I'd like to ask some questions. Any help you folks can provide will be
greatly appreciated.

We recently acquired a Sony DCR-TRV22 handycam which shoots video on
mini DV tapes. (I gather that this is a commonly used amatuer,
issued-to-you-if-you-live-in-the-suburbs sort of camera). We've used
it to shoot some home video of my sister's graduation.

I'm not totally clueless in regards to computers, pixels, digital
video, codecs and so forth, but until now I've generally dealt with
clips I've downloaded, streamed online, or watched on a DVD. This is
my first hands-on experience with shooting and then captureing home
video.

I've tried various programs to capture the video from the mini dv
tape. I was able to capture with the bundled software and also with
Ulead Video Studio but in each case I ran into some issues. Now I'm
using Premiere.

I decided to test the capture process with a small segment of video; I
captured a minute and a half of video to an avi file.

At this point, I'm not having any actual problems. The video plays
fine and it looks as good as it did on the LCD screen. But I know from
past experience that when you increase the size on the screen, the
video won't necessarily look as good. So I tried viewing it in Windows
Media Player at 200% and, of course, it looks crappier. I switched to
full screen and it got even worse.

So... my questions:

1. How can I determine the original aspect ratio of my video? That is,
the width and height in pixels of the image when it's on the digital
video tape, before it's captured and encoded. Is it 720x480? That's
what it is post capture/encoding. I want to make sure I'm not losing
any quality or otherwise screwing up the transfer process. Since I
don't know exactly what I'm starting out with, it's kind of hard to
figure out whether the transfer/capture process is the best it could
be.

2. Ultimately, I'm supposed to encode this video in an mpg format and
put it on a Super Video CD (SVCD). This will involve some degree of
compression, won't it? How much quality will I lose in the process?
Given the level of quality of the raw video, will I be able to
produce, at the end, a useable product? Is there anything I can do
along the way (i.e. encoding in a particular format, refraining from
encoding, etc.) to help preserve the quality of the video? When it
comes time to choose a codec, how do I know which one is the best one
for the task I have in mind?

3. The Premiere tutorial mentions the sample rate for the audio. As I
understand it, my audio might be sampled at 3200 mhz or at 4800 mhz.
How do I determine which of these rates my camera uses? I did find a
setting on my camera which changes from 12 bit audio to 16 bit audio.
It had been set on 12 bit, so I guess that's the setting that was in
use when the video was shot. What does that mean to me as I capture
and edit the video?


Alright... I expect that's enough for now. Thanks for listening..


Clay

"Clay Smith" <ClaySmith201@softhome.net> wrote in message
news:8e1nb0ps74v58kt38cl3m69ld1045lhdea@4ax.com...
If you're shooting NTSC miniDV, which is what your TRV22 is, it's always 720
x 480.

As long as you don't transcode, you're not losing any resolution.

Depending on how much video you want to put on the SVCD, it can involve an
enormous amount of compression. MPEG compression is variable, depending on
bit rate.

It can be quite a bit, or very little, depending on how much video you
transcode. Commercial DVDs are MPEG2, and their quality is usually quite
high.

There's no way to answer that. SVCD is a poor medium for video. miniDV/D8
uses 13 gigabytes per hour of video. A DVD holds 4.7 gigabytes. You can
usually get good quality up to about 2 hours worth of video if you do
multi-pass VBR transcoding, which is about a 7-to-1 compression ratio.

A CD holds 650-700 megabytes. You're compressing nearly 20-to-1 to get one
hour on a CD. It's not going to look particularly good.

Yes. Capture to uncompressed AVI. Edit the AVI. Don't transcode 'til
you're finished editing. Use a good transcoder (tmgenc is good and cheap,
but slow).

The Microsoft DV codec works pretty well and doesn't compress.

If it's 16-bit, you're sampling at 48.1 khz.

12-bit on miniDV usually means you can capture 2 stereo tracks, i.e. 2 pairs
of 2 tracks each. There's little reason to do that. Capture at 16-bit.

Thanks for all the information, Ptravel. It's helped quite a bit.

I've got more questions if you're of a mind to take a stab at them....

Well, that's kind of the rub, actually. See, I was using Ulead Video
Studio to capture the video and it produced enormous avi files (about
8 or 9 gigs of avi for forty five minutes of video). When I analyzed
these files with GSpot, it didn't find any codecs in them but the
files played in Windows Media Player just fine. From this I surmised
that Ulead didn't encode the avis at all during the capture process.

However, when Premiere (the software I'm currently using) captures the
video, it does employ some sort of encoding. The avis it produces are
very large (200 MB for less than a minute of video), so I guess
there's little or nothing in the way of compression... but GSpot says
that they are encoded using a codec called dvsd.

