Capturing NTSC HELP!!!

[DigitalfreakNYC]

I am using and ATA133.

I'm not worried about the "black bands" any more so much as the AVI situation. Any way to help that? Will there always be that jump cut no matter what program I'm using?

[jggimi]

No, I think your software is "unique" -- Vdub doesn't do that, for example, and I capture in 2G chunks. When I was capturing in MPEG2, in 4G chunks, the tool (it came with my AverTV card) didn't do that either.

Of course, with the latter, I might not have noticed it, considering the frame drops I was getting.

[DigitalfreakNYC]

*sigh*

ok...so i guess i'm back to square one..:( :( :(

What video card should i get that will suit my needs?

[DigitalfreakNYC]

What I meant by that was...

What video card should I get that will work correctly with the tools I am using (AVI_IO, TMPGE, Vdub) and will also capture in high quality (720x480) AVI?

Just want to make sure I don't screw it up this time...

[FredThompson]

It all depends on how much you want to spend. As I've said, most internal cards that people commonly use WILL wash out things like sand and caucasian skin. Forget trying to do a good capture from Lawrence of Arabia with those.

If you can hold for a few days I'll put together some samples to illustrate this. I'll use my ATI All-in-Wonder Radeon, DV passthru (which is more expensive, obviously), Instant DVD, and Instant DVD 2.0.

Another thing you might want to investigate is system noise. On my links page are Audacity and a free noise gate filter. Try them and see if your internal sound card will give sufficient performance. The dynamic range of videotape can be quite large. It would be a real shame to do a large colume of captures only to later decide you can't stand the squiggly underlying noise.

[DigitalfreakNYC]

thanks fred.

I guess i'm just a simple guy. At this point, i'd be thrilled if the damn thing WORKED like it's supposed to.

Don't get me wrong...i was very impressed with the resulting picture from the AVI->mpeg2. The ONLY Problem was that it dropped frames when it switched files. Which is a BIG problem obviously...
leading to my dropping the damn thing altogether.

Do all external devices go to MPEG2 immediately?

[FredThompson]

Thing is, you'll get frustrated watching a Caucasian face fade into the background when they're by concrete, a water surface getting whited out, that sort of thing.

I don't know about ALL external devices. The reason for doing it externally is the developer can control the interface (USB instead of PCI) and be free of the radio noise inside the box. Can't imagine why there would be such a device that DIDN'T have an encoder chip. They aren't all the same, however. MPEG2 is a standard with many different parts. You can have cruddy MPEG2 just like cruddy MP3s. Personally, I'd rather have lossless but the image quality of internal cards isn't as good as the external devices with dedicated encoders. At least, that's the way it is now. My guess is ATI, Matrox and other combo card manufacturers have to take a guess at what people will really pay. Most people would be just fine with a TV tuner and occasional sampling so they use cheap components.

[ooops!]

Hi Fred,

You mentioned in an earlier post that you've got your hands on the new ADS 2.0.

Have you managed to do any tests yet compared to your previous device. I also wondered if the audio sync issues had been resolved with the Mk2.

Apparently I've got to wait until July before the Mk2 is available in England.

Nice to see your still dishing out good advice.

Cheers for now.

[FredThompson]

No, haven't tested it yet. I've been working ~20 hours a day on business stuff which included rebuilding my server. Blech!! Haven't had mroe than 5 hours sleep since Sat night and I'm just a little pooped right now.

Audio sync shouldn't be any problem because there's an audio digitizer in the box. I'm interested to see what the filters can do. 2x oversampling and the comb filter seem really nice but they talk about additional filters. Been playing with some post processing stuff. I've found anything that gets rid of dropouts shouldn't be used for night shots of tracer fire or some rainy scenes ;)

As it happens, I got a multi-format VCR the other day and I'm going to test PAL and NTSC captures.
I think the sync problems were caused by the wide range of hardware and people using ... uh ... inefficient computers. (that's a nice way of say they weren't up to the task.) Granted, you could use USB1.1 but it WAS a stretch. Oh, a lot of people also wouldn't keep up to date on drivers. Duh.

The only problems I ever had were related to a computer that kept trying to share the IRQ with the scanner and some other stuff. It was only USB1.1 and that screwed everything up.
I've got to finish some stuff up and make a PAL tape tonight so will have some results soon.

[ooops!]

Thanks Fred

You better go get some sleep!

[DigitalfreakNYC]

Hey guys,

an update.

I got a hauppauge card yesterday and installed it today. It was a hit and miss situation. The good news is that Vdub and others now work and I can get picture. The bad news is that it won't let me capture 720x480 AND the picture I do get is really red and has lots of interferance coming in. I see differences in the way some of the programs are handling this card as opposed to my other card. But I do have to say that the quality of the Flyvideo was MUCH better. I may try and reinstall it and see what happens. Gonna have to return this Hauppauge to J&R music tomorrow.

The journey to capturing continues...

[FredThompson]

Well, could have told you about the Hauppage quality. They're using very low quality circuits.

ATI, Matrox or a dedicated external are know good quality. However, the first 2 won't do that great on the visual extremes.

If you want to digitize an analog signal, you need at least 2x and prefereably 4x the maximum frequency to accurately map it.

There's a LOT of electrical noise inside a computer, especially running at the speed you've got.

