How I plan to upconvert VHS to HD. I need someone to show me what it would look like.

Glenn Chan · July 9th, 2007, 01:16 PM

To add my two cents...

It shouldn't hurt to transfer the VHS tapes onto another medium (e.g. DV, DVD), and hang onto both. You should periodically check that your media is in good condition. And then in 10 years, you should probably migrate your DVD material... for example, just copy it again. Because DVDs aren't going to last that long.

Andris Krastins · July 9th, 2007, 01:34 PM

Quote:

That is true today. But, software will continue to evolve at a rapid rate.
As I stated above, "someday" in the not too distant future, there will be software that will be able to create the missing lines, and upconvert any video to HDV.
So hang onto that old video. Ye of little faith....

I can't agree. I am new to filmmaking, but not computer software.
There are no missing lines. What is original, stays original - the amount of data gathered at a particular time and place.
A computer algorithm can smoothen corners, enhance faded colours, add additional lines to fill blank spaces between A and D, but it can not go back in spacetime and witness that moment when the original data was "written down",
it can't see the tree as it was in May 2001 during a family birthday video,
therefore it can't create the leaves of that tree as it was in May 2001.

Maybe it can create a new tree in the outline of the existing one, but then it will be computer graphics - a new creation, a guess of what was there, not an actual depiction of what was seen by the camera in May 2001.

Steve House · July 9th, 2007, 02:05 PM

Quote:

Originally Posted by David Rice

"garbage in, garbage out"...

That is true today. But, software will continue to evolve at a rapid rate.
As I stated above, "someday" in the not too distant future, there will be software that will be able to create the missing lines, and upconvert any video to HDV.
So hang onto that old video. Ye of little faith....

Creating the missing lines is the easy part, line doublers exist today. But creating the information that should have been in the line instead of just copying or averaging the lines on either side of the new lines are another matter entirely. With 250 lines on a VHS frame and 1080 lines on an HD frame, you have only 1/4 of the information required to really fill in the missing lines with real data.

Glenn Chan · July 9th, 2007, 02:34 PM

It might be possible to present the information in the signal in a more useful manner / "extract" "more" quality out of the image. Since video records images over time, you can do things like temporal de-noising. I also believe you can do motion estimation on the footage to extract higher resolution... though I haven't seen any demos online.

Basically you are presenting the existing information in a more useful way.

2- On a semi-related note, there's some cool things you can do by using information from other frames.
http://www.motiondsp.com/
Image processing to get rid of compression artifacts.

If you want to do those tricks however, it is best that you copy the original as cleanly as possible, WITHOUT upconversion. Image processing like that introduces spurious information (which is not helpful) and can throw away useful information.

Time-base correction can be helpful if your tapes have that problem.

3- There are also tricks to make images look better... not all upconversion techniques are created equally.

Frank Simpson · July 9th, 2007, 03:26 PM

Quote:

Originally Posted by David Rice

"garbage in, garbage out"...

That is true today. But, software will continue to evolve at a rapid rate.
As I stated above, "someday" in the not too distant future, there will be software that will be able to create the missing lines, and upconvert any video to HDV.
So hang onto that old video. Ye of little faith....

I think the point you're missing is the difference between "image enhancement" and "resolution".

Resolution cannot be accurately changed after the moment of capture. ANY change in resolution results in the computer being asked to guess.

As marvelous as computers are (and I know, I spend virtually all of every working day in Photoshop) they cannot make decisions, let alone artistic ones. When I retouch an image I have to choose which areas to resample, at what opacity, orientation, etc., etc., etc..

Unfortunately many people who watch CSI where they zoom in on a single frame of "surveillance video" and push their magic "enhance" button" believe that such technology actually exists in this form. They never pause to consider that the "original" frame is actually a down-sampled version of the "enhanced, finished" frame. Whiz-bang technology for entertainment, but it simply is not real. (Consider that if it was, a lot of crimes would be much more easily solved.)

The computer simply does not know how to paint the missing parts of the picture beyond the algorithms it uses to guess at image enhancement.

