The strange case of Star Wars Rai Tre broadcast…

In 1997, George Lucas decided to release a Special Edition of the Star Wars Trilogy (at the time, Episodes 4, 5, 6); the “ultimate version” (even if we all know, now, that it was not the truth…).

So, every subsequent TV broadcasts of Star Wars probably used the 1997 SE master, until the 2004 DVDs eventually repaced it. Until now, only few recordings of this 1997 SE version appeared on the net; they are Continue reading

Restoration for dummies

Many people would like to start a fan restoration project – some have good ideas, I must admit – but almost no one has patience, time and skill to do one in person; they keep asking someone else to do it for them.

So, it’s time to reveal an hidden truth, feared by all of us, fan restoration project makers… the truth is…

EVERYONE COULD MAKE A RESTORATION WITHOUT EXPERIENCE!

you just need few things – no computer, fancy software, 4k monitors, steep learning curves… Continue reading

Super Resolution

Some links to interesting threads and articles about Super Resolution:

http://forum.fanres.com/showthread.php?tid=275 (free registration needed to read it)

http://originaltrilogy.com/forum/topic.cfm/Star-Wars-GOUT-in-HD-using-super-resolution-algorithm/topic/17552

http://originaltrilogy.com/forum/topic.cfm/Star-Wars-97SE-in-HD-using-super-resolution-more/topic/17781

Restoration tips: Axis-Aligned Polygon Aspect Ratio™

AXIS-ALIGNED POLYGON ASPECT RATIO™

What is the Axis-Aligned Polygon Aspect Ratio™ (or AAP A/R)?

It is an unconventional aspect ratio to represent a motion picture, using sources taken from different media, in an axis-aligned polygon enclosed inside a conventional aspect ratio screen.

How does it work?

If a motion picture is displayed in a screen with a different aspect ratio than the original – e.g. a film on tv, a 4:3 show on a 16:9 tv – there are different methods to do it. Of course, if the aspect ratio of both screens are the same, nothing will (should) be done, and the motion picture will be displayed “as is”. Continue reading

Restoration tips: Andrea’s Corollary to the Kush Gauge™

prologue: DON’T (always) TRUST COMPANIES!

Following many HDTV and internet broadcaster advices, a “studio quality” transmission for an H.264 1080p transmission could be achieved with a bitrate of a mere 6mbps; if we use the Kush Gauge formula, we can see that this is true only if the motion factor is lower than medium… for 16/9 sport material at 29.97fps, for example, a 17.4mbps bitrate is needed!

 

ANDREA’S COROLLARY TO THE KUSH GAUGE

To calculate video bitrate for a codec different from H.264, the Kush Gauge costant value should be changed accordingly to the codec used.

Continue reading

Restoration tips: Kush Gauge™

KUSH GAUGE™ -> new Kush Gauge calculator

What is the Kush Gauge™?

It’s a rule of thumb to calculate the needed bitrate for H.264 encoded video; it was written by Kush Amerasinghe, a computer scientist. In this context, the word “gauge” means “a device used to make measurements

How does it work?

Quoting Kush’s document:

to estimate the optimal H.264 bit rate value that would give what is considered “good quality” results for a given video, you could multiply the target pixel count by the frame rate; then multiply the result by a factor of 1, 2 or 4, depending on its motion rank; and then multiply that result by 0.07 (the constant, Ed.) to get the bit rate in bps (divide that by 1,000 to get a kbps estimate or by 1,000,000 to get a Mbps estimate).

The Kush Gauge™ formula:

Continue reading

Restoration tips: Overlap matching

OVERLAP MATCHING

What is the Overlap matching?

When two images of different sizes are used to improve the final result.

Could you be more specific?

Usually this is used when two different versions of the same movie are available at different aspect ratios.

