64t clip on You-Tube. Nice stuff!

[super8man]

Cool - you can see the blues turning into contrasty white instead of subtle tones like in the top picture....just above the heads is where I am looking. Also, the faces are more contrasty in the lower picture with the skin tone more stark on the face on the right (on the lower picture). The top picture is more pleasant but that would be mixing metaphors.

[mattias]

that doesn't sound right. are you watching the two above each other on an lcd screen by any chance? try putting them side by side. you're not supposed to notice any difference in color and contrast whatsoever, had there been any i would have done something wrong. digibeta and dv use the exact same color space.

/matt

[Scotness]

digibeta and dv use the exact same color space.

/matt

Okay - but doesn't digibeta and betaSP have more lines of resolution that they sample at before the drop it down to (for example) standard pal 720x576.

I thought they sampled at 600 lines - so in you're oversampling and should get a better result than miniDV when comparing motion as opposed to comparing colour in stills.

But then of course that benefit would only apply to areas of the picture where the colour range is high - because if adjacent colour elements are too similar you won't be able to distinguish them

And of course oversampling at Super 8 won't have too much of an effect if the original isn't fine grained - but it will be of much more benefit to 16mm and above

blah blah blah resolution arguements! ](*,)

[mattias]

doesn't digibeta and betaSP have more lines of resolution that they sample at before the drop it down to (for example) standard pal 720x576.

?!? of course not.

btw i'm curious why you chose that quote from my post. color space has nothing to do with resolution. it refers to which colors can be represented and reproduced. it's interesting though because the color resolution of dv is actually only half of digibeta's, so subconsciously maybe you're on to something? ;-)

/matt

[Scotness]

Hi Matt - I chose that quote to say "on the other hand" - or yes you're right about the colour but what about the lines of resolution though - does that give digibeta an edge in a different way?

And actually colour does have a bit to do with resolution - if resolution (or resolving power) is defined as the ability to tell one part as seperate to another then the ability to present those parts (pixels/grains/molecules whatever) in a broader range of colours will increase the ability to be able to tell them apart - or to see them as different to each other.



Scot

[mattias]

what about the lines of resolution though - does that give digibeta an edge in a different way?

um, the answers was no a few hours ago, and surprise: it still is. :-)

And actually colour does have a bit to do with resolution

oh, but i never meant to suggest otherwise. i said color *space*, and the point was that they are the same, thus it's not an issue.

/matt

[mattias]

anyway just look at the images. one is digibeta quality and one is dv, from the same high quality source. that's all you really need to know.

/matt

[Scotness]

what about the lines of resolution though - does that give digibeta an edge in a different way?

um, the answers was no a few hours ago, and surprise: it still is.

And actually colour does have a bit to do with resolution

oh, but i never meant to suggest otherwise. i said color *space*, and the point was that they are the same, thus it's not an issue.

/matt

Actually I think we're both right - if the colour space is the same - then having greater line resolution (whether real or apparent due to oversampling) will be of benefit only when there is a great range of closely positioned highly diverse colours in the image - because otherwise the extra line resolution wouldn't be apparent because the colour depth wouldn't be sufficient for the viewer to be able to distinguish between different coloured adjacent pixels.

So something with little colour range like the sky will appear to have lower resolution than something with alot of closely packed variation like the feathers on a parrot.

anyway just look at the images. one is digibeta quality and one is dv, from the same high quality source. that's all you really need to know.

/matt

I agree the images look almost identical - but we would have to see motion examples to really get a better idea - something like in the examples I mentioned above would be interesting



Scot

[mattias]

Actually I think we're both right

actually, no. you just don't understand what color space means. look it up.

