Windows XP Incompatibilities - post here

[LC1104]

I was wondering if anyone has installed XP and have had problems in contrast to Win2k. If you have please post your problems and complaints as I am interested in upgrading.

Jason

[Dr Mordrid]

The WPA copy protection drops out when Microsoft says it shouldn't and often won't reregister itself, even if the manual codes are entered. This is a common complaint on the XP beta site and folks are PO'ed over it.

Microsoft typically suggests that if this happens all you can do is reformat and start over. VERY user friendly

I'll give it stability, but I also find the adds and promos throughout the GUI very annoying.

I'll be sticking with Win2K 'til the WPA issue is fixed or its removed.

I'd probably avoid it anyhow until at least the first service pack. Early adopters usually are also migrane sufferers.

Dr. Mordrid

[hapz]

Video Tools, and thus the MJPEG drivers, do not install. And there is no hope in sight either as Matrox stated that they will NOT support XP ;(

[Brian Ellis]

Hi!, I'm with Doc.

I shall NEVER buy any software with WPA. It is an intrusion into one's liberty, as well as being a PITA. Furthermore, experience has shown never to buy any MS O/S until either a) 12 to 18 months have elapsed since official launch or b) one or preferably two service packs have been issued, whichever is the longer. I still use 98SE, NT4 and Linux. I recently obtained W2000, to replace NT-4, but have not yet installed it. For video, I'll be sticking with 98SE. I also have ME, which was bundled with a computer, but I UPgraded it to 98SE.

[mrtj1]

some people have been able to get mjpeg working tho havent they?

also about the wpa, if someone has the above mentioned problem, then could microsoft realistically refuse if you said you could download a copy of xp corp and use that since you obviously have a licence, i mean if there is obviously a fault where you may have to format and reinstall? or at least crack the wpa.

[Brian Ellis]

Mrtj1

Have you ever tried to reason with Microsoft? They are the most unreasonable (literal meaning) company in this world. I have had two lots of correspondence with them about very real bugs in software, which they finally admit to after many exchanges, using pressure through the brother of a close friend working in the higher echelons of the company in Redmond, but the best that they have ever offered me is a refund of the money spent on the software (after over a year), provided I returned the CD-ROM and the original receipt, in one case. I have asked for, but never obtained, a free upgrade for a fault. Nor have they replaced the 2 Intellimouse Explorers which have a design fault on the cable, despite saying they would.

To answer your question, I am 99.9% sure they would not co-operate as you suggest. I recommend boycotting all WPA software.

[Karen St Martin]

Brian,

I couldn't agree more! I had an interesting exchange with Microsoft about a bug in their TCP/IP file transfer protocol. It seems that it is faster to get a file from a remote system than it is to send a file.

I went round and round with their 2nd level support types for a couple of weeks. They were actually able to reproduce the problem on their test systems. Their final response to me? That's the way it works!

Well, live and learn I always say. So what did I learn from this? Not to waste $35 on Microsoft Tech Support!!!

Karen

[Dr Mordrid]

Amen to that!! Their "tech support" isn't as good as info you can get free from the d**n internet.

I agree with Brian about the need to wait before upgrading M$ products. I didn't go to Win2K for a very long time in my own setups because of problems I was seeing in my beta rigs.

Similar experiences in XP rigs make me caution others about it. WPA just makes a bad situation worse.

ME was a non-starter here. I installed it in ONE beta system for a go using OHCI DV and that was enough. Even a pro-choicer would agree that ME was worthy of a late-term abortion.

Dr. Mordrid

[Duty]

Maybe this is a dumb question but, what is the advantage of using the matrox RR-G stuff over a generic firewire card? I am using a Audigy for capture in XP and love it so far. Is the struggle to get the RR-G worth the time?


Thanks,


Jeff

[Dr Mordrid]

It's handy to have analog as an option because DV doesn't handle some subject materials or added effects very well.

This is because DV has a relative over-abundance of its B&W signal vs. its color signal. This makes DV footage look good to the eye because it's bright and contrasty, but it makes life difficult when it comes to artifact generation. Bright sharp high contrast edges, particularly on a diagonal, are hard to encode either to DV itself or DV-to-MPEG without generating artifacts.

Solid color regions are also more difficult for DV to properly shade because of the relative weakness of its color samples. This too can translate into problems when making MPEG's from DV sources.

The end result of either situation are block artifacts. If on a diagonal or high contrast edge they look like stair steps. In a solid color region it often looks like an animated cloud.

One way around these problems is to capture high quality analog (typically HuffYUV encoded) from the DV cams S-Video analog port. This footage can then be incorporated into the DV project with a reduced artifact level.

This technique is particularly handy with DV footage used for keyed effects or for backgrounds with graphic overlays or titles applied.

When encoding edited DV to MPEG another method is to apply a very slight horizontal blur in the MPEG encoder, if possible. TMPGEnc has such a filter (it's the Sharpen Edge filter, you just use negative numbered settings to get a blur). Vertical blurs can also be applied if there are interlace or the artifacts are persistant after adding a horizontal blur.