Let’s say you want to restore a movie that is not available in high definition, and you want to upscale a DVD or a laserdisc capture. If you are lucky, an anamorphic DVD is available, so a theoretical 720×576 (PAL) or 720×480 (NTSC) max resolution could be achievable. But, if the original aspect ratio of the movie is 2.35:1, then you have only 720×432 (PAL) or 720×360 (NTSC). Well, enough resolution left to do a decent upscale. However, if the DVD is not anamorphically enhanced, actual resolution drops to 720×324 (PAL) or 720×252 (NTSC).

But what if the only (or the best) available low definition sources are analog? In this case, a 2.35:1 original resolution in laserdisc could be at maximum 576×324 (PAL) or 576×262 (NTSC), while VHS could be 328×324 (PAL) or 328×262 (NTSC)… pretty poor…

So, how it’s possible to improve the quality of such low resolution sources? Continue reading

Restoration tips: the Slice Technique™

Restoration tips: the “Slice technique”

THE SLICE TECHNIQUE™

What’s about this so-called “Slice Technique™”?

In few words: achieve the widest (highest) image possible adding a “slice” of another video clip.

Be more specific…

Well, sometimes there are two editions of the same movie, where one has more image on one side, while less on the other side, and vice versa. So I thought to use the missing “slice” of one joined together with all the other, to obtain the widest image possible. Indeed, it could be only (e.g.) 3% wider than any each version, but I much prefer a 100% wide image Vs a “mere” 97%… (^^,)

Of course, it’s quite difficult – sometimes impossible – to use this technique with every movie, because what must be taken in account are the dimensions (eventual image rotation, width and heigth of sources and chosen slice, eventual resizing and cropping, exact point of matching), colors (both versions should have the same color grading, or one should color match the other), video quality (grain, resolution/definition, different compression used in the sources, frame whobbling in a version not present in the other – in particular at the beginning and ending of a shot), etc.

In these examples, I used a vertical slice and two sources, Continue reading

AviSynth and VirtualDub – speed improvement

Recently my old video card die… it is (was) a GeForce GTL 7600 (a custom made version by Sony, the same specs of a GT 7600, but a bit slower and with a very good pipe cooling system). It was the default video card in my Sony HTPC PC – a VGX-XL202 – and I never thought to replace it because, even if it was slow and old and with only 256MB, it was capable to play Blu-ray without stuttering (with some codecs and setting tweaks) and because AviSynth and VirtualDub use only CPU for their processing…

Or, at least, it was what I thought!

Just today I replaced it with a brand new Sapphire (AMD) Radeon HD 5450; a low profile 1GB DDR3 “honest” video card. I chose to buy this because I need a video card ASAP, it’s cheap and, as I have other two PCs, I’ll buy a better one later when I’ll upgrade another CPU.

Well, as I was working onto one of my project the day the card died, I promptly re-run the conversion, and I noted one thing: conversion speed is improved, A LOT! Now it runs at about double speed!

So, it seems that AviSynth and/or VirtualDub DO USE GPU power too, or the old GPU simply slowed down the CPU speed?

If it’s true that GPU power is important for conversion speed as the CPU, my next upgrade (CPU from Core 2 Duo E6300 1.86GHz to Core 2 Quad Q6700 2.67GHz, and GPU from GeForce GTL 7600 256MB to AMD Radeon HD7750 1/2GB) will improve speed up to 2.5x/3x thanks only to CPU; how much the video card will improve speed is to be discovered…

If someone had some similar experience, please post here… if only I knew this, I would have replaced my old video card before… years before!

PaNup™: or, how to upscale PAL + NTSC capture and live (quite) happy…

Today I was thinkering with the “pyramid” LD capture, and, when I was looking for the files, I encountered an old test script for upscaling, abandoned two years ago. Idea was good, so I managed to improve it – as now I have more knowledge than at that time – and results are quite good.

The idea is simple: take two captures of the same content, one PAL, one NTSC, then merge them to “squeeze” every bit of details from it… everyone knows that a PAL capture has 576 horizontal lines, while NTSC has 480… now, when a movie was transferred to video, usually original resolution was higher than that (not all the times, but often); so, the PAL and NTSC “received” different lines of image… see the next image (obviously intentionally exagerated…):


Continue reading