So... I'd like to refrain from transcoding as you advise but I'd also
like to continue using Premiere. What's the best way to get Premiere
to cooperate? Can I tell Premiere to leave the video unencoded? Would
I do just as well to continue using the dvsd codec since it doesn't
seem to be compressing the files at all? Is compression the only
concern here or are there other things involved in the encoding
process that can cause my video to lose quality?





Suppose I wanted to put 20 minutes of video on one CD?. That would be
a compression ratio of about 7 to 1, correct? Or suppose I wanted to
do 10 minutes... that'd be a ratio of 4 to 1 or thereabouts.

I'm certainly willing to experiment with different ratios and burn a
number of test CDs. I did something similar when I was first learning
to make Video CDs. But what I'm wondering is how I'm going to control
the compression ratio. Does this all come down to the codec I use? For
example, if I decide I want a 15 to 1 ratio, do I have to go find a
codec that compresses to that degree? And then if I change my mind and
decide I'd rather do it at 12 to 1, I have to find a different codec?
Or will tmpgenc (or some other piece of software) give me some control
over the degree of compression?


Again, thanks for all the information.



Clay

Clay Smith wrote:

The normal (Mini)DV (DV25) datastream is about 13 GB per hour.
Eight or nine gigs for 45 minutes sounds about correct.

I have not been using any Ulead products for capturing DV data,
so I cannot offer any explanation as to why you see the "DVSD"
codec mentioned in the Premiere captures but not in the Ulead
captures. (Maybe it is some issue with GSpot itself, possibly
caused by the difference between Type 1 and Type 2 DV AVIs.)

That being said, DVSD = "DV, Standard Definition". Premiere
does not touch or alter the actual DV data; you get it as it
was on the tape. There are a couple of different ways to
encapsulate the DV datastream inside an AVI file but the
raw DV data should remain the same, regardless of the
encapsulation method. You can even encapsulate raw DV data
inside a QuickTime file, or just process raw DV streams
with no encapsulation as some programs do.

In short, Premiere left the raw DV data in the original DV
format without altering anything; it just encapsulated it
inside an AVI container, and you see "DVSD" as the FOURCC
codec identifier just as you should. The actual raw DV data
inside the file is exactly the same as on the tape.

Cuts-only editing (simply rearranging the order in which
the frames will be shown, removing frames, or concatenating
footage from multiple source DV clips into a single clip)
does not require re-rendering the material, and you do not
lose any quality.

However, if you touch the actual content of the frames
altering the pictures in some way (for example, by
adjusting brightness, colors, mixing two pictures
together, adding transitions, titles, filters, effects
etc.) every frame whose content you touched will need
to be decompressed, then processed and finally recompressed
to the DV format again. This decompression-recompression
cycle is handled by the software DV codec installed in
your system, and the resulting quality loss depends on
the quality of the codec.

There are DV codecs available at least from Microsoft,
MainConcept, Panasonic, Canopus and Matrox. Leaving
the actual decompression/compression quality aside,
one particular way in which they differ is how they
handle full-range R'G'B' and "studio" R'G'B' (related
to the "white clipping" problem). You can read more
about that in here:

<http://www.adamwilt.com/DV-FAQ-editing.html#CodecProblems>

There is also an interesting (albeit a bit out-of-date) codec
comparison page (with sample images) here:

<http://www.adamwilt.com/pix-codecs.html>

Also see Adam Wilt's main DV site:

<http://www.adamwilt.com/DV.html>

--
znark

Thanks for all the information. This is all beginning to make a lot
more sense to me now. I did have one question about your response.....


You mentioned that...

But then, speaking about


you say that ...


Alright. So... if my data is in the "raw DV" format... do I still have
to be concerned about this decompression/recompression issue under
these circumstances? I don't imagine that I'll be making a lot of
alterations ("alterations" as distinguished from cuts and pastes, per
your explanation) to my images, but I'm trying to get comfortable with
this whole process, to learn where the stop signs and speed bumps are
located, as it were. It's making me a bit nervous knowing that some
casual fooling around with frames of video (hey, I wouldn't it be cool
if I inserted a still shot where my Aunt's face is superimposed on
Madonna's body?) might, without me even realizing it, screw up my
video footage.





Thanks for the links. I've read through some of this stuff and will
definitely peruse it further.



Clay

"Clay Smith" <ClaySmith201@softhome.net> wrote in message
news:enhpb0dhb7r5i31dj506n8474d87g58ere@4ax.com...
The word "raw" is just a descriptive term to describe the encapsulated DVD
data. For all practical purposes it has no impact on anything.

As Jukka said, if you make any changes to a frame of video in any way it
will be affected by the "decompression/recompression issue". So any frame
that has your Aunt's face pasted to Madonna's body is affected. The point
is that your whole video is not affected unless you change it as a whole.
For example, if you mod the brightness, color, etc... Anything that would be
globally across the video. Transitions between clips will be affected
during the transition, but not the rest of the clip.