There's no two ways around it, if you want quality, you've got to shield from the noise which means an expensive card because the engineering and manufacturing must be better or go external.

[DigitalfreakNYC]

So if you were to go internal, what would you suggest?

At this point, i'm beyond believing that ANYTHING is going to work with what I have. :(

[FredThompson]

My preference for sub-$$$$ is ATI. Matrox is also supposed to be good but ATI seems a little more popular. If you've got $$$$, Canopus, no question.

Thing is, any internal card is subject to the system interference (which you noted with the Hauppage) and a combo also has video display circuitry and a tuner. All that costs $ so it's not fair to compare a $200 internal card to a $200 external encoder. The external one will almost always win.

If you'll wait another day I can make some samples to show what I mean. I'll just go from DV to ATI All-in-Wonder and ADS2. Guess I'll have to do a framegrab and post it somehow.

The only benefits I personally see to an internal card are:

1) no external power supply
2) typically a TV tuner (but I've got DirecTV so it's useless)
3) no need for USB2

The drawback is:

1) lousy quality (in my opinion)

Just give me a couple of days and I'll have some samples to show you. If the ADS works as well as I expect, it should be very easy and more accurate than using the "normal" methods typically described here.

I know you're anxious, so am I, does this thing perform as well as they say? Dunno yet, I've literally been working 18-20 hours a day all week.

If it does, the process should be very simple and quick.

Having said all of that, you might want to consider buying a REALLY CHEAP capture card to use as a way to learn and in a little while buy something better. Even if you decide to spend $200 for an external device, if it works properly it should save a LOT of time doing cleanup that you'll have to do with the "normal" methods.

If you can wait a little while, I'll have a review done. Don't know how I'm going to handle video clips because that's an important aspect and, I suspect, any clip posted here will be a monstrous bandwidth hog.

Oh, I know, I'll post to alt.binaries.multimedia.utilities

Uh...if you don't know how to use a news reader, learn. I'll post some sample clips in the next couple of days. Will start a new thread to make it visible so people can grab them. Maybe I can also get a GeoCities account. I'll look into that while I'm running some compress jobs tonight.

[DigitalfreakNYC]

Fred,

I will wait for your comparisons. As I said, I just want something that will actually WORK on my computer. Thus far, neither did. Oddly enough, if you combine the good things about both, I'd have a winner. ;)

I can't really get a "cheap" capture card at this point to screw around with because A) i think most would agree that both of the ones I've gotten ARE cheap and B) there's nothing really to screw around with since i'm not going to bother burning dvd's of stuff i'll never watch. It's all slightly agitating.

I guess the biggest reason why i was "set" on the internal was the avi-mpeg switch. I was really thrilled with the difference that I saw from TMPGE and I was hoping to go through that again. I may try the lifeview one more time as a fresh start tomorrow in the hopes that it may work out but i'm still doubting it.

I'm just warnin' ya, Freddie-boy. If i get this external card, you may just become my official capture babysitter. :) :) :)

[FredThompson]

Originally posted by DigitalfreakNYC
I'm just warnin' ya, Freddie-boy. If i get this external card, you may just become my official capture babysitter. :) :) :) I suspect you're not the only one. Take a look at the number of "reads" for this thread. That's another reason I want to do a review properly.

[DigitalfreakNYC]

Any updates, boys and girls? :) Just curious...

i'm trying to hold off on buyin' an ATI ;)

[FredThompson]

Probably this afternoon. I had four DV tapes on the drives and needed to get them cleaned.

Think I've also found a tape with a guy walking through sand which should illustrate what I meant about things getting washed out.

Tape will be played on VHS into DV then I'll be able to show some comparison. Tape deck is a good JVC S-VHS with genlock and digital stabilization/cleaning. I don't own a DV player so this is the best you'll get from me.

[DigitalfreakNYC]

It's all good fred.

I speak for myself when I say that I appreciate all the help you've given me and thank you.

Looking forward to the comparisons and finally making my decision!

[FredThompson]

Well, I didn't test the hardware today, sorry.

I've been struggling to figure out how to properly recover film from a progressive NTSC stream that's been processed.

The goal, of course, is to restore the film frame rate so your file is smaller and accurately matches the film. Smaller file means more data available per frame and an accurate match means to don't get extra goofy smearing.

Well, I've figured it out. It's fiendishly simple. I was trying to do it from inside, while the file is being created. Wrongo, write the file then go back and corrupt the frames that need to be replaced.

You might not understand all of that. Suffice it to say one of the biggest bugaboos of working with film that was moved to NTSC has just gotten a lot smaller.

[DigitalfreakNYC]

Just postin' for a check-up.

What video cards is everyone else using? Is everyone using this external one?

[markrb]

I hope I am right on this one. I am just as new to Capturing as many of you here, but I have been reading up on some other boards.
Most on these other boards really like that FlyVideo card and I am considering replacing my X-Capture with it since the X-Capture has some picture flaws from noise on the card.

I think I might be able to shed some light on the original situation of not being able to use Vdub any higher then 352 or so.

