Video size and resolution question..

[zepolli1]

It's probably here somehwere, but what is the difference or definition of video size vs. resolution. For a while I thought they were the same. For example look @ the chart at the bottom of this pagE:

http://www.thenoisereport.com/content-14-page4.html

Sizes range from SIF to D1 and resolutions are 352x480, etc..

From working with computers forever I know what resolution is, but what is video size then?? how does it corelate to resolution?

[mats.hogberg]

In this context, they are one and the same. 1/2 D1 is is 352*576/480 (PAL/NTSC), Full D1 (F-D1 in the chart) is 720*576/480 (PAL/NTSC). So in this chart, Size is only the name for a resolution. (and the actual resolution depends on TV system)

/Mats

[yoda313]

Hello,

https://www.videohelp.com/glossary?R#Resolution

1) A measurement of relative detail of a digital display, typically given in pixels of width and height;

2) the ability of an imaging system to make clearly distinguishable or resolvable the details of an image. This includes spatial resolution (the clarity of a single image), temporal resolution (the clarity of a moving image or moving object), and perceived resolution (the apparent resolution of a display from the observer's point of view). Analog video is often measured as a number of lines of horizontal resolution over the number of scan lines. Digital video is typically measured as a number of horizontal pixels by vertical pixels. Film is typically measured as a number of line pairs per millimeter;

3) the relative detail of any signal, such as an audio or video signal. Also see lines of horizontal resolution.

Hope that helps - it's from this website's glossary

Kevin

[gshelley61]

Size (or perhaps you also meant aspect ratio?) refers the shape and physical dimensions of the video image. In the television world, there are two basic aspect ratios... classic NTSC 4:3 and the newer widescreen 16:9 shape (that's width:height ratio).

Resolution in digital video refers to the number of sampled picture elements (pixels) per image field (or frame in the case of progressive scan formats) that are described and stored in the video file. There are lots of digital video resolution formats. For example, commercial DVD's are 720x480i (interlaced) whether the aspect ratio on playback is 4:3 or 16:9 (an encoding flag in the video file tells your DVD player which it is). There are two ATSC high definition resolution TV formats... 1280x720p (progressive scan) and 1920x1080i (interlaced). Both of these HD formats are 16:9 aspect ratio on playback.

The higher resolution the video format is, the larger the physical display screen size can be before the image starts to look like crap. Low resolution formats like VCD only look OK on small screens. High resolution formats, like HD, can be displayed on much larger screens. Of course, perceived image quality also depends on the viewing distance from the screen... the farther away you are from the screen, the better the image will look. High resolution formats displayed on high resolution screens allow closer viewing distances.

In analog video, the number of scan lines (in NTSC) is 525, of which the viewable image is about 480 interlaced horizontal lines. The apparent resolution then depends on the bandwidth of the video source. Low bandwidth analog video, like VHS, equates to roughly 240x480i in perceived resolution. Higher bandwidth analog video, like laserdisc, Hi8 and S-VHS, are closer to 400x480i.

[zepolli1]

Thanks for the replies..one more..

from the chart..

whats the difference between SIF and 1/2 D1 then? if size and resolution are roughly the same on that chart. Must have something to do with the ratio?

[mats.hogberg]

SIF=VCD=352*240/288
1/2 D1=CVD=352*480/576 (double vert res compared to VCD)

/Mats

[thecoalman]

if size and resolution are roughly the same on that chart. Must have something to do with the ratio?

Here's an example, it's from footage shot in 16:9





Consider this to be what the video looks like before any changes are made to it's aspect. Also note that 4:3 and 16:9 use the same resolution. Notice they appear elongated. This is 1:1 ratio because it's displayed as the image size of 720x480. If the video from this example was shot in 4:3 it wouldn't be very noticeable since it's close 1:1. You video playing software takes the info from the file so it displays correctly. It squishes it vertically..... For 4:3 just a litte bit and for 16:9 just a little bit more.

[Spicuzza]

zepolli1 wrote:

but what is video size then?? how does it corelate to resolution?

Frame size such as 720 pixels x 480 lines, has absolutely nothing to with "resolution" if you are referring to resolution meaning video quality/sharpness.

A video with a 720 x 480 frame size and a bit rate of 1500 kbps would look pathetic compared to the same video with a 352 x 240 frame size @ 4000 kbps.

It is the kbps (bit rate) that will determine the video quality NOT the frame size, however, of course, a large frame size i.e. 720 x 480 together with a high bit rate i.e. 9000 kbps, encoded from a quality source will produce a great looking video.

[thecoalman]

Just to add a little bit more....... I'll use the image again as an example. It's 1:1 right now, let's say it's 4:4 which is the same as saying 1:1.

That gives us 4 divisions horizontally and 4 divisions vertically. If we remove one division vertically and squeeze the image into the remaining 3 divisions the image now has a 4:3 ratio compared to the original.

