Mike,
To make this a little less indigestible I will respond in two parts. My
responses are interleaved with your questions.
**************
[original message]
A few HDV questions
Posted by: "Mike Boom" [EMAIL PROTECTED] secotioid
Date: Fri Feb 23, 2007 11:33 am ((PST))
Thanks, Julian, for the illuminating explanation of HDV. It's very
well written and very helpful in understanding what's going on in an
oddly defined medium. I have a few more questions if you don't mind:
At 07:16 AM 2/22/2007, Julian Baldwin wrote:
>Regarding interlace: the principle of interlaced pictures was a solution
>devised to overcome two problems with the TV system when it was first
>developed. The problems were a) flicker and b) bandwidth. The human eye
>seems to need a minimum 'flicker rate' of about 50 flickers per second in
>order not to be aware of actual flickering. However the bandwidth needed to
>provide 50 full resolution frames per second was more than the early
>electronics (or the space on the airwaves) could provide so the solution
was
>to interlace the lines and so provide 50 half frames, called fields, per
>second. This halved the bandwidth but maintained an adequate flicker rate.
>When the USA came to develop its TV system the NTSC followed a similar
>principle but used a field rate of 60 fields per second for technical
>reasons related to the North American power line frequency of 60Hz.
I understand the problem with flicker -- trying to use a computer
monitor set to anything under 60 Hz is guaranteed to give me a
headache -- but what I don't understand are camcorders that offer
24p, 25p, or 30p frame rates. Wouldn't these also be very hard to
watch unless they were doubled somehow? Maybe I misunderstand what
these progressive rates really mean, or there's some trick to make
them seem smooth to the human eye, something like the spinning
shutter used on film projectors to double each frame.
[response]
I don't understand the point of being able to display pictures at a refresh
rate of 24p, 30p or (rest of the world) 25p, either. The only reason (apart
from the increased definition - see below) I can think of for shooting at
24p or 25p is to ease the problems of transferring the video material to
film for cinema (theatrical) release. I suppose it could be displayed on a
TV by showing each frame twice but I have no experience of seeing this and
so have no opinion about how satisfactory it would be.
J.
****************************
My other misunderstanding comes from the implication that a 30i frame
rate has a more economical bandwidth than a 30p frame rate at the
same resolution. Wouldn't they be the same? For example, a 1920x1080
uncompressed video stream at 30p would have to transmit 2,073,600
pixels per frame, 30 frames a second for a little over 62 megapixels
per second. A 1920x1080 uncompressed video stream at 30i would chop
each frame up into two fields, each field with half the pixels of a
frame (1,036,800 pixels). But the field rate is double that of the
frame rate, so that's still 2,073,600 pixels per frame, 30 frames a
second, for the same rate of a little over 62 megapixels per second.
Again, maybe I misunderstand how the progressive frame rates work.
[response]
Part of the problem here comes from use of language and, for example, what
is meant by 'at the same resolution'. The first thing to appreciate is that
a progressive picture and an interlaced picture with a given number of lines
(say 1080) do NOT produce the same vertical definition on the display.
The difference lies in the way that the camera reads off the information
from the camera's pick-up chip. Think of that grid of 1920 pixels wide and
1080 high. In the case of the interlaced picture the camera looks at lines 1
and 2 (of the 1080), mixes together ( perhaps 'averages' would be a better
term) the information from pixel 1 of line 1 with pixel 1 of line 2 and
sends out that combined information as the info for pixel 1 of line 1 on the
display. It does the same with pixels 2 to 1920. Having completed the signal
for line 1 it moves down on the pick-up chip.
Now it looks at lines 3 and 4 (of the 1080) and averages the info from pixel
1 of line 3 with pixel 1 of line 4. This forms the information sent to pixel
1 of line 3 of the display. It does the same with pixels 2 to 1920. Note
that we have not created information for line 2 of the display yet, that
comes later.
The camera proceeds down the picture using two camera lines to create one
display line. Having reached the bottom of the picture it returns to the top
where starts again but uses pixel 1 of line 2 averaged with pixel 1 of line
3 to create the information for display line 2. It does the same with pixels
2 to 1920. Having completed the signal for line 2 it moves down on the
pick-up chip.
Now it looks at lines 4 and 5, averages their pixels together individually
as above and creates the signal for line 4 of the display.
The display knows that it is an interlaced system and knows it has to put
these even numbered lines in between the odd numbered lines which it
received on the first pass down the picture.
So the picture is built up but it has taken about 1/30th (1/25th in the rest
of the world) second to do it and each pixel on the original camera picture
has been 'diluted' by being mixed with a neigbour before being transmitted.
In the case of the progressive picture the camera sends the information
about line 1 (of the 1080) to the display to be used for line 1 and then
sends the info about line 2 to be used for line 2 etc. etc. This it needs to
do 60 (or, rest of the world, 50) times per second.
So you can see that the 60 frames/ sec progressive system has sent info
about 1920 x 1080 individual points to the display twice in a 1/30th of a
second whereas the interlaced system has sent info about 1920 x 1080
averaged points to the display only once in the 1/30th second. The data rate
(bandwidth) is therefore half for an interlaced system but the picture is
less sharp.
You may ask why it is desirable to go to all the complication of averaging
adjacent rows of pixels in the interlaced system and the answer is 'to avoid
line twitter'. This is an effect which occurs when a thin line in the
picture only illuminates one row of pixels (say, line 589). Without the
averaging system this line would only appear on the display half as often as
the picture as a whole and would appear to flicker annoyingly at half the
frame rate.
Julian
Bristol, England
***********************************
See part 2 tomorrow.
------------------------ Yahoo! Groups Sponsor --------------------~-->
Yahoo! Groups gets a make over. See the new email design.
http://us.click.yahoo.com/hOt0.A/lOaOAA/yQLSAA/ADr1lB/TM
--------------------------------------------------------------------~->
Yahoo! Groups Links
<*> To visit your group on the web, go to:
http://groups.yahoo.com/group/Adobe-Premiere/
<*> Your email settings:
Individual Email | Traditional
<*> To change settings online go to:
http://groups.yahoo.com/group/Adobe-Premiere/join
(Yahoo! ID required)
<*> To change settings via email:
mailto:[EMAIL PROTECTED]
mailto:[EMAIL PROTECTED]
<*> To unsubscribe from this group, send an email to:
[EMAIL PROTECTED]
<*> Your use of Yahoo! Groups is subject to:
http://docs.yahoo.com/info/terms/