Vdub uses a Windows wrapper with these newer cards since I believe it doesn't know these cards. The Windows wrapper forces the resolution lower. If you use a program like IUVCR then it can directly attach to the drivers and use any supported resolution of the card. It does not use a wrapper if the card is internally supported. I have the X-Capture card and Vdub does the same thing. If I use IUVCR (commercial) it works at the full 720x480/576.
The problem comes in that IUVCR doesn't hold 29.97 very well. It will often stray by a few decimal points.
I just use Vdub to adjust the frame rate to a real 29.97 from something like 29.963 or the like.

I have not gotten to the joining of files yet. I use NTFS, but since I am doing Laserdisc I decided to make each side a seperate file and join them later. I happen to have a fairly large (240GB) partition for this. I am almost at that stage. Right now I am adjusting the framerate of the last file to match. Hopefully it will combine and everything will be nice. Then it's on to CCE for some DVD conversion.

Mark

[FredThompson]

You shouldn't have any problems with Win2K or XP and a decent card with VirtualDub. Dunno about NT. DirectX has taken care of the VfW-WDM bridge. I still think it's a driver issue but the Fly card, in particular, doesn't seem standard. When I did some research I saw the video chips are very oriented towards PAL. Given the original post is about capturing NTSC, it's probably not a great idea.

wrt, sample grabs:

I've finally gotten some space at Geocities and here's what I'm going to upload tonight:

torture footage of a white guy walking through sand

1) VHS plain
2) VHS with genlock, comb, digital stabilization

Both of those will be from DV connected to the S-VHS deck

3) ADS Instant DVD 2 sample of #2 from camcorder s-video
4) All-in-Wonder sample of #2 from camcorder s-video

I have no intention of making a plain VHS grab. If you're going to buy a video grabber at $200, you should also get a decent playback deck. Not being snotty, being reasonable. You'll spend far too much time trying to clean up composite feeds. It will all be plain when you see the clips if you don't already understand.

The new ADS has 2x oversampling and a pro DVD encoder chip. We'll see how it performs, huh?

I might also do the Instant DVD 1.0 for comparison but that would have to come later.

Oh, so will PAL VHS samples.

[DigitalfreakNYC]

Mark,

Good luck with the flyvideo. As i said, i was VERY impressed with the quality of the capture but REALLY aggravated with not being able to use any other program to capture besides what it came with. As far as your capturing, do you do it in MPEG2 or AVI? That's my biggest problem. The joining. If it would only join properly, I'd keep the damn thing and not go through all the trouble that I have as of late. I spent $1500 on a computer that I can't use to do what I got it for!! :(

Fred,

Where can I get the ADS Instant DVD for a decent price should I be satisfied? Is there an online tutorial on how to optimize? And will there be a way to clean it up like TMPEG does?

Thanks!

[markrb]

@FredThompson I haven't upgraded to DX9 so maybe this is my issue. When I get the chance I will. I know that my Brooktree card will not with the Windows Wrapper in any resolution higher then 352 x 240 for NTSC under DX8 in XP.

All I know is that in Vdub if I try to select any resolution greater then 352 x 240 I get an error saying that the card does not support this resolution.

I will try DX9 and see if that helps. I need to upgrade anyway as my vidcard was made with DX9 in mind.


@DigitalfreakNYC I am using IUVCR to capture to an AVI @ 720 x 480 with 48 audio.

Mark

[DigitalfreakNYC]

Any pics? Responses? I need to just purchase the ADS i think. I'm sick of my computer just sitting there. :(

[FredThompson]

Sigh, yes, I'm trying. I'm a week late getting some PAL stuff converted and the ADS was to do it. (ATI cards can read both PAL and NTSC but the image qulaity sucks in comparison.)

Yes, I think the ADS is going to be a far better choice. The ATI and other internal cards are not capable of the kind of quality you can get from a dedicated MPEG encoder chip. I don't know how to explain other than to make ana alogy of a DVD to an AM radio.

It's up to you. Right now I'm caught in the middle of an argument between an American and a Frenchmen and the phone lines are about to melt. Both think they're correct and neither wants to yield. ;)

If you can hold off a day longer, I'll get some samples. I have to use the ADS to export the converted NTSC to PAL VHS anyhow.

[DigitalfreakNYC]

OK fred...i can hold off.

But a couple of quick questions:

Where can i buy the ADS?

What do you think about the Canopus and Datavideo as choices? as opposed to the ADS?

One other question: just did some research on 2.0 and it says it only captures at 704x480. It can't do 720??

[FredThompson]

adstech.com or do a web search. The 1.0 is going for around $100 now. List on the 2.0 is $200, upgrade was $130 but you had to already own and have registered a 1.0 version. Dunno street price.

Canopus is very good stuff but costs twice as much. I think you mean the ADVC-100? Performs very well but is DV so the files are larger. As far as I know it doesn't have any filtering to remove noise and it's reported as modifying dv that is exported as analog. Problem with that is you don't know what you're really getting if you play through it. My new camcorder does DV passthru. I'm sure the performance isn't as precise and firewire connectors looks like they're going to crinkle at any minute but you can probably see why I didn't buy the Canopus.

Canopus does make USB MPEG-1/2 converters with a list of around $450 or so. When money isn't as tight, I'd love to have Canopus. Well, probably get one of those firewired JVC HDTV recorders first...

The ADVC-100 was actually my first choice but the size of DV is a real bear. 13G doesn't seem TOO bad until you start using editing files then it's a bear. It's not uncommon for me to have around 35G tied up during editing a single session. That also gets awkward because it slows a drive down. Best to use 2 drives but then your work files are spread all over and it can be confusing.

