Thanks for all the responses; I didn't want to start any DV vs MPEG-2 argument. :)

I may consider a ATI TV Wonder due to the relatively cheap price.

I actually have a spare PCI-E slot, and have noticed the HD600/650 are available as PCI-E.

The spec sheet mentions 'Hardware MPEG-2' capture, only on the 650's. What is that all about?

Thanks for all the suggestions.

Any suggestions for a cheap TBC?

^ Are you wanting to stabilize horizontally or vertically?

:huh1: How would a computer create a whole color channel out of nothing? You are misunderstanding what the notations mean. Fact check.

4:2:0 doesn't mean one of the chroma channels is discarded. If you threw out Cb or Cr, you would be left with red-tinted or blue-tinted video. The scaling is applied equally to both channels regardless of which subsampling scheme you use.

With 4:2:0 you get 1/2 horizontal and 1/2 vertical. With 4:1:1 you get 1/4 horizontal and full vertical, which means you shouldn't have to worry about scaling interlaced chroma differently than progressive. 4:1:1 is considered equivalent/superior to Betacam SP's chroma resolution, so it sure as hell exceeds the pitiful color of (S-)VHS. More fun comparing pro formats here.

If you're going to DVD from 4:1:1 you end up with the worst of both systems: 1/4 horizontal and 1/2 vertical (the equivalent of 4:1:0 I guess, but stored as 4:2:0).

It's all academic because VHS color is low-res and noisy. If you take a look at my examples, the artifacts caused by an intermediate conversion to RGB have far higher resolution -- and those are invisible. I was surprised to find that 4:2:0 is insufficient for highly-saturated color transitions captured from LaserDisc, though. I think 4:1:1 would smear the color bars horizontally but I haven't tested that yet.

It's still hard to type, but a conversation like this almost begs/screams for my input! I can't sit on the sidelines for this one...

Yes, MPEG is (can be) an end-product only. Keyword "can be".

• It's used as a delivery format (Amazon, iTunes, Hulu etc) -- we've done work with them in the past (through another party), something most "video transfer services" cannot claim, FYI.
• Before HD came onto the scene, it was an editing format used by Sony, studios, amusement parks rides (IMAX-like), etc. Circa 2000-2004 timeframe. I used to have contacts in that industry, before most retired.
• The GOP can be not just IPB, which is the end-product format, but IP and I-frame only. When done as I-frame only, it's the same as MJPEG, which was a popular lossy/lossless format pre-Huffyuv. And being 4:2:0, it's better for NTSC than DV and this spec without a doubt.

I'd say Sony MicroMV failed for many other reasons. It had nothing to do with the MPEG. We offer that transfer service too, but it's a complete PITA because the cameras are lousy to work with. The video looks good, as does DV, but those home cameras really look the same as my $100 AVI (MJPEG) Canon camera from 2009. All of those home cams are a joke compared to my pro Nikon digital SLRs with pro glass.

I don't really think Betacam SX or MPEG IPX are anything other than aborted formats, to an extent. It's just something you read about on Wikipedia more than anything else. MPEG is not a good shooting format anymore than DV is a good conversion/intermediary format.

PAL and NTSC both need to be as close to 4:4:4 as possible. Due to the color method, 4:2:2 is just as good, and a close approximation to the analog signals. 4:2:0 is next best. Bringing up the rear is 4:1:1, which just really smears the color on anything converted with that format (i.e., not shot). Maybe if you converted 4:4:4 to DV, it would be okay, but converting 4:2:2 is dreadful. Real life -- aka shooting -- is like converting 4:4:4, and that's why it's fine. 4:2:2 converts far better to 4:2:0 than it does 4:1:1.

Do note that some DV codecs suck, as there's not a single way to decode it. Nothing was ever standardized, as it was too short-lived.

This site is far from the only ones that suggests lossless AVI (uncompresses AVI), or to a lesser extent, MPEG, and not DV, for converting analog video (especially VHS sources).

• VideoHelp does, though it's mostly about converting downloads, copying discs, and other low-end user needs. Many of those users are far too cheap (EzCap, VHS VCRs, converting tapes instead of buying DVDs when they exist, etc), which makes it really aggravating. Especially for long-time members like myself.
• Or Doom9, though it heavily focuses on DVD copying and Avisynth use -- especially the latter in more recent years.
• The only sites that are pro-DV are the sites for the cameras themselves, mostly for shooting and how to's. But those are dying as the users migrate to newer models. Several are already gone. The biggest site is/was dvinfo.net, but those once daily conversations are now weekly or even monthly. It wasn't common to see conversion topics on those sites.

