Capturing VHS to PC with a high definition resolution...?

[edDV]

I would agree the loss from 4:2:2 to 4:2:0 is minimal, to human eyes. The first two numbers are the important ones. This is why I (and many others) take issue with 4:1:1. It was a stupid mistake. If PAL can be 4:2:0 then there is no really compelling reason for NTSC to not also be 4:2:0, aside from greed of compression space.

I generally disagree and welcome the debate.

First the facts then the pro/con (later this evening).

1. 4:1:1 and 4:2:0 take the same amount of data space. There is no "greed of compression space".

2. 4:2:0 has color pixels spatially interpolated and is intended to be that last step in the process (e.g. DVD final encode).

3. 4:1:1 retains chroma pixel spatial relationship to luminance (co-sited with first luminance pixel) and undergoes no interpolation through multi-generation. 4:1:1 has near zero multi-generation loss in those areas of the frame that are not processed (e.g. backgrounds). 4:2:0 needs to have chroma pixels reinterpolated for each pass and multi-genertion chroma losses are great. For this reason, NTSC and PAL DVCPro formats use 4:1:1 sampling. They are considered intermediate formats, not end display formats.

4: 4:2:0 was developed theoretically to make large interlace displays look better especially for PAL and interlaced HDTV. There is no problem when it is used for that purpose, but there have been a host of problems surface when 4:2:0 is used in production.

5. Europe (and other PAL areas) are having severe problems mixing 4:1:1 and 4:2:0 sources in production and the EBU is considering a standard that forces all multi-format DV editing to be upconverted 4:2:2 to solve the problems.

6. In NTSC areas we have been free of these issues until the introduction of standalone DVD recorders, DVD camcorders and ATSC DTV tuners. Now we have to contend with 4:2:0 inputs for consumer material. Fortunately all pro formats avoid 4:2:0.

7. Current format lineup:

NTSC DV, DVCAM, DVCPro are 4:1:1
NTSC DVCPro50, Digital Betacam, (other D1 based) are 4:2:2
NTSC DVD, DVD camcorders, DVD recorders, ATSC SD-HDTV are 4:2:0 (all intended to be final display formats)

PAL DV, DVCAM are 4:2:0
PAL DVCPro is 4:1:1
PAL DVCPro50, Digital Betacam, (other D1 based) are 4:2:2
PAL DVD, DVD camcorders, DVD recorders are 4:2:0

The above are my understandings. Lets discuss the accuracy so far before going subjective.

[lordsmurf]

Yes, 4:2:0 is an end product. And when converting 4:2:2 sources, best leave it in a 4:2:2 format, not quash it into 4:1:1 for editing, then to 4:2:0 for a DVD. Less conversions are preferable.

I do not understand how 4:1:1 is a "intermediate" format, it's application is as a cheap DV shooting format. It's a source format.

We do best when sticking to 4:2:2 for work, and 4:2:0 for final production.

I won't disagree that 4:1:1 has less multi-gen loss, but then first gen already stores the loss. The damage is done on the first gen, so what if later gens are all the same? Lesser of evils discussion?

And yes, creation is FAR FAR FAR more fun than duplication.

[edDV]

Yes, 4:2:0 is an end product. And when converting 4:2:2 sources, best leave it in a 4:2:2 format, not quash it into 4:1:1 for editing, then to 4:2:0 for a DVD. Less conversions are preferable.

I do not understand how 4:1:1 is a "intermediate" format, it's application is as a cheap DV shooting format. It's a source format.

We do best when sticking to 4:2:2 for work, and 4:2:0 for final production.

...

OK I generally agree with that. DV is a source format but 4:1:1 DV can also be used as a "poor man's" intermediate format, without loss under certain cases.

Once transferred to DV-AVI we have our source files on HDD. When these files are loaded into Vegas or Premiere Pro "DV project format" they are actually "losslessly" converted to 4:2:2 YUV or 4:4:4 RGB behind the scenes when previewed. If all we do is cut clips together, we can "losslessly" output to 4:1:1 DV tape. We can also repeat this process N times without loss. N is a big numerber.

If we do the same with a graphic overlay, then the overlay portion of the picture is processed in internal YUV or RGB but the background is unprocessed. In this case the background pixels would be recorded with no loss to 4:1:1 but 4:2:0 recording would probably take a generation hit as the chroma pixels are reinterpolated.

It all depends on how the editing programs work internally.

added: DVCAM and DVCPro are considered intermediate formats or even end release formats in news and some production. Much of the time editing is done in DV format and the final air tape is DV. Other times the edited DV tape is loaded onto a 4:2:2 playout server.

[vhelp]

To continue the flavor of the original question ...

In this ongoing effort to determine once and for all, what is the best
format / process for capturing VHS (or any source) with the purpose
of converting to a high defination resolution (or closest) (or alternative)
(or alternative method) ?

