DV Info Net

DV Info Net (https://www.dvinfo.net/forum/)
-   Open DV Discussion (https://www.dvinfo.net/forum/open-dv-discussion/)
-   -   Trying to get best VHS quality. (https://www.dvinfo.net/forum/open-dv-discussion/16501-trying-get-best-vhs-quality.html)

Paul Doss October 31st, 2003 05:24 PM

Trying to get best VHS quality.
 
Working with captured VHS tapes through Canopus ADVC-100. Printing back to VHS tape through ADVC-100.

Since digital uses a larger screen area, when you fill that area with footage captured from VHS isn't it being stretched? If you then go back out to VHS, aren't there less pixels being used?

If the answer to these questions is yes wouldn't you get better resolution if you slightly reduced your frame size in order to keep as many pexels as possible?

Mike Rehmus October 31st, 2003 07:30 PM

Understand that analog video has no pixels. It does have lines but then the number of lines are a constant (sort of) regardless of format. The signal along a line is a changing voltage level.

So what you end up doing is sampling each line at the rate supported by the format. When you put it back out to tape, the digital to analog converter changes it back to an analog signal that potentially has more detail (contains higher frequencies) than the VHS system can record. So it does its best and there you are.

Glenn Chan October 31st, 2003 08:31 PM

Paul, are you confusing this with the overscan area of your TV? All TVs will crop off some of the sides and the top of the picture.

Paul Doss October 31st, 2003 09:38 PM

Mike,
I understand that analog video is comprised of lines. Within each line there is a certain amount of information by the voltage rise and fall. There are so many lines = x amount of total information.

Glenn,
I'm not confusing it with overscan but asking if you take the above mentioned "x amount of information" and spred it out over the digital image area including the overscan and then lose the overscan area. Have you lost some potintial resolution.

What got me thinking about this was noticing that pips seem to be so much sharper and better looking. I got to wondering if part of that is because you are using more, if not all of the original information.

Let's pretend we are scaning in an image of say 9 x 12 inches and the image is actually made up of 400 horizontal lines. If you printed that image out at 9 x 12 but framed it in a 6 x 9 frame. You not only can't see 1/3 of the image but you've lost detail that you would have have bee had you reduced the image size to 6 x 9 and printed that out. All of the lines are being used so all of the available information is there.

Maybe I'm way off base here but I know that taking 2 similar clips and making one into a pip makes the full size clip look bad. Just trying to figure out why.

Ron Evans November 5th, 2003 07:41 PM

The active picture area (width by height )involved for digital and analogue can be assumed to be the same, at least as far as the viewer is concerned. The difference is in the capabilities of each system to record detail. A VHS tape system has a hard time recording detail equivalent to 250 horizontal lines, SVHS and Hi8 about 400lines and DV about 500. If these systems are fed inputs at 230 lines they will all record and playback the 230 line images. Feed them 450 lines and the VHS will give you 240 ( or so, you will loose the detail) the SVHS and Hi8 between 380 and 400 and the DV 450. The horizontal and vertical sizes are all the same but the frequency response for analogue systems and sample rate for digital systems governs the detail that is seen. The systems always record at their capabilities whether or not there is anything to record!! When you feed in your VHS to the computer it samples at a rate that could reproduce 500 lines at 55db S/N however the feed from vhs is only 230 at less that 40db. When you play it back out almost nothing will have changed. There will be technically small changes due to the encode and decode that you will find close to impossible to see. In terms of your radio its a bit like listening to the same tune on Am or Fm Stereo. The song is definately the same, lasts the same length of time but there is a definate difference in the quality of the signal. This is the frequency response and for video this can be translated directly into lines of resolution . Same tune, same artist, same length big diference in quaility. Another way to look at it is by considering how an LCD TV screen shows the image. Say there are 720 across by 480 high pixels. Regardless of whether this screen is feed a 230 line image or a 500 line image the picture will have to fill the screen. For the low resolution image blocks of pixels will appear the same and for a higher resolution image the pixels either respond individually or the pixel groups responding the same way are smaller. Imagine the VHS picture overlayed on the fine detail pixels of the LCD. It means a lot of the pixels are responding in the same way retaining the VHS detail level. The playback electronics translates the input to provide the neccessary coverage ( decode ). As a computer screen it is different, feed it a 640 by 480 picture and that is what you get. However play a movie and ask the playback software to go full screen and you get exactly what a TV does for you. You get all the image spread over the screen. The ultimate is to ask the PC to do this with a very low resolution image and the screen becomes a mass of blocks when the decode is unable to create a usable image at the screen size.
I hope I haven't totally confused you!!

Ron Evans

Paul Doss November 6th, 2003 08:26 AM

Ron,

Thank you for the reply.

Glenn Chan November 6th, 2003 07:12 PM

Quote:

Let's pretend we are scaning in an image of say 9 x 12 inches and the image is actually made up of 400 horizontal lines. If you printed that image out at 9 x 12 but framed it in a 6 x 9 frame. You not only can't see 1/3 of the image but you've lost detail that you would have have bee had you reduced the image size to 6 x 9 and printed that out. All of the lines are being used so all of the available information is there.
If you resized the 9X12 at the best theoretical quality then the 6X9 image should have 1.5X400 or 600 horizontal lines. The higher resolution image will look sharper. But in this case the resolution is so high that humans will have a difficult time telling the 2 apart at normal viewing distances (assuming 1 horizontal line = 1dpi). Also you'd need a printer that can reproduce that much resolution.

If you play computer games you'll know that the tiny thumbnail screenshots on the box of the game look gorgeous. The actual game's graphics may be disappointing though because your computer screen is so much bigger than that tiny thumbnail.

Quote:

Maybe I'm way off base here but I know that taking 2 similar clips and making one into a pip makes the full size clip look bad. Just trying to figure out why.
Resizing the VHS source to a tiny window increases its resolution by a lot, and it also may have some side effects like reducing noise and crap around edges.

DV however has a maximum resolution based on its pixels and how you do the resizing. The maximum resolution of DV (on the computer) is 540 lines. Resizing is not always efficient and you can lose some resolution in the resizing process. When you resize down you usually don't have to worry about the resizing algorithms used, just the limitations of the DV format and aesthetics (it makes other footage look bad).

Paul Doss November 7th, 2003 03:03 PM

Glenn,

Thank you for the post. I really liked seeing "just the limitations of the DV format and aesthetics (it makes other footage look bad)." I now know that I'm not completely crazy.


All times are GMT -6. The time now is 05:47 PM.

DV Info Net -- Real Names, Real People, Real Info!
1998-2024 The Digital Video Information Network