Giroud Francois · July 9th, 2007, 03:44 PM

between simple backup of your tape and video restauration there is a large gap.
While you can easily backup to a better media (like DV tape or DVD), applying simple luma/color correction, image restauration require studying the damage (noise, drop, ghosting, colour shift etc..) and applying filters.
the problem is these filters usually can multiply the conversion time to a huge factor (like 100 for 1).
you would easily do that for a 20 sec clip but managing it for hours of video seems a bit too big.
Usually when i need to uprez, i try to copy the original to the best format (uncompressed video), then deinterlace (if needed), then uprez to a at least 1.7 or 2.3 the final size (fixed multiple are useless because, then downsizing would just skip pixels), apply filters, then downsizing to the final size.
virtualub is a great free utility to do that, with lots of filters.
neatvideo allows great noise removal and uprez is nicely done by "video enhancer" http://www.thedeemon.com/VideoEnhancer/
from the economic point of view it is cheap, but the time consumed is terrific.

Robert McGee · July 9th, 2007, 06:46 PM

As long as you keep the original source materal in good condition, your going to have it for future technology to enhance it in anyway possible. You also have to think about this. Several television shows where shot on several video tape formats, in order to avoid the coast of having to have what was then expensive film to video transfers so they could be broadcasted on TV. Many new processes came along that made it easier and less coastly. If you shoot something for big screen you use 35mm film or HD Tape, same is true for television.

When digitizing an anolog source like VHS it's not always about what level of image resolution or sound quality but what video codec are you using, the playback device, and your computer's software and hardware. So you have several choices, whether you want the results to mimic the original or look better then the original. You not always going to get what your after but in the end or going to have something to compare it to.

July 9th, 2007, 01:16 PM	#16
Glenn Chan Inner Circle Join Date: Jun 2003 Location: Toronto, Canada Posts: 4,750	To add my two cents... It shouldn't hurt to transfer the VHS tapes onto another medium (e.g. DV, DVD), and hang onto both. You should periodically check that your media is in good condition. And then in 10 years, you should probably migrate your DVD material... for example, just copy it again. Because DVDs aren't going to last that long.

July 9th, 2007, 02:34 PM	#19
Glenn Chan Inner Circle Join Date: Jun 2003 Location: Toronto, Canada Posts: 4,750	It might be possible to present the information in the signal in a more useful manner / "extract" "more" quality out of the image. Since video records images over time, you can do things like temporal de-noising. I also believe you can do motion estimation on the footage to extract higher resolution... though I haven't seen any demos online. Basically you are presenting the existing information in a more useful way. 2- On a semi-related note, there's some cool things you can do by using information from other frames. http://www.motiondsp.com/ Image processing to get rid of compression artifacts. If you want to do those tricks however, it is best that you copy the original as cleanly as possible, WITHOUT upconversion. Image processing like that introduces spurious information (which is not helpful) and can throw away useful information. Time-base correction can be helpful if your tapes have that problem. 3- There are also tricks to make images look better... not all upconversion techniques are created equally.

July 9th, 2007, 03:44 PM	#21
Giroud Francois Inner Circle Join Date: Feb 2004 Location: switzerland Posts: 2,133	between simple backup of your tape and video restauration there is a large gap. While you can easily backup to a better media (like DV tape or DVD), applying simple luma/color correction, image restauration require studying the damage (noise, drop, ghosting, colour shift etc..) and applying filters. the problem is these filters usually can multiply the conversion time to a huge factor (like 100 for 1). you would easily do that for a 20 sec clip but managing it for hours of video seems a bit too big. Usually when i need to uprez, i try to copy the original to the best format (uncompressed video), then deinterlace (if needed), then uprez to a at least 1.7 or 2.3 the final size (fixed multiple are useless because, then downsizing would just skip pixels), apply filters, then downsizing to the final size. virtualub is a great free utility to do that, with lots of filters. neatvideo allows great noise removal and uprez is nicely done by "video enhancer" http://www.thedeemon.com/VideoEnhancer/ from the economic point of view it is cheap, but the time consumed is terrific.

July 9th, 2007, 06:46 PM	#22
Robert McGee Regular Crew Join Date: May 2007 Location: York, south carolina Posts: 24	As long as you keep the original source materal in good condition, your going to have it for future technology to enhance it in anyway possible. You also have to think about this. Several television shows where shot on several video tape formats, in order to avoid the coast of having to have what was then expensive film to video transfers so they could be broadcasted on TV. Many new processes came along that made it easier and less coastly. If you shoot something for big screen you use 35mm film or HD Tape, same is true for television. When digitizing an anolog source like VHS it's not always about what level of image resolution or sound quality but what video codec are you using, the playback device, and your computer's software and hardware. So you have several choices, whether you want the results to mimic the original or look better then the original. You not always going to get what your after but in the end or going to have something to compare it to.