Today I had the pleasure of receiving a CD with 6 VCD files on it from fellow poster Doug (email: MazinZ@aol.com - not a false email address like a lot of others on here).

This posting is going to be a long one.

Three of the files were ordinary rate NTSC vcd's and 3 were highrate NTSC vcd's at 2500kb/s from his Dazzle mpeg-1 capture card. Only two of the files could be compared for the effects of high and low data rate processing. I am a PAL user but with the equipment to play and view in native NTSC.

In a recent posting which was begun by Mr VCD I posted that I was amazed at reading the comments from those that really get stuck in about the "shitty" quality that the Dazzle card produces, well read on..........

Unfortunately Doug does not provide any ordinary home brewed video vision on this CD, nor does he provide a comparison file based on the dazzles performance when capturing at the low 1150 kbit/s rate of normal vcd's. These files were all based on capture data rates of 2900 kbits/s with re-encoding in the Panasonic standalone encoder at either 1150 or 2500.

I burnt a VCD in Nero 4.8.03 using a Kodak CD-RW with the normal rate files first, knowing my Philips DVD will not play the HR VCD - when it gets a HR vcd track it simply locks up on the first frame and will not let me go beyond that, important to burn the low rate files first.

The source files were from a VHS master tape (not quite sure what Doug means by that), a DVD and a 14 year old ex rental VHS tape. Therefore all were images from professional productions and source material, none were consumer level sources.

So here in lies the first problem when comparing quality, you cannot compare "apples and oranges" so you cannot compare a DVD source with Professional vhs and or with consumer level images either. You can only do that in the individual catagories. We should be more careful of what catagory we are talking about when we talk quality. Why? Well its simple a DVD source has a lot less artifacts and noise than any VHS production tape will have and if the VCD process amplifies faults then the lower level sources will suffer more. Its difficult to place consumer level sources in the pecking order, particularly DV source material.

I have several methods for playback and viewing (all were viewed from the VCD as burnt in Nero and not from the hard drive):

1) Computer with 17" monitor at 800 x 600 viewed from 20 inches. 2) Multisystem Sony Professional Studio Monitor - 14" viewed from 32" (450 line capability at the centre) connected in svideo mode. 3) multisystem consumer 26 inch TV - German Loewe, viewed from 10 feet, composite connected to the DVD player.

VIEWING ON THE COMPUTER:

I used the 10x Pioneer 104 DVDrom for playback from the computer, so there is no effects that result from high speed CD readers "mucking" up the image.

I used firstly the Cyberlink Wizard PowerPlayer at full width in 4:3 mode on the 17" screen - there was no noticable blockiness, not even in the sky or in scenes with smoke involved. The Cyberlink player would not play the HR VCD's so no comaprison of image quality was possible. There were no interlaced combing effects in the playback even in lots of movement.

I then used the RealMagic Hollywood Plus mpeg decoder card and software to play and fill the full 4:3 aspect ratio screen - the same comments as above apply except that all 6 files could be played and the hardware decoded image was noticably of higher quality than that produced by the Cyberlink software player.

There was NO "jump out at you" difference between the HR image and the ordinary data rate image in the DVD source material comparison. The image was marginally better at the higher data rate. The vhs images were all noticably down on the quality provided from the DVD source material and contained noticable video noise.

VIEWING ON THE STUDIO MONITOR:

Using the RealMagic decoder card I could play the vcd in native NTSC or in the on-the-fly conversion to PAL. There was no noticable difference in quality between the systems and I think such conversions from NTSC to PAL are better than the reverse as there are no padded frames added to make it jerky.

Again there was a noticable image quality differences between the DVD and VHS images BUT no jump-out-at-you difference in the DVD based comparisons of high and low data rates.

Actually its more difficult to compare a wide screen image than a full 4:3 image. The 2.21 wide screen image will easily give a false impression due to its reduced height, image magnification is simply not the same. Beware of the apples and oranges again. I would prefer to compare a pan and scan DVD image and I suspect you will find a very big down grade in the pecking order for that image.

There was no sign of blockiness on the TV monitor and remembering that it is svideo connected to the card that is quite an achievement from what some say is a heap of "shit"

In my view all VHS examples required a higher level of noise and video filtering to get a better image. Then again some of that could result from the source vhs vcr which was probably composite connected instead of Svideo connected - Doug your comment required here. The blue screen option on the professional monitor indicated quite high levels of noise in the VHS source material but the 14 year old tape showed quite a bit less. The DVD images were quite clean.

In each case the image on the professional Sony monitor was far better than that achieved on the computer screen which is a 17" 0.28 pitch screen.

ON THE CONSUMER TV:

Only the ordinary data rate files could be played and as expected the DVD source material was again noticably better - no blockiness at all was evident on the 240 line TV as viewed in full screen NTSC and as with all the other viewings the VCD played smoothly and was a good example of what I would call a high quality end product. Beware here also the comment about wide screen images being compared to 4:3 images, it will be misleading.

SUMMARY:

DOUG you should again list exactly what you do, I know that you have said it several times on this site but tell these people again exactly what you do in the process because what I have seen tonight is far from a heap of "shit" and I have not been paid by Dazzle to bend the reporting.

All you Mpeg dazzlers should take a lot of notice of what his guy says because he speaks from experience and more to the point, calls a spade a spade and does not beat about the bush, he can, even more importantly, support what he says with actual examples.