Don't know about the Datavideo. Do you mean the DAC? It's also DV, not MPEG.

I also don't remember if DV has the same colorspace as analog sources. Something to consider because editing with multiple steps can change the nature of the color during conversion. I'm still working on my methods but if you're output is going to be MPEG2, shouldn't you capture and edit with the same colorspace? I think so.

ADS also makes a DV passthru.

FWIW, you can get a DVD+R writer for around $130 in the States now. http://www.techdeals.net/ I want DVD-R or a combo. Those prices should drop in the next month as well.

[DigitalfreakNYC]

Fred,

Yes, i'm talking about the Datavideo DAC-100 which is rumored to be just as good as the Canopus.

U also didn't answer about the 720 versus 704. Can you do 720 with the ADS?

I guess my biggest hesitation is the fact that it does "on the fly" MPEG2 and i really am a huge stickler for quality. I haven't read good things about this thing on other websites and everyone seems to be pointing towards the Canopus. Does the Canopus save to AVI? Does it work differently?

[FredThompson]

yes 720x480 or 720x576 (or whatever PAL is)

on-the-fly isn't really a proper description. MPEG encoding is temporal, meaning, one frame is usually somewhat dependant on the encoding of a neighboring frame or frames.

What you've seen is people talking about general-purpose or low-quality encoders. If you're encoder is high-quality and there's a big enough buffer, inline encoding is fine. I wouldn't necessarily use it for the ultimate squeezing of byte usage in a limited space but it can be very, very good.

All satellite TV is MPEG2-based and the vast majority is encoded on the fly.

ATI and other similar captures us the computer's CPU to do MPEG encoding. It's not optimized for that. You could, to make an analogy, use a compute inline to act as a distortion device for an electric guitar. Compare that to a pedal effect. Given $50, which is going to give the better performance.

Bandwidth is also important. If you pull MPEG-2 at 6000-8000 bitrate through a good hardware encoder you're probably not going to see much difference from DV. Having said that, DV is locked at 720x480 and your source ISN'T doing that resolution so it's not as accurate a grab. I used to say grab at as high a resolution as possible and there are very valid reasons for that. However, I have a suspicion about how the oversampling works in the new ADS and that might not be a valid thought anymore. If it isn't, capturing at the source resolution with a high bitrate should yield a proper-quality file at about 1/4 the DV size. That's no small issue.

Can't compare an apple to an orange.

AVI is a file container, it can hold many different types of encoded files. If you pull video through a DV converter you're probably going to store as AVI but the content is DV.

The Canopus is good, no question, also costs more and the files are a lot bigger.

--

You DO realize you'd have to downmix DV files and they're captured at a higher bitrate than the final destination, don't you? You aren't comparing DV to capturing as MPEG-2 with your CPU at the final resolution, are you?

[FredThompson]

OK, well, here's what I can tell you:

There is no comparison between the ADS and the ATI. The ADS uses a pro-level dedicated MPEG2 encoder and does up to 15 Mb/s throughput for video and up to 448 kb/s for audio, locked, and external to the box. The ATI has no native sound support. By way of comparison, that's about 1.5x the max "legal" datarate of DVDs.

The ADS also have some built-in filters tha can be real helpful including a simple temporal and a spatial filter.

I can't find the footage of the white guy, dressed in white, walking up a sand dune nor did I see any good sun-on-water shots on the travel channels tonight but I did get 2 short segments I think are worth comparing. One is the title shot to a thing about Mt. Everest, lots swirling white dots on a black background. The other is Steve Irwin in some water with reflected sun.

Here's the thing, I can easily setup the ADS to blow away the ATI, no question about that. Having thought about it, though, there's no way to really judge what's right for you without understanding more about the computer you're using. I'm going to find a good example frame or two and will pull them from DV, the ATI and the ADS. Don't expect color swim like you'll see with composite, this is going to be DV played from the camcorder and the original source was DirecTV.

I can find nothing in the book about the oversampling. This unit DOES capture at 2x oversampling which means there are two "samples" per pixel. The ATI probably does 1. If you visualize a pixel as being a sine wave and it's only sampled once, you'll see how it's not accurate compared to 2. 4 would be ideal but that's not an option. Basically, if you want to digitize an analog source, you have to sample it at a greater frequency. The ATI only samples at the same frequency. (This would be a lot easier with a picture. I'll scribble one up and put it with the stream grabs.) Manual is available at http://www.adstech.com and this stuff is around page 22.

There is also an ADS unit with DV in/out so you can go from DV to MPEG-2. Uh...I don't see the benefit to that. All the DV I shoot has to be edited anyhow and I want to save it on DVDR in the original format. Dunno what ADS was thinking.

[DigitalfreakNYC]

Fred,

Some very good advice there, thank you.

Here are my final thoughts. Is there a way of setting up the ADS to optimize it? I seem to remember seeing a post of yours for doing this. Would it be the same process for the 2.0 version? Also, I found 2.0 online for around $170. Good price?

Is there ANY benefit to going to AVI first? I just want to make sure I wouldn't be better off getting the Canopus.

