Did anyone else notice some omissions and inaccuracies in the Building a Digital Video Capture System Part 1 guide?

here are a few random thoughts I had while reading it:

"Also, some hard drives perform a thermal recalibration every few seconds. "

A problem in '93 maybe, but really, I can't think of a single drive from a major manufacturer that needs recalibrating since 1995. I think the tone of the article makes it sound like an issue you might encounter when buying a new hard drive today - which is false. Also, suggesting that a 5000 rpm drive is adequate is a little reckless. Very few 5000 rpm drives are up to the job.

The whole discussion of FAT16/32 NTFS is circa 1997. There's no mention of the 4 GB file size limit for FAT32 - a big issue for video editing because it equates to only about 20 mins of DV capture. Granted every editor gets around this by splitting the files into 2G chunks but all this is worthy of inclusion. Even better would be quality tests/benchmarks highlighting performance differences between fat32 and NTFS. There is a lot of misinformation out there and some controlled Tom's Hardware tests would be very worthwhile.

The discussion of SCSI controllers is circa 1998:
"Just about every video professional or serious hobbyist I know uses an Ultra SCSI 2 controller and drives...big, big drives. "

Welcome to the 21st century! IDE Raid controllers are a viable, faster and cheaper alternative than all but the most extreme scsi raid setups. Certainly superior to any non-raid Ultra SCSI setup and cheaper too. It may be worth pointing out the difference between so-called Raid cards like the Fasttrack and real Raid cards like the 3ware line and the Adaptec 2400a. The difference being a dedicated controller chip like an i960 chip on the *true* hardware Raid cards.

The article's stance is that cpu power isn't really a factor because capture and playback is implemented in hardware. True - however, it's a pretty big omission to forget about the scurge of computer editing - rendering. Something that is most certainly CPU based. Even with more expensive systems that have dedicated dsp's to render realtime effects, you are still likely to encounter having to wait for scenes to render in a project, you''l just be waiting for the transitions not covered in your realtime effects. Because in practice it's hard to keep every effect within the confines necessary for real time rendering - even if it's just rendering the color bars. That is unless you spend megabucks on a realtime system with virtually no limitations, but on a reasonably priced home system (like the RT2500) you'll be rendering like everyone else.

How about some CPU benchmarks with Athlon/P4/Dual P3 systems to highlight the differences in render time? That's what I'd be looking for in an article from Tom's.

Here's my beef: If this article had been published in 1998 or 1999, then it would be fine - except for the overstating of the thermal calibration problem. But a lot has changed over 2 years, and that's what's missing - the 2001 perspective on system configuration and options for PC video editing.

That said, there's a lot of good information in the article, and I absolutely agree with the premise: setting up an adequate PC for video editing is not trivial.

If the article were tweaked a little it could be THE authoritative source on the subject.

I forgot to mention the article should state buying a new PC from Super Mega Multinational PC Builder Inc. will (probably) not give you a PC system suitable to do video editing on. The article implies this, but should really say it explicitly.

Really if you want a PC for video editing (that's still affordable), you need something customized for the task. For what it's worth, here's what I bought and self-built a few months ago. Final price was around $3000 US.

1.2 Ghz Athlon
Asus A7M266 (AMD 760 chipset)
386MB PC2100 RAM
3Ware 6400 Raid controller (4 channel)
4 * IBM 75GXP 46G drives - In Raid 0 array
Matrox RT2000 - Firewire DV editing system
Matrox G450 PCI - drives 2nd monitor

Last piece of advice: before buying a video editing system, check out the user forums for the editing card you're thinking of buying and lurk for a few weeks. It may save you from buying a system configuration that is known to have problems.

Alex

well, i was looking at matrox video web forum, and there is a lot of complaints about rt2000 compatibility and stability. there is a few really good cards from canopus and dps that don't have so bad feedback, and people should try some of them. i have tried canopus, it's great. no problems with compatibility with any kind of hardware.
thermal recalibration of the hard disks today doesn't have to be the problem, but i've seen on some pinnacle system performance test of a good scsi ibm, and it wasn't good enough. every seagate scsi drive was ok. if you work with dv, you don't have to use scsi drives, and allmost every today's disk will do the job. i tested fujitsu 10 gb ata66 at 5400 rpm, it was ok. it comes to problem when you want to do some more serious job with a lot of elements on the timeline, i.e. 2-3 transitions, some titles and some ext audio tracks, then you will need some more serious disk, maybe even ide raid. also, when it comes to rendering, i personally recomend p4. some users reported that a certain non-realtime dv cards have show some significant realtime features when they upgraded the system to p4. infact, video is the only thing where p4 is better than athlon