Well done Doug and thanks for the opportunity to view the stuff, its just a pity it contained no home brewed video vision for those of us that deal only in that phase of the VCD process.

In my view the Mpeg capturing process based on professional source material for producing VCD's is a lot better than some of the postings here would indicate provided you do it at the maximum possible data rate the card will give followed by a Panasonic encode, the best of bunch right now, BAR NONE! Wonder how the home brewed stuff comes up?

Cheers

Ross McLennan Adelaide South Australia

-- Ross McL (rmclennan@esc.net.au), March 20, 2000

Answers

Wasn't Doug the guy that captured with his Dazzle and encoded with the Panasonic encoder? Nobody here has ever questioned Dazzle's ability to capture to AVI. I had the Dazzle DVC and it did a great job. The point was why buy a Dazzle for $250 if your only going to use it as a capture card? There are tons of quality capture cards out there for a lot less than the cost of Dazzle's fine product

-- Al McCraw (amccraw@ix.netcom.com), March 20, 2000.

I own a Dazzle (USB), but it's taken it's rightfull place in the closet. It just couldn't produce the quality I was looking for. Even using ultraclean source (DVD) VCD legal captures look blocky. Yes, higher bitrate captures looked pretty good, but if even after re- encoding I wasn't happy with the results. Granted, this was before I had access to the Panasonic encoder. I'm sure I'll find a non-VCD related use for it, but trying to make VCD's with it was just too frustrating.

I do think it would be a great idea to establish some testing standards that we can put various VCD production processes through and publicly compare the results. The related files could be posted in alt.binaries.misc or a donated FTP space so those of us that wished could download the related files for comparison. 30 second test clips result in 5-6 meg files, which even on a modem isn't that bad. Those lucky enough to have broadband access could even share the source files. Everyone has slightly different hardware, and everyone does things a little differently, so it would be great to see how all the permutations affect image quality.

Then when a new encoder or capture device comes out, we can all see for ourselves what it can do with one of the standard test files.

-- Sean (sean@magnuminvestments.com), March 20, 2000.

Thank you for the comments!!! yes when i was not using the dvd player for an image (super vhs), i used my regular vhs player which was connected via composite. i do not own a supervhs. What i meant by vhs master tape was an actual original copy of a movie, not a copy from a copy (although as i stared one clip from my highrate vcds was indeed this). i was going to put some hi8 stuff i did on the cd as well, but i did not get the chance to, nor some other regular homebrewd vhs stuff ( but hey it was pretty cheap to send you a cd, so another might be in the mail to you Ross). As i have said before, i never knocked avi at all. I have one card that lets me capture/save in avi and thats the studio mp10. But i found the dazzle easier and actually better to use. Its also very convient for me to capture in larger files then the 20 crap (hate you bill gates)2gig limit they give us now.

-- Doug (mazinz@aol.com), March 20, 2000.

oh, Sean i like your idea.

-- Doug (mazinz@aol.com), March 20, 2000.

Thanks guys, I expect some negative comments.

I think this whole quality thing was about the ability or otherwise of Dazzle in Mpeg-1 as the capture mode. Doug, as I understand it, does not consistantly use avi and that aspect has not been discussed in detail in his postings. So I do not know how avi's got into this discussion. In any event what he gets from the dazzle in mpeg mode is way above the impressions left by other posters on this site.

I agree we all have different equipment and computers, in fact I would suggest to you that I have the lowest level of computer making VCD's on this web site but I have screwed it to get results and its strange that some put things in the closet where others get extremely good results from self generated get arounds. That is a fact of life, but do not you wonder why - gezzzzzz I do and this site has been an enormous help in achieving effective get arounds.

Take these facts shall we: I updated my video card to a Voodoo3 3000 and it lasted 3 weeks before I went back to my old Diamond Stealth 2000 with only 4M on board because the Voodoo gave me blocky results on the computer screen from the same source material. $150US in the waste bin. LSX encoder $300US not used because version 3 does not perform anywhere near the Panasonic. I bought a RealMagic decoder card because Matias in Beunos Aires posted that for the price of $80US it was a fabulous asset - he was so right. I downloaded a latest drive from their site and promptly went back to the previous one because the update produced blocky results on the computer. Finally after 4 years of struggling with a miro DC20 in analogue which cost me $980US thats also in the wastbin replaced by a $100US 1394 firewire card that provides the best home brewed results yet.

Not mentioned in my posting above is the fact that under the simple blue screen test which effectively removes the color overlay to see the underlying hash there was just some evidence of blockiness in the VHS source and it would not take much with mis matching equipment for that to be quite a problem and one really has got to ask where does it come from? OK I will suggest that the problem can lie as far back in the process chain as the VCR because VCR's as noise generators are like chalk and cheese even in the SVHS class there are big differences between brands. Perhaps it could also be the result of not using enough noise filtering or video filtering in the encoding process.

The best method of checking how much comes from the source is to author mpeg stream stills in a NLE program as that will cut out the source equipment and the capture card. In fact you will find LSX for example has more trouble encoding stills than it does with motion. Get the mpeg stream stills right and the video will follow.

I guess I am saying that it maynot be the card that is at fault and Doug has developed the right combination to make the mpeg based capture system work. Maybe some of us have the wrong stuff in the closet OR are simply not bothering to try what others find as successful. Poor quality is then self inflicted!

-- Ross McL (rmclennan@esc.net.au), March 20, 2000.

Whatever

-- Al McCraw (amccraw@ix.netcom.com), March 21, 2000.