Log in

View Full Version : How do releasing companies change 4:3 old trashy movies into clear 16:9


Halliwell
5th December 2008, 20:15
I have seen that alot of the old video nasties have been released these days as fantastic clear widescreen format copies.

I have heard so many people say that you can't convert 4:3 into 16:9 but I was wondering how it's done because they seem to be able to do it.

And also how do they remove all the grain darkness and dirt that seems to plague old movies?

Thanks

Flux
5th December 2008, 20:53
Most movies are recorded on the film. They simply scan the film again and make an 16:9 version from the movie or wider depending how wide the original aspect ratio is. Usually original one has 2.35:1 ratio or even wider. Televisions with 4:3 ratio ruled the world for tens of years. They probably thought that 4:3 version was fine...until now.

Another way (bad way) is to cut some of the picture to create wanted aspect ratio. Even if film resolution is enough to avoid blur, some parts from the picture is not visible anymore.

Technically one can convert 4:3 to any aspect ratio. Practically it is not reasonable because picture looks weird with wrong aspect ratio.

I don't know details about dirt removing but they probably do it "by hand" with some image editor? Frame by frame?

Blue_MiSfit
5th December 2008, 20:57
Well, they go back to the source :)

In most cases, there's a film print around, somewhere! Film is almost always some wide aspect ratio, be it 2.35, 2.4, 1.78 etc, so you can get the full frame to fit in a 2k or 1080p scan.

So, you throw it on a telecine, and scan it, at high resolution! Then you can go into a professional color correction / cleanup workflow to bring the film back to life, and encode the sucker into a nice master.

In my opinion, it's best to leave as much of the film's original texture as possible. Of course, if the film is literally dirty, it can be cleaned before scanning, but I don't like most noise reduction / de-graining. BluRay has enough bitrate to faithfully reproduce all but the most extreme cases of grain. If you look around on the AviSynth forums here, you will see that extremely high quality denoising and degraining is quite possible. Unfortunately, it's dog slow for HD content - at least via AviSynth methods. In an encoding house, folks usually have access to very expensive, hardware based (real-time) image processing devices. These can do any number of things, but they're quite good at doing upscaling and denoising / sharpening. Still, I prefer to leave the look and feel of the original film! Grain is good!

Anyway, this grand master can then be duplicated and transcoded as necessary to meet the final distribution format requirements.

For example, my company gets high bitrate MPEG-2 or MJPG sources from studios via an intermediate encoding house that usually takes an HDCAM or D5 master and produces a digital mezzanine file which goes to us (via secure internet connection or physical delivery) and the studio's archives for backfill. We then take these mezz files and encode them using H.264 into various SD or HD formats for ingest into our streaming video network.

A very similar situation happens when BluRay or DVDs are being produced. The only difference is this time they bring together all the extra material, build menus, and produce a master image which goes out for distribution.

That's kinda how it works :)

~MiSfit

Dr.Khron
8th December 2008, 16:42
Now that we've beaten movies to death.... What about TV?

Now that TV content is digital, do TV networks "re-encode" digital streams? What about channels that provide both HD and SD streams?

My first thought was that maybe this is all done by the cable companies, but then I realized that the networks are responsible for their own digital terrestrial broadcasts.

As for movies on TV, I suppose that the movie studio does all of the encoding/processing before sending a finished file to the TV broadcasters?

CWR03
8th December 2008, 21:30
As for movies on TV, I suppose that the movie studio does all of the encoding/processing before sending a finished file to the TV broadcasters?
I would think so, because even some movies need to be done "pan and scan" and fixed for 16:9 fullscreen playback, otherwise people will complain about the black bars on their TVs (They STILL make fullscreen 4:3 DVD versions of 2.35:1 movies, for God's sake).

I would be interested to know how TV shows that are shot on digital media are made ready for HD broadcast.

Mug Funky
9th December 2008, 04:49
for old movies:

- someone finds an archival copy of the original film (best case is the internegative used to produce the release prints, or an interpos, or an AB roll, or even the original camera negative if it's still to be found and in a good enough condition to be used)

- the film gets assessed for condition, cleaned (3 runs in an ultrasonic cleaner should get rid of most of the crap) and possibly re-spliced if needed. depending on the damage a new, wet-gated copy might also be made to conceal scratches in the film base. emulsion scratches have to be removed digitally though.

- it gets scanned (usually 2k, ie 2048x1556 for super35 4-perf, 1828x1556 for academy. sometimes they might want to wank off and go 4k at 4096*xxxx but it's not really worth it)

- the parts get assembled into one edit - all the best quality bits of the film that could be found are conformed to match the original cut. this process can take a long time depending on how lost all the bits are, or it can happen immediately if a copy of the interneg was found (as that's the finished film in one hit).

- it gets graded, usually supervised by people who worked on it originally if they're still alive. director and cinematographer preferred. they all sit in a nice room with a very good reference projector and make the film's colour and contrast sit the way they remembered it, or do something new with it depending on how faithful they want to be.

while the grading takes place, digital de-spotting happens underneath. the dirty frames are replaced by clean ones (backups are kept of course). techniques and software used vary depending on the damage. photoshop is often needed for splice marks, or registration problems or whatever. de-spotting is usually done by anyone in the place who is able to operate the software. automatic tools don't cut it, but can help greatly if there's the ability to check and correct it's results. basically it is long, boring frame-by-frame work, and you'll be seeing white spots in your dreams if you do it too long.

- sound is fixed separately, again from whatever is the best source available. noise reduction, dynamics, and some re-editing are probably done. dunno - i'm not a sound guy.

- when the grading and de-spotting is done, it's all "onlined", ie titles re-done if necessary, a few names or dedications added to the credits (ie "dedicated to such-and-such who died in 19xx", or "restoration by these awesome individuals"), etc.

- deliverables are spat out of the online system - dpx files on LTO-3 tapes for archiving (digital film files for making new prints), digibeta for standard def use (TV, DVD), HDCAM SR or D5 for high-def use, etc.


old TV shows go through a much less sophisticated process, as all the above steps are very expensive and usually charged per hour.

[edit]

oh, btw, for TV broadcast usually digibeta or HDCAM/SR tapes are sent to the TV station, and they handle the encode side themselves. either dumping everything onto a server or streaming it live from the output of the tape deck into an encoder. Ads are a bit different - in Australia at least they are distributed online from ad distributors (basically a large dub house with a good internet connection) as MPEG-2 standard def, high profile 422 at 25mbps. the TV stations receive this and do whatever they want with it. no ads in HD here yet, except for 1 which was a test at channel 9.