What happens to digital video when you edit it?

Trying to sort out issues related to digital capture - i.e. done via a
Digital8 camcorder -vs- something captured from an analog VCR or cam.

Given that everything that happens inside a computer is in the "digital"
realm, it seems that the difference is where the digital to analog
conversion takes place, either inside the cam or via a capture card.
Essentially correct so far?

If I capture video off my Digital8 camcorder via Firewire either as a
pass-through or using a tape in the cam with digital video on it, how is it
different than analog video captured through a capture card?

Now, I take the file created via a Digital Cam onto the hard drive and I do
something to it. Let's say either adding reverb via Soundforge, or altering
it in some way with Virtualdub.

In the case of the Soundforge effect, I can save it either as the same file
or a different file with a different name. In the case of Virtualdub, it
appears I have to save it as a different file name. With Virtualdub, I have
to choose among a number of codecs when saving it. Whatever codec I use, the
size of the file seems to always be different than the original "digital"
video file.

In either case, what's the difference between the original file and the
altered file?

Further, if I subsequently upload the altered video onto the camcorder from
the computer, what's the difference between this "re-captured" video and the
original? Is there a way to do it to avoid any loss of quality?

Thanks for all input.

 

Depends on how good your (unidentified) D8 camcorder is vs. how
good your (unidentified) video capture card is. Since you have not
identified either of them, only you can answer that question.

If you are only modifying the audio, hard to imagine why the file size would
be different *given identical file type/codec in and out*. OTOH, making
extensive changes to the video, or editing (shortening, lengthening, etc.)
the video or audio would be expected to change the file size, even with
the same file type/codec. And certainly if you were using a *different
file type and/or codec*, one would expect the file size to change.

Assuming you are using DV encoded video from your D8 camcorder
and stored without further compression into a file type like AVI....

Most video editing applications will alter only the video that you
have added effects to (dissolve, wipe, titles, color correction. etc.)
A simple "cuts-only" edit will return every frame of video to the
camcorder tape in exactly the same quality as on the original tape.
One of the prime benefits of digital recording and processing.

 

I'll rectify that.

Cam- Sony TRV-240 Digital8
Capture Card - Pinnacle DC10-Plus
Firewire integrated with Audigy2 sound card
Capturing and project editing using Pinnacle Studio 9
OS - WinXP

I did an experimental short capture/edit of about 30 secs. The original
file is 281 megs. When I add the reverb with Soundforge I did a "save file
as" once with "no recompression" and once under "NTSC DV". Both of these
come up as 283 megs, 2 megs more than the original file. I assume the extra
2 megs is because of the extra information that the reverb adds.

Looking at Virtualdub, I don't see any codecs labeled as "DV". There's a
codec labeled as "Uncompressed RGB/YCbCr" among others. If I save under that
codec, the resulting file comes up as 2.26 gigs, almost 10x the size of the
original file. Clearly, not the same thing as "no recompression".

So, I guess the question is how can I save the file as the exact same format
under Virtualdub?

 

"Doc" wrote ...

I have no first-hand (or even second-hand) experience with
the Pinnacle, but given the nature of the Pinnacle posts that
I've seen in this neighborhood, I'd expect that the quality of
the Sony TRV240 A/D conversion will generally exceed
that of the Pinnacle. The "quality" of the Firewire port
is of no consequence (assuming that it is actually functioning
properly.)

Dunno. I am not a VirtualDub user. But there are many around here.
Perhaps you should repost this specific question with a better subject
line to attract their attention.

 

Use the "Direct Stream Copy" option in the video menu. That way
VirtualDub will not reencode the video stream in any way.

You may need to installa a VfW (Video for Windows) style DV codec into
your system, though. VirtualDub cannot use DirectShow codecs (such as
the Microsoft DV codec that comes with the system.) For example, there
is a Panasonic DV codec circulating in the web, or you could purchase a
VfW DV codec from MainConcept (there is a demo version available for
download at their website.)

The latest versions of VirtualDub come with a built-in DV decoder, but
not with a full codec with which you could encode new material into DV
format.

 

The DC10 plus has a hardware codec which encodes to the MJPEG format.
You can select the quality/compression ratio of this encoding; the
higher qualities are as good as DV (though if the video signal is coming
off an analog tape format, the picture will be poor to start with).

The camera will have a hardware codec which encodes to the DV25 format.
This is a fixed compression ratio.

 

The Pinnacle DC10plus is an analog card which captures using the Pinnacle
MJPEG codec - Zoran chip I believe. The Pinnacle Studio 9 (if it's the AV/DV
version) allows capture of both MJPEG analog and DV capture over a 1394
device. When choosing the option of DV capture your choice for input/capture
should be the Pinnacle DV codec.

DV files have a data rate of 3.62MB/sec. which means a 30 second clip should
be only about 115MB in size. MJPEG captures can be done on a sliding scale
(at least that's the way all the other Pinnacle analog card worked) with a
default setting which yeilded about 6 minuites per gigabyte - a data rate of
about 7MB/sec.

According to the Studio 9 AV/DV manual you can actually capture using the
MJPEG codec over the 1394 device. Why this is I do not know because the
highest data rate of the natural DV stream is the same no matter what. If
the MJPEG codec is used and the highest setting is chosen then the system is
doing the encoding on the fly and giving a larger file size where it is not
necessary.

From what I gathered by reading the manual was that you can also capture to
MPEG formats on the fly. You may want to compare notes for input and output
to ensure those settings are for DV.

 

Just one little additional point. In general, computer editing is
nondestructive, what the edit decisions do to captured video is merely
create a list of "pointers" or addresses to the various parts of the
footage... a long playlist of "First play this bit up to this point,
then jump to this point and play x amount..." Edit changes change the
pointers, not the data. like skipping around a CD album to play the
songs in a different order.

Except.

Except where there's an effect more complicated than a cut or dissolve
(and sometimes dissolves are a problem too). Where the manipulation of
the images is more complicated, like across the length of an effect,
the product is rendered, re-written to a new file. The old components
remain (nondestructive) but the pointers now describe a newly-created
chunk of video that may well be recompressed, at some loss of quality.

Transferring the Dv program back into your camera by firewire, the
data is unchanged for the clips that were just cuts. The rendered parts
across transitions or special effectsDo take a quality hit. Usually
it's not noticeable.

Every time you record something, you're already compressing it,
throwing away picture information. Is DV codec better than using motion
jpeg? It's more efficient, uses way less drive space, but i don't think
it's better qualitatively. I'd rather try to do a chromakey in motion
jpeg than DV. And I do.

 

Since dissolves *change* the image throughout the length of the effect,
they are decoded, modifed, and re-encoded just as if you had applied
any other kind of effect. No "sometimes" about it. Always.