My hesitation is with the creation of the MPEG2 immediately rather than going from AVI-MPEG2. The comforting thought is that "stand alone" machines do it "on the fly" when you hook them up to your TV and I'm impressed with what I've seen in that respect. But someone had mentioned that this capture card does it at a fixed bit rate and not a variable bit rate which didn't make sense to me because isn't that something that happens during the burning process?

You've almost got me hooked, Fred! Thanks!! :)

[FredThompson]

The 1.0 hardware has an upper standard bitrate of 3000 and was designed for USB1.1. That's why the "hack" to override the settings. No need with the 2.0 hardware.

Understand that even with a very high bitrate any MPEG-2 device is probably not going to give pixel-per-pixel accuracy. However, with a very high bitrate and the oversampling, you'll probably never see the difference. DV is also lossy so no big deal.

Personally, I think given your destination is MPEG-2 you're BETTER grabbing as MPEG-2 and doing your editing within that color model if at all possible. There is no 1:1 relationship with RGB, the most common method used for AVI. AVI is more universally used by PCs but there are plenty of ways to access MPEG-2. VirtualDubMod and the new AviSynth can both do it. Re-compressing within a color model is faster and more accurate than adding a translation step.

The ADS can do CBR or VBR. For quality, you'll grab at a real high rate and do whatever editing you want then re-compress to fit the target storage medium. If it's not critical, you COULD just grab to fit the target medium, that's your choice.

Again, AVI->MPEG usually changes the color information somewhat, MPEG->MPEG should not. It WILL if you do some editing with a non-MPEG-friendly method.

[DigitalfreakNYC]

Fred,

So if i get the ADS 2.0 it will be fine out of the box and I won't have to do any tweaking?

What do you mean by "recompressing with a color model?"

Does it hurt to recompress? How can I do that? How's the quality if i capture at VBR?

I know that when i transferred from AVI-MPEG via TMPEG, I loved that it color corrected and looked better in the resulting MPEG. I just want to make sure i'll be able to clean it up.

[FredThompson]

Originally posted by DigitalfreakNYC
Fred,

So if i get the ADS 2.0 it will be fine out of the box and I won't have to do any tweaking?I don't think you'll have to do any tweaking to capture. If I recall correctly you have a decent setup. Someone with a P2-400 and ATA33 drives wouldn't be happy. This thing can firehouse data and you need to be able to handle it.What do you mean by "recompressing with a color model?"There are different ways of numerically representing the image data. There is not necessarily a 1:1 relationship between them. If, in the process of editing, you convert the source between different formats, the result will be the least common denominator of color range. To illustrate, it is impossible to print on paper a fully accurate representation of parts of the visible spectrum. Forgetting for the moment the difference between direct and reflected light, you cannot, for example, print a full representation of some types of stained glass or a rainbow.Does it hurt to recompress?All editing and compression (except, sometimes, removal of frames) changes the data somewhat. The more you do this, the lower the quality. That is why you should capture at as high a resolution as is feasible and edit then reduce to your destination format, same as with audio mastering. There is always a diminishing return. It would make little sense, for example, to store an AM radio transmission with 96bits but it sure makes sense if you're capturing a live performance.How can I do that?I use DVDV2SVCD and SVCDBuilder. SVCDBuilder looks like it will be the better solution for groups of capture as there's active development of a batch mode. I like the interface better, also. How's the quality if i capture at VBR?Dunno. You'll have to try it. I can't say which settings are optimal for your needs and taste.I know that when i transferred from AVI-MPEG via TMPEG, I loved that it color corrected and looked better in the resulting MPEG. I just want to make sure i'll be able to clean it up.Color correction is a different function than the color encoding method. You can adjust color many different ways. The 2.0 includes some settings you can adjust before you capture if you wish. Perhaps there is a good practice of setting white and black points. Haven't experimented with it enough to find one and don't want to say along those lines until I've got the process really refined. Like I've said, I think there will be a switch away from internal capture cards to external devices because of the quality/price and LOTS of people will be asking LOTS of questions. I'd rather respond with, "Read part X of the article."

[DigitalfreakNYC]

Hmmm...

well...looks like i'm going to have to bite and go for it. I guess I'll let you know how it turns out!

Wish me luck! I'll probably order it tonight.

[FredThompson]

I think you'll be pleasantly surprised.

wrt the goal of mass migration of analog to DVDR, I've been doing a LOT of tests with DV made from DirecTV and also VHS-C camcorder. I've got S-VHS also but haven't tested it as thoroughly yet.

What I've found is an hour of DirecTV easily fits in 1.17G at SVCD resolution using DVDPatcher, no problem. That's actually a waste because it's recompressing anyhow. I'm still playing with noise removal to get rid of the MPEG artifacts but I don't think it's really possible unless you're working with simple animation.

With analog source, 1/2 D1 (CVD) resolution is just fine and you can fit an hour in 0.78G

That means you can easily put 6 hours of converted analog on a DVDR and still have room for menus. The bitrate is so high that I can't see any benefit to VBR so the encoding is a LOT faster. My 2100 Athlon is doing it at greater than 1x speed.

That's why I was talking about running in batch mode. Let your computer do the final compression on 4-6 hours of source material while you sleep.

