logo Sign In

Post #95276

Author
Moth3r
Parent topic
Letterboxed Widescreen vs. Anamorphic Widescreen Discussion
Link to post in topic
https://originaltrilogy.com/post/id/95276/action/topic#95276
Date created
4-Mar-2005, 5:38 AM
Anyone who saw the sample from my DVD on a.b.starwars will know that I've finally decided to go with an anamorphic transfer (as I had originally planned to do, only for a long while I was considering doing a letterboxed transfer - all the JPEG screenshots were letterboxed).

The main reasons are purely personal:
- Our TV is in 16:9 mode most of the time, so it saves me from having to press a button on the remote control to change to zoom mode.
- We have a portable player with a 16:9 LCD screen, which has no zoom mode.

But I'd like to discuss some common (IMO) misconceptions that have cropped up in other threads recently; rather than add my thoughts to other threads I thought it'd be better to have all 4:3 vs 16:9 discussion in one place.

"You're not gaining anything by making an anamorphic DVD, because the laserdiscs are letterboxed the information isn't there in the first place."
&
"If you have a widescreen 16:9 TV, you can view letterboxed DVDs by setting the TV to zoom mode, the quality is just the same as resizing during post-processing."

Assuming most people have CRT TVs; when you watch a letterboxed film, setting the TV to zoom mode, the TV stretches the picture vertically by increasing the distance between the scanlines. Some 25% (?) of the total No. of lines (480 NTSC/576 PAL) end up in the "overscan" areas off the top and bottom of the screen - but of course these lines only contain black bars anyway.

Now when you process the video to create an anamorphic picture, you use a process called upsampling (or interpolation) to generate additional lines of video. The upsampling algorithms used - bicubic, lanczos, etc. - are much more sophisticated than simply doubling up certain lines or using linear interpolation. The aim is to increase the size without blurring the edges too much or producing jaggies, and these algorithms do this sucessfully (lanczos is said to be sharper, but I don't really notice a difference between bicubic, lanczos, Mitchell, sinc or any of the advanced upsampling filters).

So when you watch a DVD containing a straight letterboxed transfer, you see a stretched out picture. When you watch an DVD with video from the laserdisc converted to anamorphic, you see an upsampled picture. I don't need to tell you which looks better.

For those people with a home cinema setup incorporating a plasma or LCD screen, or when you watch a DVD on your PC, the image will always be upsampled, either by scaling circuitry in the display device or the video card in your PC. In such cases, there should be little or no difference between a letterboxed or anamorphic version.

"If you have a regular 4:3 TV, then you lose 1 out of every 4 lines of resolution when playing an anamorphic DVD."

This is partly true. If you set your DVD player to 4:3 mode, then it will downsample (decimate) the active image and add black bars to the top and bottom. I've no idea if this an actual scaling process, or simply removal of every 4th line.

However, many modern 4:3 TVs now have a 16:9 mode. In this mode, the distance between the scan lines is reduced, compressing the picture vertically. The black bars top and bottom are not part of the video - they are black because there are is no electron beam scanning that area (like your TV is black when you turn it off!). With the DVD player set to 16:9 mode, you still see every line of the anamorphic video, and you don't lose anything.

"You need a higher bitrate for anamorphic video."

Of course you do - there are more lines of video and fewer lines of solid black. Solid black compresses well, actual video information needs more work. (Try this: use notepad to create two text files, exactly the same length, one full of spaces, the other containing random text. Now zip each one. Which is smaller? How much smaller?)

However - and this has been discussed at length by Cowclops - laserdiscs cannot reproduce the same level of detail as a DVD. The picture will always be softer, and a softer picture compresses better than one with lots of sharp details. Unfortunately laserdisc captures contain analogue noise; this noise, because it is random, does not compress well. That is why we must use noise reduction during post-processing before encoding to DVD.

So a de-noised laserdisc capture will compress reasonably well, and this is the reason why Cowclops and myself both consider that single-layer DVD is more than adequate for an anamorphic transfer, with no noticeable increase in quality by using dual-layer. But that's another discussion...

:: Edited for typos