[trevlac]

This is a favorite topic of mine so I just have to throw my hat into the ring ....

You need to follow the Analog to Digital process thru, and know about the standards to understand how these terms are thrown around ...

Let's do Audio 1st:

Analog

There are 2 things to consider ....

Time - for which we wil just pick 1 second

MaxFrequency - Which is a bit more complex ... have you ever seen an audio equalizer with the LED lights that jump up and down. Well this is a spectrum analyzer which means it took the audio signal and broke out some specific frequencies. The max one is usually around 20k cycles per second aka 20kHz ... this is considered the limit of human hearing ... so there is little reason to worry about higher.

Digital

Here we add a new thing ... samples.

To digially represent analog you take measures at given points of the signal. To capture a given frequency (like 20khz) you need to have a little more than 2 samples per cycle. So 40.1 samples per second can digitally represent a 20khz frequency. To make sure we get the highest and have some headroom in case our equipment is not exact ... give yourself 10% room so go after 22khz and sample it at say 44.1khz.

Now you have this 1 second signal stored in 44.1 samples. If you don't want that many samples you can chop it in half and have 1/2 second of audio that still contains the full high frequency or you could resample to say 22.1khz and loose all frequencies above 11khz but still have 1 second of audio.

On to Video:

Analog

Time - Well this would be picture width in video. There happens to be a standard for digital video that says your TV should take 53.333 microseconds to 'paint' one line of video picture on a 4:3 TV. Without sticking to this constant it is hard to talk about this stuff.

MaxFrequency - The same standard (for standard def video) says the max frequency is a little less than 6.75MHz. For VHS or BroadcastTV or LaserDisk, it is actually lower .... but lower is ok, right? High frequencies show up in video as sharp details.


Digital

Samples (aka Pixels) : Same story as Audio - we need to sample a little more than 2x the highest frequency to keep the details.

So the standard says do 13.5Mhz. And 13.5 * 53.333 = 720 pixels.

Now hopefully you see where we are going .....

If you take a 720 frame and crop it, you could get 640 pixels containing the full details. However, you'd loose some of the picture ... and it probably would not be displayed correctly.

Instead ... you generally want to resample. Resample basically means re-create the full analog signal and then take a new set of samples (pixels). If this new set is less then you loose high frequency details if they were there to begin with.

------------------------
So what is resolution ?

Here is where the 'standards' come into play. If the video in question does not follow the standard ... you have to provide more info to make a comparison. If it does, you can leave out the details and make some assumptions. Also a little knowledge of the amount of detail in the Analog source would be nice.

720 vs 704 vs 352 and a VHS example

Let's say VHS source can contain a Max frequency of 3.2MHz

If you sampled a 53.333 microsecond video line of your VHS at a rate of 13.5MHz ... you get 53.333 * 13.5 = 720 pixels (aka samples). You also have captured frequencies almost up to 6.75MHz ... well above the details in the VHS.

Then if you crop the 720 pixels to 704 you have not reduced the high frequencies but just chopped the edges off of the picture.

Finally, if you resample the 704 down to 352 you have now cut the highest frequencies that were in the original image down by 1/2 .... or at best you could keep ~3.3MHz.

So what tells you resolution ?

Frame width (like 352 or 720) is good enough assuming the source has the max and you know how you got to those values. Strangely enough, 352 is half of 720.

----------------------------------------
A few notes:

- The way you resample matters. Xtra resize(resample) steps are not for free. Different methods cut out different amounts of details ... and some dont cut out any details but turn them into digital noise.

- Most lossy compression is based upon the idea of removing high frequencies that may not be there or that you may not see. So a highly compressed source can have a lower resolution .... but this is not a given ... because it depends upon how many details were there in the 1st place ... so compression as a measure of resolution is like frame size ... you need to know more info to compare.

[zepolli1]

Thanks for all the help. I will try to digest it all !

[lordsmurf]

I disagree on the 704 = cropped 720
It's simply a resize of the pixel dimensions. Some players crop 704, but those are stupid chips. We have LOTS of stupid chips in lots of equipment. At least half of all capture cards are crap for this reason alone.

[trevlac]

I disagree on the 704 = cropped 720

You just want to be disagreeable!

It was just an example. I did not mean to suggest that the only way to get from 720 to 704 was to crop. But if you do resize .... 704 can no longer hold as much detail as 720. If you crop, it can.

Ah .... and I do recall that DVD Demystified states that 704 is a cropped version of 720. But wait! ... now you've got me arguing the minutia again.

BTW: I've been recently goofing with a virtual dub filter that does a spectral analysis on an image. With it you could more or less 'measure' the amount of high frequency details. Really just a work in progress ... but thought maybe someone would like to see? Foo you still out there ?

[lordsmurf]

I gotcha now.

Yeah, 704 has slightly less detail.
Of course, whether you can notice depends on source.

I didn't even notice it was you.
Good to see you around again.