"Peter" wrote in message
...
I can report that making a VCD quality (576 lines or whatever the
exact figure is) movie (with Pinnacle 10.5) from 640x480 stills does
generate a movie of reasonable quality. Certainly at least as good as
normal TV reception.

I found other quality limiting problems however, which made me realise
that the pros that do this have to be pulling a few tricks.

For a start, the slightest amount of camera shake or subject movement
screws up the result. Whereas of source if you take one still shot,
then given a fast enough shutter, movement doesn't matter.

So, let's say you want to do one of those classics e.g. a flower
opening up, over several hours. The camera obviously has to be on a
solid tripod (easy) AND the flower has to be completely shielded from
wind (not quite so easy if you also want it to open, which needs
sunlight...); I guess the pros either do it indoors with a lamp, or
they do it outdoors and surround the flower with a perspex enclosure.

The stills came from a Ricoh Caplio R6 whose fastest interval shooting
is 5 secs. I needed under 1 sec and got rid of the camera. (the
battery incidentally lasted for 1400 shots, not bad). Next, I will try
some VGA webcam - there appears to be a selection of PC software which
can capture periodic stills from a webcam. I hope not all webcams are
as crap as all those I have seen so far... I have an Axis 205 webcam
right here, which wasn't at all cheap, and the colour quality is truly
crap compared to a cheap compact camera set to 640x480.

You can look at what is probably mostly the lower cost range
of web cameras he

http://www.geeks.com/products.asp?Cat=VID

But now this thread has truly been hijacked.

Luck;
Ken

"Ken Maltby" wrote in message

I was providing a simplified explanation as to why the generalized
assertion the DV is totally lossless, is obviously wrong.

As was I. ;o)

Your explanation was far more technical than mine, but we had the same
goals. I was simply responding to Trev's idea that because the images are
smaller than DV resolution, they won't suffer from the compression when
combined into a DV AVI.

The best way for the OP to combine his stills into a video file is to use
some form of image sequence that utilises his original JPEGs. Compression to
any format, DV included, should be avoided.

"G Hardy" wrote in message
...
"Ken Maltby" wrote in message

I was providing a simplified explanation as to why the generalized
assertion the DV is totally lossless, is obviously wrong.

As was I. ;o)

Your explanation was far more technical than mine, but we had the
same goals. I was simply responding to Trev's idea that because the
images are smaller than DV resolution, they won't suffer from the
compression when combined into a DV AVI.

The best way for the OP to combine his stills into a video file is
to use some form of image sequence that utilises his original JPEGs.
Compression to any format, DV included, should be avoided.

He would also be best to avoid lousy formats (such as JPEG) when
saving image stills, not sure what his various software supports but,
he should be saving his image sequences in an uncompressed format such
as TIFF, (Photoshop) PSD or failing that Bitmap.

Ken Maltby wrote:
wrote in message ...
G Hardy wrote:

wrote in message ...

...Drop a couple of hundred stills on the time
line and render as an DV quality AVI file. Repeat this as many times
as necessary. Once all the stills are in movie form, you can use Studio
to splice all the AVI files you created back into a single movie. Because
you are working with DV quatilty AVI files, there is no loss in quality
up to this point.

Huh? So what about the huge drop in quality when you render the stills to
DV?

There is absolutely no loss in quality when rendering to DV quality AVI
files. It is the same format that Mini DV and D8 camcorders use. You
can edit, write to DV and re-edit as many times as you want. You are
probably thinking of MPEG editing where the quality drops each time.
-Bill

DV-25 is at least a 5:1 compression of any image data being
encoded into it. The image sensor on a typical Mini-DV camera
supplies a frame image to be encoded into DV-25 that is a little
larger than 720x480 (NTSC). It is supplied as two interlaced
frames. 720x480=345,600 pixels

A still image can have several million pixels, that is per image,
comparable to per frame.

I'm afraid that DV will suffer compression losses and artifacts
if abused just like any digital format. It may take more abuse
and have smaller, more limited, impact, but there can be very
noticeable impact. And that is in relation to a more compressed
format, not an original image. It is always something less than
any original image data. DV-25 at 25Mbps won't/can't encode
all the image data of a quality still image....snip

To refresh everyone, I got involved in this discussion when someone was
having trouble adding a bunch of still photos to a timeline and having
his editing program hang up. Because I had ran into the same problem
myself, I suggested adding a hundred or so stills at a time, and
rendering them as a DV AVI file. Once all the pictures were saved in
the AVI files, they could be brought back into the time line and the
editing program would threat them as any other movie, and should render
the project with no problems. This is a solution that has worked for me
in the past.

Others suggested that there would be a quality loss with this method; so
I decided to do a test and see if my methodology was sound. I took a
still picture (one frame) and rendered it as File01.AVI. I then started
a new project and used File 01.AVI as the source and rendered it to a
new file called File02.AVI. I repeated this nine times. I then started
a new project and imported all nine files to the timeline. My movie
was now 9 frames long, and each successive frame was a copy of the
previous with the last one being a 9th generation copy. I then placed
a copy of frame #1 (the 1st generation original) at the end of the movie
so after rendering to DVD, I could step forward/back a frame at a time
and do a comparison of the quality. Guess what...using the HDMI cable
of my DVD player, I could detect no difference in the quality between
the 1st and 9th generation still frame on my 50" HDTV.