"FLY135" <fly_135(@ hot not not)notmail.com> wrote in message
news:g65vc.34052$zO3.25279@newsread2.news.atl.eart hlink.net...
of what you watch on reality TV is shot with Pro and Prosumer DV cameras.
That footage is captured to either MacIntosh or Windows based systems. The
clips are cut and pasted and layered and processed with all the necessary
recompressions. When rendered back to tape those programs are shown in
large and small venues as well as broadcast. Most people agree that the
results compare favorably with BetaSP, the previous (and still in many
cases) standard for such programming. It is good to want to understand the
dos and don'ts, but most of the quality loss you will experience is going
from DV standard files in your computer to other formats, i.e. VCD, SVCD,
and DVD. In choosing the parameters for these transformations you will want
to experiment a bit to see how much to cram into the medium (in program
length) before the quality no longer meets your standards for the specific
purpose for that program.

Maybe the more important thing to keep in mind is that your original clips
captured into your computer will retain all of their quality. Working in
NLE is not destructive to original material. You can always go back. Have
no fear. Get on with it.

Steve King

Clay Smith <ClaySmith201@softhome.net> wrote in message news:<l6aob0ldo79b3mvpo90mhnsnhpt55radcn@4ax.com>. ..
8 or 9 gig for 45 minutes sounds about right. The codes won't be in
the file, they'll simply be packaged in a form which can be read by
the appropiate codec. Unfortunately, I don't know what the Ulead
product does.


You need to check your compression and codec settings on export in
Premiere.

When you export, go into settings and make sure you're using the
Microsoft DV codec and it's set to not recompress.

From what I recall, SVCDs are limited in the bit rate. To get good
video, MPEG2 needs to be running at about 4k bits -- anything less
won't produce satisfactory results.

No. For transcoding, it will be a function of a number of settings.
Premiere (which uses the Main Concept transcoder) doesn't offer
complete control over transcoding. Tmpgenc will, however. There are
a number of websites devoted to setting up tmpgenc to produce optimal
results. I only know one for producing DVDs -- can't help with SVCDs,
but google should find them.

Mostly, you'll just change the bit rate. You don't change the codec.

Steve King wrote:

There is one notable exception, though: the "white clipping"
problem (which I already mentioned in my previous message)
from which some DV codecs suffer.

When using a DV codec which has forced Rec. 601 / legal color
safeguards (and no user interface to turn them off), this can
become annoyingly visible e.g. when making a transition during
a shot where the sky is visible - often leading to an unwanted
effect of the sky suddenly getting a duller color during a
transition or when displaying a title, then reverting back to
its original shade when the transition is over or the title
has disappeared.

--
znark

"Jukka Aho" <jukka.aho@iki.fi> wrote in message
news:c9j40j$4o3$1@plaza.suomi.net...
Tell me more, because the effect would, indeed, be annoying. What DV codecs
seem to be susceptible (or have this as a 'feature').

Steve King

Steve King wrote:
See <http://www.adamwilt.com/DV-FAQ-editing.html#CodecProblems>.

--
znark

"Jukka Aho" <jukka.aho@iki.fi> wrote in message
news:c9kvvi$a3s$1@plaza.suomi.net...
Thanks. I'll read it.

Steve King

On Mon, 31 May 2004 16:49:47 -0500, Clay Smith
<ClaySmith201@softhome.net> wrote:

If you use firewire to capture (same thing as Sony iLink, different
name), you get 720x480, no other options are allowed. That is what is
recorded on tape.

If you use USB, you'll get a lower resolution.

Windows Media Player has several settings which affect video
playback quality. Seems to me that the default was a low res (and
lower CPU use) version of the video. 200% of 720x480 is 1440x960 - to
see that you'd need to be running 1600x1200 desktop resolution --
which is pretty uncommon on anything smaller than 19" monitors.

Why SVCD? DVD is a better option if you want to keep the quality
up, and will play on more players.

mpeg2 at a high enough data rate will look pretty much like your
original DV material. Mpeg isn't good for editing -- you don't want
to recompress it, as it loses quality each time -- but it is the most
practical way to encode video for playing.

You should set the camcorder to 16 bit/48k -- 12 bit is lower
quality. The DV spec puts in 12 bit as the default, and on some
camcorders (not yours), you can record a second 12 bit track (for four
tracks of audio).

Since any additional audio you use will be added on the computer,
the 12 bit setting isn't helpful at all.


--
*-__Jeffery Jones__________| *Starfire* |____________________-*
** Muskego WI Access Channel 14/25 <http://www.execpc.com/~jeffsj/mach7/>
*Starfire Design Studio* <http://www.starfiredesign.com/>