.
.

We all agree that the best method of obtaining near-original quality
is by capturing in a lossless format.. ie, 422 format.

At the moment, this is hard to determine, because I have not seen
anything in writing that specifically says that such a given
capture device is capturing internally in one of two most popular
format.. RGB vs. YUV.

Its my understanding that sources such as:

** cable tv; satelltie tv; laserdisc; dvd players; vcr;

..all output in YUV 420 format.. because that is the end product, or
as has been termed here, final production

But, when a given capture device is put in between, then there is the
phonomina/mystery of the capture machinics, which (in my limited
understandings) will sample the pixels (bits or whatever at the time of
reaching the capture card) and organizing the data into a fashion to
be formated.. that remains to be determined weather or not the final
output is YUV or RGB [without codec involvement]. That is the question.

Then, a sub-question would follow.

** If not RGB (which would ultimately be 444) then, if YUV is the final
** output production to the capture device (see Fig A. below) then what
** is the actual format of the sample when captured data is finalized to
** the capture device, before transfering to a container (avi file)
** without any codec/compression/sampling envolvement..
**
** 444 or 422 ??

Possible combination of setups, BEFORE codec process:

input_source[YUV 420] --> capture device --> output_source[RGB 444]
input_source[YUV 420] --> capture device --> output_source[YUV 444]
input_source[YUV 420] --> capture device --> output_source[YUV 422]

----- ----- ----- ----- ----- Fig A. ----- ----- ----- ----- -----

Where input_source could be a VCR -> capture device -> stream (before avi file)

However, what is important is the users current equipment on hand.

If the user's capture tool is 422 capable then IMO there is no limiting
factor here, and from an archival standpoint, this is the first
stage or link in the lossless capture process. This is Person A.

** 422 capture tool = DV 422 devices.. ie, dv external boxes; capture cards
** in conjunction with codecs supporting 422 format.

But, if the user's capture tool is 411 then that is the limiting factor,
and from an archival standpoint, this is the first stage or link in the
lossless capture process because this is the best tool available to the
the user at the current time of process.
.
Note, this setup does not mean that it is lossless, but that given the
limitedness of the user's current tools, that this is to be considered
the closest to being lossless at time of process for this person's current
limitedness.. though we all know in reality that it is not lossless. This
is Person B.
.
Many of you and I's are sometimes in this last group, depending upon
our current priorities and/or process endeavors.

** 411 capture tool = DV devices.. ie, dv cams; dv external boxes; etc.

I for one, would like to find out what is coming out of a given
capture device before any codec or compression finalizies to an avi
file. Is it RGB or YUV ??

last..

Cautious contridiction warning..

Please bare in mind, that what is considered lossless is in the eye
of the beholder and his/her current setup/equipment/tools and comprimise
of final product. And, what is considered lossless today, could be
determined lossy tomorrow, and back n forth, and so on and so forth

Thus, if a user chooses to use VHS as the argument of quality, and finalizes
with either 422 or 411 or 420 or whatever to be closest to lossless, then,
at that user's current time of comprimise; skills level; process; current
choice; is to be considered as close to lossless as determined then.


-vhelp 3542

[FulciLives]

Let me try to ask something easier and more simple than vhelp's last post. Let's say you had the following 2 choices:

1.) Capture NTSC using DV (such as a Canopus ADVC-100)
2.) Capture NTSC using hardware MPEG-2 at 15,000kbps CBR (such as Hauppauge WinTV PVR 250)

Which would be better quality wise?

I'm afraid that for myself I am down to using one of those two methods. I just can't seem to get a PCI card TV tuner card (on my new computer) to work and keep audio sync. I love the picture quality of capturing YUY2 with HuffyUV or even PICVideo MJPEG but it's worthless without audio sync.

Currently I just bought (as of today) the ADS INSTANT DVD 2.0 USB device (hardware MPEG-2 with audio sync lock function) and I've done (so far) 2 captures at 15,000kbps MPEG-2 CBR and the quality is very nice but looks slightly "soft" to me (although currently it isn't doing "I" frame only encoding ... need to hack it ... need vhelp for this my friend if you could msg me).

My other option is a Canopus ADVC-110 (since the 100 is hard to come by) or the DataVideo DAC-100. Have yet to try this solution. Unfortunately don't know anyone with one I can borrow so if I want to "try" first I have to buy first.

At least I bought the ADS INSTANT DVD 2.0 at Circuit City ... 30 day 100% return policy if I don't like it.