If anyone else would care to repeat this test I would be interested in
their results.

-Bill

wrote in message ...
Ken Maltby wrote:

snip

If anyone else would care to repeat this test I would be interested
in
their results.


Apart from the suspect methodology, as outlined by Peter (re .avi
'wrappers' and what they actually contain), might I suggest the test
to be done by sending the play-out to both a vector scope and
histogram rather than to a (LCD/Plasma) TV that will have all sorts of
error correction built in.

wrote in message ...

To refresh everyone, I got involved in this discussion when someone was
having trouble adding a bunch of still photos to a timeline and having
his editing program hang up. Because I had ran into the same problem
myself, I suggested adding a hundred or so stills at a time, and
rendering them as a DV AVI file. Once all the pictures were saved in
the AVI files, they could be brought back into the time line and the
editing program would threat them as any other movie, and should render
the project with no problems. This is a solution that has worked for me
in the past.

Others suggested that there would be a quality loss with this method; so
I decided to do a test and see if my methodology was sound. I took a
still picture (one frame) and rendered it as File01.AVI. I then started
a new project and used File 01.AVI as the source and rendered it to a
new file called File02.AVI. I repeated this nine times. I then started
a new project and imported all nine files to the timeline. My movie
was now 9 frames long, and each successive frame was a copy of the
previous with the last one being a 9th generation copy. I then placed
a copy of frame #1 (the 1st generation original) at the end of the movie
so after rendering to DVD, I could step forward/back a frame at a time
and do a comparison of the quality. Guess what...using the HDMI cable
of my DVD player, I could detect no difference in the quality between
the 1st and 9th generation still frame on my 50" HDTV.

If anyone else would care to repeat this test I would be interested in
their results.

The bad news is you didn't render it nine times. As was mentioned elsewhere
in the thread, virtually every editing program will "smart render" video* -
DV video especially - meaning that unless you do something to the video
frame (or turn "smart render" off) each generation will be almost a byte for
byte copy of the previous one. The only time you rendered your stills to DV
was the first generation. After that, you were just copying the file.

Try it again, this time adding a subtle title to the bottom of each
generation, showing the number of that generation. It doesn't really matter
if each generation's digit overlays the last one, but it will look a mess,
so you might want to start at one side and work your way to the other. the
purpose of adding the title is to force your video to be rendered, not just
copied.

I've got an old set of generational DV frames that show the effect of
repeated true rendering. The first image is at
http://homepage.ntlworld.com/gareth.hardy1/mug/ft0.jpg and is the frame
captured directly from the camera.

http://homepage.ntlworld.com/gareth.hardy1/mug/ft1.jpg is the same frame
after I forced it to render, then checkerboarded against ft0 to show the
difference. You can just see the checkerboard pattern if you look hard
enough, but you probably wouldn't be able to see it if you watched the
uncheckerboarded version and the original video side by side. This shows
that the quality loss is there, but whether you'd be able to see it is
another matter.

http://homepage.ntlworld.com/gareth.hardy1/mug/ft2.jpg is the same frame
after its second forced render, again checkerboarded against the original
frame to show the difference.

Anyway, the full list of links is:
http://homepage.ntlworld.com/gareth.hardy1/mug/ft0.jpg
http://homepage.ntlworld.com/gareth.hardy1/mug/ft1.jpg
http://homepage.ntlworld.com/gareth.hardy1/mug/ft2.jpg
http://homepage.ntlworld.com/gareth.hardy1/mug/ft3.jpg
http://homepage.ntlworld.com/gareth.hardy1/mug/ft4.jpg
http://homepage.ntlworld.com/gareth.hardy1/mug/ft5.jpg
http://homepage.ntlworld.com/gareth.hardy1/mug/ft6.jpg
http://homepage.ntlworld.com/gareth.hardy1/mug/ft7.jpg
http://homepage.ntlworld.com/gareth.hardy1/mug/ft8.jpg
http://homepage.ntlworld.com/gareth.hardy1/mug/ft9.jpg

It illustrates the effect of continual re-renders of the same video. There's
another file, http://homepage.ntlworld.com/gareth.hardy1/mug/ftz.jpg, which
is the same as ft9, but without forced render: No visible difference. It's
interesting to note that in the 31,804,228 bytes of smartrendered AVI, only
seven bytes change with each generation.


* As long as the source video properties match the editor project settings.

[ apologies if this message is a duplicate ]

wrote in message ...
Ken Maltby wrote:

snip

If anyone else would care to repeat this test I would be interested
in
their results.


Apart from the suspect methodology, as outlined by Peter (re .avi
'wrappers' and what they actually contain), might I suggest the test
to be done by sending the play-out to both a vector scope and
histogram rather than to a (LCD/Plasma) TV that will have all sorts of
error correction built in.