/matt

[mattias]

I agree the images look almost identical - but we would have to see motion examples to really get a better idea

you'll see the same thing, but here goes. from a tv pilot i shot a while ago. the first is straight from the digibeta master to mpeg-2, the other has been converted to dv, then to mpeg-2. you can definitely see a difference in a side by side, but to connect with the original premise of this thread it doesn't really change the quality. sharpness is the same, colors are the same, there's just a slight "difference" in the image structure due to an extra generation of dct compression.

http://www.mattias.nu/stuff/lalaunc.m2v
http://www.mattias.nu/stuff/laladv.m2v

/matt

[Scotness]

Actually I think we're both right

actually, no. you just don't understand what color space means. look it up.

/matt

I know what colour space means - and they both use the same colour space - didn't I agree with you on that about 6 or 7 posts ago? What I'm getting at is the apparent improvement because the digibeta is essentially oversampled in the camera to begin with before being output as a standard definition signal.

If you re-read my statement

having greater line resolution (whether real or apparent due to oversampling) will be of benefit only when there is a great range of closely positioned highly diverse colours in the image - because otherwise the extra line resolution wouldn't be apparent because the colour depth wouldn't be sufficient for the viewer to be able to distinguish between different coloured adjacent pixels.

I'm saying in some circumstances it will appear that digibeta is better, in others it won't. It depends upon the detail and colour variance in the (orignal) clip or scene - regardless of the colour space used - this is because of the oversampled original.

It was probably confusing of me and unecessary to bring in colour the way I did in that earlier description. Probably what I should have said is:

having greater line resolution (whether real or apparent due to oversampling) will be of benefit only when there is a great range of closely positioned highly diverse colours in the original image - because otherwise the extra line resolution wouldn't be apparent because the colour variance in the original image won't have been enough to make apparent any extra detail that was captured by the oversampling of the original image

It's like saying if you film a flatly painted and lit wall and then film a whole lot of different coloured flowers - all the same focal distance, f-stop lighting plan etc - the flowers will appear to be much higher resolution than the wall.

I'll have to look at the .m2v's later - can't do that now on the PC I'm on

Scot

[mattias]

digibeta is essentially oversampled

no. (is anyone counting?)

I'm saying in some circumstances it will appear that digibeta is better, in others it won't

of course. it's not the fact that you're wrong about, it's the reasons and implications.

this is because of the oversampled original.

it can't be, because it's not.

captured by the oversampling of the original image

you really have to explain to me, let's pretend i'm five years old or something, what you're talking about. first you ask if digibeta uses "oversampling" and when i say no you start bulding an entire case as if my answer had been the complete opposite.

/matt

[super8man]

I am looking on a samsung 20inch lcd, yes. Let me fgo back and look again...perhaps lcd reveals the true nature of the photos? Strange but I still stand by my original statement. No big deal either way.

[Scotness]

To put it simply:

A) If you have more lines of resolution (either real or apparent via the benefits of oversampling) then in order to see those extra lines you have to make them look different

B) The only way you can make them look different is by having adjacent lines (or sections of them ie. pixels) different colours. They can be different colours by two methods:

C) The colour space or depth being so great that minute differences in colour that normally wouldn't be presented can be presented -- but this isn't the case here as they both have the same colour space - and neither has the advantage here

D) So it falls down to the actual image you are trying to capture - if it naturally has alot of colour variance (like flowers) the chances are those lines will look different and you will notice more detail. If the original image has little colour variance (like a flat wall) then the chances are adjacent lines (or pixels) will look similar and therefor will look like a thicker line - or something shot at slightly lower resolution.

So as you say the whole arguement depends on whether or not digibeta oversamples. You say it doesn't - I thought it did. I'm sure I read somewhere that it has 600 lines of resolution which it samples down to 576 to create the SD PAL image. I can't find anything to support that on the net so maybe I am wrong, but if I'm right then surely my logic above is correct.

Scot

[mattias]

perhaps lcd reveals the true nature of the photos?

no, i've checked the histograms and they are identical. the problem with lcd's are that they have different contrast between the top and the bottom of the screen, depending on viewing angle.

/matt