Dr. Mordrid

[Flying dutchman]

Hi Doc,

still this makes me wonder. If DV captured over an analog port looks better, one should be able to obtain the same result using digital capturing and some filtering. After all, digital-> analog conversion shouldn't IMPROVE the signal, should it? Maybe it just LOOKS better, because the DV macroblocks are shifted with respect to the MJPG macroblocks?

[Dr Mordrid]

It's not so much a matter of "improving" the DV signal by capturing it to analog.

In fact its quite the opposite: you're actually trying to selectively DEGRADE portions of it, at least in terms of DV's contrast scale and the excessive sharpness caused by its over-abundance of luma vs. chroma.

The idea is to soften the sharp, contrasty edges often seen in DV footage. These high frequency areas are focal points for the generation of DCT artifacts both when the raw signal is encoded to DV itself and when transcoding DV to MPEG.

By softening and/or reducing the contrast of these edges the DCT compressions used in both DV and MPEG can handle them with a minimum of "gribbles" or other block style artifacts.

The same kind of contrast reduction or blurring also helps mitigate the moving block clouds often seen in solid color regions and dark backgrounds of DV => MPEG encoded footage.

DV's macroblocks aren't really shifted vs. MJPeg. Both are 8x8 and DCT encoded, although with slightly different Huffman encoding and DCT parameters.

One big issue is that the color samples in MJPeg are spread across just 2 horizintal pixels while DV's are spread across 4 horizontal pixels. In a 720 wide frame this gives MJPeg 360 color samples and DV only 180. Score one for MJPeg.

This is not too much for the naked eye to adjust to, but it can make MPEG encoders and compositing engines go nutsy. It can also bring about DCT compression errors in those high frequency areas mentioed earlier. It's also problematic when a colorspace conversion is done.

When encoding DV to the MPEG profiles used for VCD/SVCD/DVD there is a 4:1:1 to 4:2:0 colorspace conversion in addition to the a DCT recompression AND a macroblock conversion from 8x8 to 16x16.

No wonder it artifacts

Dr. Mordrid

[Brian Ellis]

I'll add a wee bitty of another thought. If you capture MJPEG from analogue at 704 (or 720) pixels wide, do all your editing etc. at this width and only at the end reduce it by half in each direction (e.g., for VCD) as you render it into MPEG1 or whatever, you will have as much chroma and luma as the system can use. If you do the same with DV, the chroma will still be lacking, although the difference may not be noticeable to the eye of a casual viewer.

You see, the eye doesn't care two hoots about chroma resolution. When I was a student (1949), I did my dissertation on colour TV and I believe that the team I was on was one of the first to realise this. To demonstrate it, we set up a still life bowl of fruit with fairly harsh 45° lighting. We took a Kodachrome shot of it (had to borrow the prof's pre-war Leica and plead with Kodak for the film!). With the same lighting and set up, we also took a mono photo on Ilford FP3 film. We then re-arranged the lighting to flat frontal lighting and took a series of Kodachromes, with over-exposures of 0, 1, 2, 3 and 4 stops. (Actually, we did it in a different order for practical reasons, but I am explaining it more logically here.) We developed the b/w film using a reversal technique, and waited for the Kodachromes to come back. We begged three 2" x 2" projectors from various faculties (difficult, as they were still mostly the 3.25 x 3.25, at that time). As it happens, they were all Aldis ones with identical lenses, so we were in luck. We set up two screens. On the LH one, we projected the first Kodachrome, which was to be our reference. On the right hand one, we started by carefully focussing the 1 stop over-exposed Kodachrome, which was a slightly wishy-washy, uninteresting image. We then switched on the b/w image and carefully superimposed it over the flat colour image. Lo and behold, it sprang into life and looked nearly identical to the ref image. We then de-focussed the flat colour image: it made absolutely no visible difference, even though, on switching off the mono projector, it had become an almost meaningless and unrecognisable splash of almost pastel shades. We repeated it with the 2-stop over-exposed flat image, with very little difference. The colours started to appear less intense with the 3-stop over-exposure, but were still acceptable. At all times, a very slight defocussing of the mono image degraded the visual impression, even if the flat colour image was sharp. We therefore concluded that the luminance signal had to be always as good as we could get, but that the three chrominance signals (we kept them separate in our experiments at that time) could be degraded in both intensity (provided all three were the same) and bandwidth without degrading the visual impression. In fact, our Luma amplifier had a full 3.5 MHz (or Mc/s, as it was then) bandwidth (405 line interlaced 25 fps), whereas our three chroma amplis had only 750 kHz bandwidth. We derived the luma signal by averaging the three camera output signals at full bandwidth. We hadn't thought of using the green as a luma and just sending the blue and red as low-quality signals! I wish I still had my carbon copy of the dissertation, but it got lost in a house clearance in the 1950s and the University doesn't have the original, either. I've therefore been aware that the eye is very tolerant to chroma signals for over 50 years, now!

[mrtj1]

as to the wpa issue in the start of this post,

i have emailed microsoft asking them if i could use the corp version. wonder what their response will be.

[Brian Ellis]

mrtj1

Silence!