(DV magazine is apparently gone now, while every issue of Broadcast Engineering still has lots of info on MPEG. Even Videomaker magazine still has MPEG-mentioning articles, which is a member of this site, even in the era of Blu-ray and HD.)

One of the big problems I've had for years is that people don't think non-"Hollywood" video can be any good. But it's caused by crappy VCRs, crappy capture cards, and crappy DVD recorders. So in addition to those who want to defend their purchase, there are those that thinks it's the best possible when it's clearly not.

It's not so much an argument as an educational post, with some disagreements. I'd rather things like this be made known, rather than people make quiet assumptions.

I'd not really anti-DV, but there are limitations that you/anybody should be made aware of before going down that path. For some, DV is the only option. For other, there are really better choices to be made. But regardless, you should know everything first -- not just read the propaganda put out their by Canopus. (Note that since Grass Valley bought out Canopus, much of the BS is gone. But the old Canopus info still lingers online, and with its users.)

Do not buy a 650 card. It has uncorrectable AGC issues.
Get the 600 card. Preferably, get the ATI 600 USB card, perhaps not the internal PCI card.

Gosh -- this took me forever to write. :mad4:

That's a really good explanation. :salute:

I always neglect to include that in my posts. Unless you plan to watch DV as DV -- which is pretty much impossible -- double converting is really where the issue lies. The end format is always going to be DVD, Blu-ray, H.264 streams, etc.

However, I would note that even analog source 4:2:2 has issues converting well to 4:1:1. It's probably not so much the issue with DV (in theory) but with the converter boxes/cams themselves, and/or the codecs. I can always tell when something was DV converted, because the colors are cooked (IRE and gamma off), because the color detail is lacking, and because there are often artifacts in the chroma.

Not to go too far off topic, but this is a good read: http://www.dvxuser.com/V6/showthread...-2-0-and-4-1-1
It's rare that somebody prefers 4:1:1 (for conversion or shooting -- but especially conversion).

Getting back to the original topic is important. Let's not overlook/forget the OP. ;)

It seems to me that the 4:1:1 color space itself is superior to 4:2:0 for storing old home video formats. It's a shame that the hardware to capture to it apparently sucks and the hardware to play it back is nonexistent. It would be nice if MPEG had offered it as an alternative for 1/4 color storage.

Do you have any experience with the ATI HD 750 PCIe card to check for the AGC issue? I know there are complaints about the USB stick, but that's apparently different hardware (no 3D comb filter).

I've been evaluating inexpensive USB sticks and I prefer the Diamond VC500 to the ATI 600. It doesn't have the issue that was recently pointed out of the 600 filtering S-Video chroma bandwidth, and at least with my DVD player it seems to get the levels closer to the original file.

MicroMV failed because of typical Sony "NIH". It was a format addressing a problem nobody asked. Yes, it was small for 2002, but was quickly eclipsed by flash media recorders and it was pretty darned expensive. The higher end cameras were on par with the MiniDV camcorders Sony was selling at the time. Also the format had a major weakness, it wasn't closed GOP! Every time you stopped and started recording, the camera didn't write an I frame, so you had garbled macroblocked video for a few frames like a low bandwidth web stream.

Review with sample footage:

No, MPEG is the worst capture codec both with hardware and software. I've done several of my own tests over the past ten years, and MPEG I always find has crushed blacks, even when I've lowered the black levels, and it has multi-generational issues very similar to S-VHS. Plus I can always tell when stuff has been captured via MPEG because the color smears when played back by anything but composite. I've seen a number of DVD's made by my competitors here in the Ottawa area, and the vast majority of time,e ven on playback via composite, the Colors are smeared (and a number of them hav come from set-top DVD players as the menus can attest).

I've even run these DVD's through the ADVC-300 (besides VHS) and I've achieved results that are equal to or close enough to look like Betacam SP.

DV is the superior of the two codecs. I'm not arguing that uncompressed is the ultimate way to transfer, but MPEG is garbage. I worked at an independent production studio years ago, and we would occasionally get footage in in MPEG, and we would always hate working with it, since even using Handbrake and Fnal Cut Pro, the footage never looked as good as the stuff we were shooting in DV/DVCPRO (at first we used DV then went to DVCPRO), and even when it broadcast it looked even worse. But that's not surprising with MPEG's 10:1 compression ratio vs DV's 5:1.

DV is the superior format.

Then you have never seen high def. captured mpeg, quality is very good.