So far though I am fairly happy (knock on wood). I don't know if the "softness" is due to the source (digital cable channel) or not. Maybe I should try to capture from a clean high quality DVD source as a "test" since some of my digital cable TV channels ain't so hot looking to begin with. Also I need to figure out how to do "I" frame only encoding. vhelp said he hacked his somewhat very similar ADS DVD Xpress to do "I" frame only encoding. Hope I can do the same with my ADS INSTANT DVD 2.0

- John "FulciLives" Coleman

[edDV]

....
...
Its my understanding that sources such as:

** cable tv; satelltie tv; laserdisc; dvd players; vcr;

..all output in YUV 420 format.. because that is the end product, or
as has been termed here, final production

-vhelp 3542

The disconnect here is all of those devices output in analog. Once D/A has occcured these outputs revert to analog specs. Typical analog "formats" or standards follow these descriptions.

Composite video standards (studio spec and tuner decode spec)
-----------------------
NTSC
PAL
SECAM

Y/C video standards (luminance separated from modulated chroma subcarrier)
------------------------
NTSC
PAL

Component video standards
------------------------------
YUV, aka Y, Pb, Pr
RGB, RGBHV
YUVA (adds alpha)
RGBA (adds alpha)

It is technicaly incorrect to apply terms like 4:2:2 to analog, but many use these terms to indicate the relative "bandwidth" of the analog components.

The capture device resamples the analog components into digital component YUV. Sampling theory (Nyquist Theory) suggests that sampling frequencies that double the analog bandwidth are adequate to describe that signal and that oversampling is of no benefit. Going by that rule, 4:1:1 is more than adequate for sampling decoded NTSC, PAL or SECAM and also more than adequate for any standard definition component YUV output found in a consumer environment.

As an example, 4:1:1 is even adequate for Betacam SP. Betacam SP has a luminance bandwidth of 4.5MHz and 1.5 MHz for U and V. 4:1:1 can handle the following

----Sampling Freq.-----Max analog bandwidth-----
Y____13.5 MHz_________6.75MHz______
U_____3.4 MHz__________1.7MHz______
V_____3.4 MHz__________1.7MHz______

So in sampling terms, 4:1:1 is adequate bandwidth for any home analog source other than HDTV.

A/D sampling is identical for 4:1:1 and 4:2:0. The only difference is pixel arrangements of chroma U and V within the 720x480/576 raster.

[lordsmurf]

So in sampling terms, 4:1:1 is adequate bandwidth to for any home analog source other than HDTV.

In theory .... that does not always carry out to be practical application.

[SatStorm]

@FulciLives: I tested many hardware mpeg 2 capture cards and all I can say, is that I don't like the picture quality they all offer!
Also, call me old school, but USB devices for capturing, is a big NO for me. So the same for USB DVB Cards.
Use a PCI based solution when you can. You never loose!

[edDV]

Let me try to ask something easier and more simple than vhelp's last post. Let's say you had the following 2 choices:

1.) Capture NTSC using DV (such as a Canopus ADVC-100)
2.) Capture NTSC using hardware MPEG-2 at 15,000kbps CBR (such as Hauppauge WinTV PVR 250)

Which would be better quality wise?

Very good question. The two techniques are similar, that is realtime CBR capture in a first pass, then edit, then encode to DVD MPeg2. The question should also be extended to how the two techniques differ for high quality input (say a good cable box 480i S-Video out of a HDTV channel) and a VHS input.

I'll dump all I know about the pro and con of each.

DV capture process

1. Composite video is separated to Y/C or Y/C external input is used
2. Y/C is decoded to YUV. Some models levels correct for NTSC setup at this point (e.g. Canopus ADVC). Some models add proc amp/TBC modes at the this point (e.g. ADVC-300).
3. YUV passes through prefilter and A/D and emerges as Y, Cb, Cr (4:1:1 for NTSC and 4:2:0 for PAL)
4. Compression: DV uses 5x intraframe only at 25Mb/s CBR. Codec is same as used on expensive broadcast models.
5. Intraframe only compression allows for high quality editing, filtering and effects.
6. DV format is highly supported for IEEE-1394 transfer by DirectShow and Apple OS.
7. Quarter chroma sampling, although adequate for bandwidth and display, may have problems with extreme effects processing vs. 4:2:2.
8. Software DVD MPeg2 encoding follows editing. Hardware MPeg2 codecs could be added.

MPeg2 hardware codec capture process

1. Composite video is separated to Y/C or Y/C external input is used
2. Y/C is decoded to YUV. Some models may levels correct for NTSC setup at this point. Some models may add proc amp/TBC modes at the this point.
3. YUV passes through prefilter and A/D and emerges as Y, Cb, Cr sampled 4:2:2. Signal is passed internally to MPeg2 codec.
4. Compression: DV uses combination of intraframe and interframe compression at 4-15Mb/s CBR. Codec is usually low cost consumer Conexant. http://www.conexant.com/products/entry.jsp?id=9
5. Interframe compression makes editing and effects more difficult and less compatible.
6. External hardware MPeg2 encoders use proprietary USB2 drivers and application software.
7. Using first pass compression above DVD 9Mb/s standard requires a second pass recompression for DVD spec MPeg2.

