempt.  Not unexpected.
• YouTube Set to Dominate the Boob Tube: A story pointing out how YouTube is partnering with all these general purpose STB and becoming very dominent in this space.
• Apple Should Open Up: More comments from about how AppleTV needs to be open to take off.
• Scoop: Boxee to Go Alpha on Monday: Yet more developments towards open STB with the very well know and popular XBMC (X-Box Media Centre) open source project launching into this space.
• I would also like to mention ADOBE’s release of AMP (Adobe Media Player), all the DRM technology and that Flash Player for embded devices as been made free.  Obviously something is cooking here.

I would like to point out that these stores have all come out in no less then a month.  Thats a lot of action in the STB area.

Still, I find this all very worrying as unlike the past, there is no one driving the development of what I like to compare to as the Analog TV Tuner of past to todays Internet-TV-Tuner of the future.

In the past, organisations such as SMPTE where formed so that the development of TV technologies was based around standards.  Content was structured and produced in a documented way as so it could then be trasnsmitted in a standard way. Any TV manufactuer could simply follow these standards and have a working TV set on the market.

Altho these standards bodies exist and an attempt by them has been made to address the issues of Internet TV, the current state of play is that it is a complete mess.  We have many suppliers using proprietory technologies, trying to leverage their products into a dominent position.

These organisations are very powerful and have, in my opinion, stifeled any movement towards simple open standards for which would allow the adoption of Internet TV to take off.

There is a lot of potential power and money at stake, so no one is willing to “play ball” so to speak.

Fortunatly the technological tuth to all this is that open based STB are inevitable.

For Example.  Lets look at DVD region coding.  This was a way to let the distributors control the release of content around the world.  The movie would be released in the US, for example, under REGION-1.  Here in Australia, Region-6, we would not be able to play that DVD.

Strangely enough, I have never had a problem with this as REGION-FREE DVD players came out very quickly.  REGION-coding today is next to useless.  Consumers eventually learnt about these issues, and today, the consumer electronics stores simply will not stock Region-restricted DVD players are they simply cannot sell them.

The AppleTV for example is already in this bucket.  A consumer who is interested in a AppleTV is as, he is knowledgeable enough to understand that purchasing an AppleTV only really means he can watch shows blessed by Steve, of pod-casts specifically designed for it. (You can hack it to do anything however)

A potential consumer of an AppleTV would understand Torrents, and that alternative content is around, and that obviously, anything he purchases, he will want to also play that content.

This may look to be a trivial matter, but from my experience it is quite the opposite.

The standard TV set of the future is going to need a TV tuner of the future, ie an Internet TV tuner.  This tuner technology needs to be cheap but at the same time as powerful as possible.  I am a big fan of the Adobe mythology as it is possible that each CHANNEL can load its own user interface designed specifically for the target consumer. It can also be easily updated.  Ie the way you use the channel can evolve with the user.  The complexity can evolve with the user.  The Social aspects of the channel can evolve with the evolution of social networks.

The STB of the future still looks like its a fair way down the road. I just hope the incumbent players who are trying to take advantage of this realise sooner then later that we need to get together and head down this road collaboratively soon.  As the longer they procrastinate, the more likely we are to damage the production industry and ultimately the TV shows we love to watch.  At the current rate of development, nether side will end up with very much.

http://www.youtube.com/blog?entry=Em6c795bzc0

This announcement makes Youtube a portal for NEWS by independent reporters.  Social technologies will obviously be used to bring hi-quality and items of social interest to the top of the channel.  It will be interesting to see how well this goes.

This development could have extensive ramifications for all news reporting organizations.