Tom is wrong - and every expert on the forum seems to concur


back to the OP:
all in wonder cards for PCI-E using XP are the x600 x800 x1300 x1800 and x1900
if you are using 7 use an ATI600USB

troll ebay for cheap TBC's
i scored a Datavideo TBC-100 PCI card recently for $70 and free shipping

the TBC-1000 and AVT-8710 are recommended
i have had a For.a FA-125 and a Keywest BVTBC-10 and thought they where pretty good units too.


dont mess with any of the old big rackmount TBC's

None is this is due to MPEG, but rather other factors in a workflow -- colorspace when encoding/editing is most likely. IRE (0-255 vs 16-235) is another. Remember that NTSC (USA version) video should appear "wrong" on a monitor vs. a TV, even when using an excellent calibrated LCD IPS monitor. Using a cheap non-IPS LCD is, of course, unreliable and a horrid idea for even serious hobbyists.

There's nothing special about S-VHS that's any different from any other analog format. After one SP generation, you can always tell when something is a copy of a copy -- VHS, S-VHS, 8mm, Hi8, Betamax, BetacamSP, U-matic, etc.

The only thing that can be noticeable is macroblock artifacts. Most of that has to do with content and bitrate -- i.e, not enough bitrate, or a kind of content that looks bad with any DVD-Video compression (wrestling, for example).

That's not at all suggested. If you want to fix MPEG, use Avisynth. Analog methods are not going to do anything other than reintroduce noise (or a new noise), and blur out details.

You used a Mac in a Windows world -- that's why. Mac is a DV only workflow. And that's why Windows -- Windows XP to be specific -- is the preferred OS, due to the tools available. On a Windows machine, editing MPEG would have been a painless process, even 10+ years ago. Understand I'm not anti-Mac, and have used Macs for 25+ years, but this is a Windows task.

No, I've gone through all the settings, and I know that the video looks different on TV's vs. computer monitors.

And I'm no hobbyist. My work has been seen all across Canada, the US and Europe. I am a professional.

You can spot tiny errors, but nothing like S-VHS on the first generation.
The stuff I was transferring was not wrestling or high-action stuff. We are talking about 'talking heads'.

I have used Avisynth, Xoom MovieClone and a variety of other software converters. They are all garbage, since I was always ending up with choppy video, unsynched audio, video that looked like it had come from VHS SLP tapes.

Personally I prefer the Premiere Pro line (although I can use any editing system), and I even asked Adobe why I couldn't get decent quality MPEG transfers (and I even tried reinstalling the MPEG codec for PPro). Adobe told me that I needed to remove the MPEG compression completely, otherwise the MPEG codec is so compressed at it's base that it is not good for capturing (and the only way was to either capture it uncompressed digitally or send it out via analog). Adobe suggested that I use a hardware codec (not a software codec), such as the ADVC-300. I used it and I was impressed at the quality that it offered over the software codecs. And I have seen High-Definition MPEG, and it is good, but we are talking Standard Definition here. For Standard Definition MPEG is garbage.

For my own personal work and that of my business I use PC and XP for my projects. And I find that, even on PC, DV is the superior performer. I have tried numerous MPEG capture cards as well and I was not getting the quality that I was looking for.

I found this interesting article written by Adam Wilt online regarding 4:1:1 and 4:2:0:

Again, for the OP, in my Professional Opinion, I would recommend importing his VHS footage by DV rather than MPEG, and using a hardware-based encoder such as the ADVC-300.

i feel very sorry for your customers who obviously cant tell the quality they are missing out on.
anyone using 4:1:1 DV to capture customer videotapes is not a pro

you cant make chicken salad out of chicken **** :wall2:

this thread should just be locked - OPs questions have been answered - i sincerely hope decided against DV

That's very offensive and a low blow. I could say the same about you, because I find, through all my tests and those of other people that I know, that 4:1:1 DV is a lot better than 4:2:0 MPEG for capturing.

This may be one argument where as Professionals we must agree to disagree, because I know that I am an expert and Professional and I only offer High Quality Professional Transfers, and I agree with Adam Wilt there about interlace sources already suffering from a loss of vertical resolution quality due to the nature of interlace and I find that 4:2:0 just adds to that loss, not to mention that the MPEG generational loss does not help; so why add to that loss with 4:2:0 just to recompress it again for DVD?

In this case it might be wiser to suggest to the OP that he try both and decide for himself.

i apologize - that was a pretty obnoxious post.

i guess if i was paying a pro service i would want them to be capturing lossless.

the OP is capturing PAL.
does the canpous device capture PAL sources as NTSC 4:1:1 DV or PAL 4:2:0 DV?
because if it is PAL DV this entire argument was pretty moot

The ADVC's capture PAL DV 4:2:0.

But even for NTSC I would still go with DV.

Disagreements are fine -- but behave.:buttkick:

I had to rename the post title to include MPEG vs DV. It is a pretty good discussion, although it did derail the thread some.
The OP is free to open another thread for more questions.

Thanks for keeping it professional. :)