An experiment

[Brian Ellis]

I have about 5 min of video (no sound) showing a scientific phenomenon (filmed through a microscope). I wanted to put this onto a CD format playable on ordinary stand-alone DVDs (suitable, modern, ones, of course).

I burnt a PAL SVCD, but this had just sufficient "jaggies" to detract from the observation I was wishing to point out.

I then had the idea of rendering it in DVD MPEG-2 format (704x576, instead of the 480x576 pixels). With the MSP 6.5 DVD plug-in, it was not possible to record it to a CD format. With Nero, I tried to do the same and it told me it was an incompatible format, but it gave me the chance to override this problem and burn it, anyway. It played delightfully in the stand-alone, without a glitch. Interestingly, I would have thought that the file would have been about 1.5 times that of the SVCD (based on no. of pixels), but it was about twice the size, which puzzled me at first.

I then, out of curiosity, repeated the experiment with a couple of minutes of ordinary video. This time, the video was perfect but the sound was chopped into, very roughly, one second on and off bursts, so this technique is not valid where sound is a required factor. This also explains the difference in file size: the standard obviously was adding "blank sound" to fill in the chops.

However, I did what I wanted to do

[Ant]

I'm intrigued, what is this "phenomenon"?

[Brian Ellis]

Ant

The formation of "dendrites" between two conductors of a printed circuit under certain humid conditions. These are tree-shaped whiskers of metal, typically 1 micrometre across, which can grow in a matter of a few tens of seconds and then just as quickly disappear, as they fuse with the current passing through them. They are a common cause of intermittent (usually) failure of electronic equipment operating under damp conditions if the assemblies haven't been suitably cleaned and protected.

In this case, their size was such in the image that some of these dendrites appeared more as a "staircase" than a line in SVCD.

[KuroNeko]

be aware

Brian, please be aware that your CD is out of spec and will not play on every standalone deck.

I have some SVCDs that go a bit over 2600 Mb/s (very slightly out of spec) and they play fine on my Pioneer decks, but they'll lock up on most other brands I tried, especially on lots of those cheap no-brand Hong Kong players.

Just thought I'd mention it in case you plan to distribute this or use it as a demo someplace wiithout bringing your deck with you.

Neko

[efty]

Where did u do this experiment

Just wondering if your experiment is work related or just for fun. Did u perform the experiment yourself in Cyprus? If yes where?

[Brian Ellis]

Work related. Anyone can do it anywhere, provided they have
a) a microscope
b) a printed circuit board with two bare conductors about 0.5 to 1 mm apart
c) a 9 V battery
d) a 10 kilohm resistor
e) a drop of purest water
Wire up the battery in series with the resistor across the conductors. Place the PCB under the microscope with the edge of the negative conductor in the field of view. Place a drop of 18 megohm water across the conductors, in the field. Watch! It will happen but the speed depends on the insulator material, the conductor finish, the cleanliness level of the PCB etc. It is a classic test developed by a Charles Jennings in the 1970s at one of the US Govt. labs (I think Sandia). He made a 16 mm film of the phenomenon, but materials have changed since then. A company in England of which I happen to be Chairman was commissioned by a client to produce a video, which they did (on VHS tape). They were then asked whether it was possible to have it on a medium they could play on a DVD player, to get a better resolution. As the guys there knew that I was into video editing etc., I was therefore asked how to do it ...

[zeb7]

Brian ,
This may be a multiplexing issue i think you may have multiplex to DVD not to SVCD ( if you're using Ulead)
I have made many SVCD's with 704x576 / 720x576 and 352x576 resolutions
the only multiypexer that was successful is the one in TMPGenc (Mpeg2 SuperVCD vbr)

if you have time , here is the fastest way to complete the test;

1, Download tmpgenc mpeg encoder at http://www.tmpgenc.net/e_download.html
2, setup and run the program , then from the file menu at the top of the screen select "mpeg tools"
3, select the "Multiplex" tab and set "type" to (Mpeg2 SuperVCD VBR)
4, add you're mpeg file(you may load directly from the svcd if no longer on the HD )
It will now re-multiplex a svcd stream to a new file.
If you get an error message like "buffer overflow" you will need to encode the original AVI using
tmpgenc program.
5, Use nero to make a new SVCD ( xSVCD )

One other thing ,if you do re-encode i would suggest
the tmpgenc program with 2500kbs CBR with 64bks mono audio for maximum compatibility

---
On a different mater , maybe the apple mac DVD authoring software has a better multiplexer
than PC software ? > see ; Wally: glitches
---

zeb.

[Brian Ellis]

Zeb7

Thanks. The company's client has received the SVCD and he says that it's perfect, so everyone's happy with the results.

I don't have time to experiment further, because it is a one-off affair, unlikely to come up again, but if anyone else has succeeded in recording SVCDs in 704 x 576 (or 480) WITH correct sound, for short clips, it may be helpful to publish the MO here.