That comes back to the issue of storage. If you sample at 6000 bitrate CBR you're probably just fine and the inline noise filters could be handy. 6 hours as DV is almost 80G. It's a lot smaller with MPEG2, around 10-11G. You could even bump that up to 12000 bitrate but might not see an appreciable difference because the analog source isn't THAT clean.

Starting to see the benefit vs. a DV converter? It would take up even MORE space using a lossless codec like people typically use with internal capture cards.

[DigitalfreakNYC]

ummmm....


what the HELL did you just say?


;)

Everything i'm putting onto DVD-R is from laserdisc or SVHS so I want the quality to be the best it possibly can. I'll never put more than 2 hours on a DVD-R and SVCD quality just isn't good enough.

Other than that, I honestly have no idea what you just said.

[FredThompson]

There's a point at which having a larger file doesn't give you anything because you've surpassed the resolution of the source material.

S-VHS doesn't have the equivalent of 720x480. It's actually less then half of that.

I don't see any real difference between DirecTV re-encoded at 3000 and 6000. S-VHS has a lower resolution than DirecTV.

Same thing with VBR and a high bitrate. It takes a lot longer to produce and I can't see any difference.

The point is you should be able to put at least 6 hours of S-VHS source on a DVD-R. Yes, you could save 2 but you won't see any quality difference, just a higher storage cost.

When you're comfortable with the process of capturing, cleaning and encoding, you'll understand.

[DigitalfreakNYC]

Fred,

I guess my confusion comes from being used to "stand alone" machines that record DVD's such as the Pioneer and especially Panasonic. On those machines...no one really uses any modes over the SP (2 hour mode). I guess i'm confused as to why it would be different on a PC burner.

[FredThompson]

Oh, yeah, those are totally different beasts.

Remember when I was talking about AM Radio and why mastering it wouldn't need as much resource as mastering a concert hall recording?

That's similar to what I'm saying about the limit of the information on your source. An S-VHS stream, no matter how good it is, has an upper limit of complexity of the image which is LESS than, say, film or any of the digital video formats.

Given that, you'll get to a point where storing at a higher bitrate doesn't give you anything other than larger files and greater storage cost.

The standalone units are probably locked at a very high constant bitrate because they don't have the buffering/speed available to do proper inline encoding. I'm guessing here but it's a reasonable assumption a consumer DVDR device isn't as capable as a professional one. I also wouldn't be surprised if this was done to keep the cost relatively high to keep the studio legal departments away. You know how music CDRs cost so much more than regular CDRs but they're really th same structure?... Come to think of it, locking in a high bitrate means the components are cheaper. Oh, and they have to handle a digital camcorder as source. That type of input has at least twice the information per frame as your S-VHS.

Getting back to color modes, you'll probably find using VirtualDub and the filters is well worth it. I was just poking around Donald Graft's site again and saw an updated delogo filter. Hoping to test that soon, it claims to properly remove alpha-blended logos, the semi-transparent ones some stations use. Wish there was a way to peel away the animated History Channel ones ;)

[^^-+I4004+-^^]

>Personally, I think given your destination is MPEG-2 you're BETTER grabbing as MPEG-2 and doing your editing within that color model if at all possible. There is no 1:1 relationship with RGB, the most common method used for AVI. AVI is more universally used by PCs but there are plenty of ways to access MPEG-2. VirtualDubMod and the new AviSynth can both do it. Re-compressing within a color model is faster and more accurate than adding a translation step.


hmm,fred you get it wrong quite often ,don't you?

nearly everyone here uses .avi as a source video for later encodings( to format of our choice) because all know that mpeg2 capturings look very poor.......different color mode is the least of a problem (capturers usually use "YUY2" (NOT RGB as you suggested) which is "stronger" color format than mpeg2's YV12!)
the image is better because of pure DCT system (mjpeg) or even better lossless huffyuv compression!
mpeg2 is dct coupled with motion estimation/compensation so it MUST look worse......also it has much lower bitrate etc.
editing in mpeg2 is pain in the butt too..........
recompressing mpeg2 to mpeg2 is surely NOT a way to go too.......
color space conversion is a small price to pay when u see how much better .avi video is!
(so small you won't see it at all)

so the way is to capture with the better quality (that's .avi in any case) and then take time to encode it woth decent software encoder...
that way results are better then thtis reencoding joke.....

also:is there some particular reason why this got all out of hands?
i mean,did nyc captured his video?(huh,too long thread to search)

btw. fred,have you uploaded the samples?
(>I've finally gotten some space at Geocities and here's what I'm going to upload tonight: )
any jpeg's fo rus modem folks (remember?........the 56k modem?heheh)

/ivo

[FredThompson]

read the entire thread, comprehend and don't troll. You look monstrously stupid in light of the rest of the thread.

Almost everything you've said is irrelevant or shows a total lack of knowledge of the situation being discussed here.

VirtualDub and VirtualDubMod filters = RGB

VirtualDubMod and AviSynth load MPEG-2.

Cheap capture cards give cheap results, regardless of codec used. We're talking about inline, oversampled, hardware encoding at 9,000-15,000 rate with (hopefully) no editing and re-encoding to about 3,000. (Had you READ and COMPREHENDED you would know that.) Lossless capture codec with a capture card incapable of the entire range due to inferior electronics just means he would have a very large lossless single-sample version of an incorrect sample.