G Hardy wrote:

wrote in message ...

To refresh everyone, I got involved in this discussion when someone was
having trouble adding a bunch of still photos to a timeline and having
his editing program hang up. Because I had ran into the same problem
myself, I suggested adding a hundred or so stills at a time, and
rendering them as a DV AVI file. Once all the pictures were saved in
the AVI files, they could be brought back into the time line and the
editing program would threat them as any other movie, and should render
the project with no problems. This is a solution that has worked for me
in the past.

Others suggested that there would be a quality loss with this method; so
I decided to do a test and see if my methodology was sound. I took a
still picture (one frame) and rendered it as File01.AVI. I then started
a new project and used File 01.AVI as the source and rendered it to a
new file called File02.AVI. I repeated this nine times. I then started
a new project and imported all nine files to the timeline. My movie
was now 9 frames long, and each successive frame was a copy of the
previous with the last one being a 9th generation copy. I then placed
a copy of frame #1 (the 1st generation original) at the end of the movie
so after rendering to DVD, I could step forward/back a frame at a time
and do a comparison of the quality. Guess what...using the HDMI cable
of my DVD player, I could detect no difference in the quality between
the 1st and 9th generation still frame on my 50" HDTV.

If anyone else would care to repeat this test I would be interested in
their results.

The bad news is you didn't render it nine times. As was mentioned elsewhere
in the thread, virtually every editing program will "smart render" video* -
DV video especially - meaning that unless you do something to the video
frame (or turn "smart render" off) each generation will be almost a byte for
byte copy of the previous one. The only time you rendered your stills to DV
was the first generation. After that, you were just copying the file.

Haven't you just proven my point. If you are just combining a bunch of
stills
into a movie, and the program does not re-render the project, then there
is no
loss of quality. That was what I was saying in the first place

Try it again, this time adding a subtle title to the bottom of each
generation, showing the number of that generation. It doesn't really matter
if each generation's digit overlays the last one, but it will look a mess,
so you might want to start at one side and work your way to the other. the
purpose of adding the title is to force your video to be rendered, not just
copied.

I do not even have to try that to know that it will degrade the video.
Anytime
you add an overlay to the video, you are going to degrade it a little.
The
object here is to just get a balky program that chokes when it has to
process
a bunch of stills on the timeline to work. My suggestion will work with
no loss
in quality of the original stills.

-Bill

:Jerry: wrote:

[ apologies if this message is a duplicate ]

wrote in message ...
Ken Maltby wrote:

snip

If anyone else would care to repeat this test I would be interested
in
their results.


Apart from the suspect methodology, as outlined by Peter (re .avi
'wrappers' and what they actually contain), might I suggest the test
to be done by sending the play-out to both a vector scope and
histogram rather than to a (LCD/Plasma) TV that will have all sorts of
error correction built in.

The guy is using a $99 editing program. I doubt if he has a vector
scope in his basement to prove his eye is lying to him. I think he just
wants to get his project done. My suggestion will allow him to proceed
with a balky program. The fact that he has not chimed back in suggests
to me that he is busily finishing his movie.
-Bill

wrote in message ...
G Hardy wrote:

wrote in message
...

snip

The bad news is you didn't render it nine times. As was mentioned
elsewhere
in the thread, virtually every editing program will "smart render"
video* -
DV video especially - meaning that unless you do something to the
video
frame (or turn "smart render" off) each generation will be almost a
byte for
byte copy of the previous one. The only time you rendered your
stills to DV
was the first generation. After that, you were just copying the
file.

Haven't you just proven my point. If you are just combining a
bunch of
stills
into a movie, and the program does not re-render the project, then
there
is no
loss of quality. That was what I was saying in the first place

No he hasn't, what do you not understand about the fact that the DV
codec compresses (@ 5:1) the source, be that your image file being
frame-served or the out-put of the CCD stage of a camera - what, AIUI,
Gareth was demonstrating in his post is the fact that DV is *not*
'none-lousy' (although being digital it's a bite for bite copy when
transferred to computer or another tape, unlike analogue which suffers
from generational loses), by revealingly rendering the same (DV) .avi
file you will see that each render introduces compression artefacts.


Try it again, this time adding a subtle title to the bottom of each
generation, showing the number of that generation. It doesn't
really matter
if each generation's digit overlays the last one, but it will look
a mess,
so you might want to start at one side and work your way to the
other. the
purpose of adding the title is to force your video to be rendered,
not just
copied.

I do not even have to try that to know that it will degrade the
video.
Anytime
you add an overlay to the video, you are going to degrade it a
little.
The
object here is to just get a balky program that chokes when it has
to
process
a bunch of stills on the timeline to work. My suggestion will work
with
no loss
in quality of the original stills.


Whoooosssshhhh....

That's some codec if it doesn't compress the first (normally in
camera) render but then compresses subsequent renders - think about
what you're first admitting to and then attempting to claim in what
you said above!

--
Jerry - on an different NNTP server.
Someone managed to break the other one!