Color will always be the lowest common denominator of whatever models are used throughout the entire process. As his destination is MPEG-2 for DVDR (had you READ and COMPREHENDED the thread you would know that) any larger colorspace doesn't really yield him anything. At most, given how he's described his source, he'll lop off some leading or trailing junk and that can be done in a GOP-oriented manner but I'm also considering AviSynth ignoring the leading and trailing junk and framserving to the transcoder/encoder.

JPEG of captured frames to compare quality? No. (Didn't you just complain about quality now you want LOSSY images?!?!) Lossy compression would obscure the real character. I'm actually planning some short clips and haven't figured out how best to do it, maybe compositing side-by-side, dunno yet.

Gotta find good torture clips. As I said earlier in the thread you didn't read yet feel qualified to comment about, I've been using DV instead of my ATI capture card for quite a while because sub-$1000 internal capture cards return horrible quality. Maybe that's changed in the past year but the only decent internal options to accurately capture the stream have been from people like Canopus. The ONLY correct thing I see in your post was the bit about capturing as RGB. Big deal, who cares? It's irrelevant because he can't get a clean capture using any cheap internal card so the options are external MPEG2 or DV.

Gotta find good torture clips and it's going to take a while. The Steve Irwin ads didn't have any real good examples for water shots. There's a few seconds and it didn't really show the problems. I got the VHS of Ken Burns' Lewis and Clark yesterday and hope to capture it tomorrow. That might have some good water motion segments. Don't remember where the sand ones are. I went looking and couldn't find the tape I was remembering which showed that.

The other challenge is how to properly present this given the bandwidth restrictions and codec issues. Screen grabs are ok for what they are but won't show real quality. I'm not sure how to balance the monthly download cap with the huge file sizes of DV, HuffYUV, and MPEG-2 captures. Changing samples to fit 10-year-old telecommunications limitations is not going to happen. It would be a monstrous disservice to those people who wish to see the true nature of the hardware performance.

Oh, yeah, read, comprehend and don't troll. (Something about that in the general rules a few times, isn't there?)

[FredThompson]

DigitalfreakNYC,

You know, maybe 720x480 IS the best way to capture your S-VHS. Here's why:

The ADS 2.0 has 2x oversampling, right? That means it's taking 2 horizontal samples per pixel. S-VHS doesn't really have the equivalent of 720 horizontal pixels, it's more like half of that.

So...the returned 720x480 capture should yield four samples of each pixel from the S-VHS tape; 2 from the increased resolution and 2 per each of those for the oversampling.

Remember how I mentioned 96bits for audio? I don't know if you're done pro audio mastering but the sound guys you know can probably explain why that's important for accurately digitizing a curve.

If I'm right about this, and I need to talk with ADS again to verify, it would mean you'll be getting about four reads per analog pixel (for lack of a better term) and that should yield a very accurate read of the original signal and quite a bit of associated noise removal. Re-sizing to 1/2 D1 or 352 horizontal pixels should give you very good results becaue it would take (what I'm describing as) 4x oversampling into account.

[^^-+I4004+-^^]

fred,fred,fred....if this was some US based forum (like ars) i wouldn't hesitate to use stronger words on you,but as this is not such place i'll behave (although your lack or comprehension doesn't deserve it) and will try again (as i did with modem stuff) and hope for the best:

i believe i'm not troll or stupid (as you said) when i say that
mpeg2 capturing can easily beaten by .avi system (sam as nyc said some while ago,or better he asked himself what's better)
since then ,your voice is the only one in this thread,so i wonder if the guy will get a wrong impression on this complete thing....

direct mpeg2 capturing is for losers who think it's the beast it can be etc.
sure enough,it's easy to do,but can you capture moto race with your gear with satisfactory results,and how bout cutting those commercials on mpeg2 stream?

also,i don't have to read the complete thread to see where are you going with it.....

>We're talking about inline, oversampled, hardware encoding at 9,000-15,000 rate with (hopefully) no editing and re-encoding to about 3,000.

as i said:you capture moto race,and do that to it,then post few hi motion scenes over ,so all of us can get a lot of laughs...(will you?)

do you really think that mpeg2 capture can stack up to multi-pass mpeg2 encoding as made with tmpgenc or cce for which .avi was used as a source?

to lighten you up a bit,bt8x8 is oversampling too (initial waveform is sampled at very high frequency)...but i have bt datasheet handy,so i can give real numbers,and not just assumptions as you do:


"The Bt879 requires an 8*Fsc (28.63636 MHz for NTSC and 35.46895 MHz for PAL) reference time source. The 8*Fsc clock signal, or CLKx2, is divided down to CLKx1 internally (14.31818 MHz for NTSC and 17.73 MHz for PAL). CLKx2 and CLKx1 are internal signals and are not made available to the system.
UltraLock™ operates at CLKx1 although the input waveform is sampled at
CLKx2 then low pass filtered and decimated to CLKx1 sample rate."
(from datasheet)

as you see,initial sampling frequency is 35.4MHz for PAL....at was raster is ADS scanning the video?

but as i said today in another thread,you won't really see the difference between the ADC's but different between the codec's...
mpeg2's temporal compression is making all the difference although mpeg2 in Iframe mode only is still inferior to mjpeg OR huff!

> Lossless capture codec with a capture card incapable of the entire range due to inferior electronics just means he would have a very large lossless single-sample version of an incorrect sample.

say what?
what have you just said here?hmmmmm

>any larger colorspace doesn't really yield him anything.

i would rather start with PROPER color instead of YV12 color....
also,you have stated that it would be re-encoded...you think that's wise?how well does your mpeg2 card handle interlaced stuff anyhow?
but concentrating on color space is missing the point:mpeg2 makes for lousy capture because mpeg2 is AN END format...not an intermediate one...

i can understand your point in one way only:for people who want to make it simple...capture to mpeg2,burn dvd,you're done....but you're talking re-encoding etc.


>It's irrelevant because he can't get a clean capture using any cheap internal card so the options are external MPEG2 or DV.

what cards/codecs have you tried before this?
"clean capture"?
in my mind it should be closer to the source quality and sharpness,but i can see that you love "clean" mpeg2 or DV codecs that smooth too much anyhow....guy on the forum just said he prefers bt device to DV capturing....i believe he knows what's he talking about...

>Gotta find good torture clips and it's going to take a while

the faster the motion the harder it gets for mpeg2....


>Oh, yeah, read, comprehend and don't troll.

i'll eat every single of my words if you prove that mpeg2 capture look better than DV or huffyuv...wether it be single frame or video sequence......i really will!
capture same source with different devices,post short samples and you're off......

you're talking like mpeg2 is holy grail of capturing etc.
very few folks here appreciate mpeg2 as a capturing system...prove all of us wrong,and you'll deserve my respect......

where are your samples?where's your proof? till date i have seen MANY lousy mpeg2 captures,never a decent one.......

>So...the returned 720x480 capture should yield four samples of each pixel from the S-VHS tape; 2 from the increased resolution and 2 per each of those for the oversampling.

what?you lack the basics of A/D conversion......"4x" oversampling because of the anticipated lower res. of the source?

also,svhs tape doesn't have "pixels"..and signal can only deteriorate after the ad conversion (any a/d conversion),and if cheap bt device is oversampling too (as i established here),what's <our point with all the sampling stuff?
codec flaws come to the sight,not ad conversion flaws.....(ie. 90% of artifacts are caused by the codec...don't worry about the conversion but about intermediat ecodec)

you're talking about cheap mpeg2 capture system as it cannot be beaten with anything else.........present the evidence or quit,it's not true......you're not helping the guy if the guy will ever snug a peek at some children's mpeg2 encodings made with cheap cards and avi as a source....

why am i talking to a person who said 4-6hrs will fit one dvd-r?

i'm off from here,as i have only repeated the stuff that the FAQ and the Guide already have........

and you,digitalfreaknyc buy yourself few devices and tell us what you saw....i believe fred's suggestions won't help you much anyhow....
mpeg2 is not even close to production system anyhow....(perhaps one day..)

[FredThompson]

You didn't read or comprehend the parameters or discussion, continue to attack and attribute to me things I did not state, and have made it very clear you have no intention of reading the thread. I will not respond to any more of your attacks or demands. You can, certainly, post to the thread but I will not respond to you.

[SeeMoreDigital]

Eee Gods! I hope thats over and done with!

Anyway, I'm all in favour of performing high bit captures direct to Mpeg2.

Mpeg2 at 6000kbps (or even Mpeg1 at 6000kbps) is more than good enough. After all you don't hear many complaints about the image quality of an Mpeg2 DVD!

And I would rather end up with a file at around 45MB per minute, as opposed to 225MB per minute for a DV.avi. Or even more for an uncompressed .avi

Sure a .avi looks good and who knows in a couple of years time when we all have 500GB hard drives that spin at 40,000RPM with a built in 500MB of cache, I might change my mind.

But as Fred pointed out, that was not what this post was all about.

I would like to see some decent 2pass VBR Mpeg1 and Mpeg2 encoding software though!

[FredThompson]

DigitalfreakNYC,

Any updates? Have you got your new hardware?

When you get yours, I'll be interested to know the performance. I had to determine different capture settings as the default generated a dark image on mine. I've also seen some funny interaction between the TBC and digital filters on the S-VHS deck and some worn tapes. This isn't anything new, happened with the ATI card as well.

I've also run into a problem with my IDE drivers. For some reason, the optimized ones aren't loading so I'm stuck with the default Windows ones until I figure it out. Unfortunately, this limits the speed at which I can capture. I've asked Asus for help but no reply yet.

I HAVE decided the internal filters aren't that great. Didn't think they would be at high resolutions and speeds but it was worth a shot.

[FredThompson]

Well, are you still around?

I finally have a good test clip. The opening credits for WWE Smackdown! last week had a huge amount of intentional over saturation and rapidly-changing mix of natural and CGI video. My DirecTiVo is busted so I couldn't get the original MPEG feed but I did manage to capture it onto DV. I've put together a sample of about 4 minutes length that's going to be a very good stress test. I have an XSVCD of it ready but I'll not post the URL until you get back to me. Want ot make sure you get it then I'll open up the site. There's a monthly download cap and I'd like to make sure you get the made-from-DV version.

I'm still playing with the settings on the ADS until I get something that looks great. Have been trying to compare with reference color bars. When I'm happy with the settings, I'll let you know what I've chosen and why. Tried the THX Optimizer and I don't like it. It's too limited a configuration test.