Plaease comment and add more.

[FulciLives]

@FulciLives: I tested many hardware mpeg 2 capture cards and all I can say, is that I don't like the picture quality they all offer!
Also, call me old school, but USB devices for capturing, is a big NO for me. So the same for USB DVB Cards.
Use a PCI based solution when you can. You never loose!

The reason I picked USB 2 was so that I wouldn't have to deal with using an audio card. I had been using a cheap PCI TV tuner type card and thought I had finally figured out the whole audio sync problem but now with a new computer (but same capture card) I just can't get decent audio sync even with VirtualVCR dynamic audio sampling.

So my choices where (in my mind) either the Hauppague WinTV PRV-250 or the Hauppauge WinTV PVR USB2 (which I've heard works just as well) but I've heard such good things about the ADS Instant DVD 2.0 (a USB 2.0 device) that I couldn't pass it up.

So far the image quality is most impressive (it has 9 bit video digitizer with 2x oversampling and 4 line comb filter) just that it looked a bit soft on one of my two captures. More testing is needed.

- John "FulciLives" Coleman

[edDV]

So in sampling terms, 4:1:1 is adequate bandwidth to for any home analog source other than HDTV.

In theory .... that does not always carry out to be practical application.

That may well be. I think 4:1:1 is fine for quality sources. I'm trying to understand the various issues for low quality VHS.

[FulciLives]

Well here is what I intend to do with my hardware MPEG-2 USB 2.0 device:

1.) Capture at 15,000kbps CBR MPEG-2 with 16-bit 48k 384kbps MP2 audio.

2.) DGIndex to demux audio and get D2V project file

3.) Convert the 16-bit 48k 384kbps MP2 audio to 16-bit 48k PCM WAV audio

4.) Load D2V and WAV audio into an AviSynth script.

5.) Load AviSynth script into VirtualDubMod to edit

6.) Import EDITS into AviSynth script

7.) Extract edited WAV audio ... normalize ... convert to 256kbps AC-3

7.) Open AVS script into CCE and do a multi-pass VBR to make it standard MPEG-2 DVD compliant.

8.) Author MPV (video) and AC-3 (audio) and burn to a DVD-R

Now if I had a Canopus ADVC-110 I would be doing basically the same thing except I could just use the AviSource command to load it into AviSynth and not have to do the DGIndex thing ... plus the sound is already 16-bit 48k PCM WAV.

The ADS INSTANT DVD 2.0 can capture 16-bit 48k PCM WAV but then there is a possible chance for audio sync error. It only works with "locked audio" when you capture to a muxed MPEG-2 file which means using MP2 audio.

According to some things that I have read and picture and video samples ... it seems to me that a 15,000kbps CBR MPEG-2 capture is a "better quality" capture than a DV capture (i.e., capture from an analog source using something like the Canopus ADVC-110). This has been based on the concept that NTSC DV 4:1:1 is "bad" and since this thread has come down to the 4:1:1 issue ... that's why I decided to bring it up.

One last thing that I am sort of "confused" about. A DV capture is YUY2 but a MPEG-2 capture is YV12 ... correct?

So in my AviSynth script I have to use "ConvertToYUY2(interlaced=true)" so I can use the YUY2 version of Convolution3D (noise filter) plus CCE works best with YUY2 input.

I wonder what "hit" in quality there might be when doing the YV12 to YUY2 conversion?

- John "FulciLives" Coleman

[TooLFooL]

Sort of like asking "can I drop a V10 Viper engine into a Dodge Neon"?

Damn, that would be AWESOME! I wonder if it would fit?!

[Wilbert]

@edDV,

Are you saying there's no visible difference between the following?

analog -> 411 -> 444 (upon display)
analog -> 422 -> 444 (upon display)

[edDV]

@edDV,

Are you saying there's no visible difference between the following?

analog -> 411 -> 444 (upon display)
analog -> 422 -> 444 (upon display)

My #1 point was that in a home environment we have no SDTV sources that will benefit from 4:2:2 sampling other than stills and computer generated animation. Oversampling inputs like VHS does not improve quality or a least nobody has offered a case that it does.

A #2 point is...

According to all the past standards research, 411 -> D/A to analog display shows no difference vs. 422. The TV would do the Y, Pb, Pr to RGB conversion.

What I can't tell you is how big that monitor was and if the research would get the same result when using an upconverted 1920x1080i display.

Here is an off cable sample of the ADVC-100 480i 4:1:1. I don't see many obvious chroma sample issues.



At 200% 4:1:1 is holding well


At 400% the Y sampling is more visible than wide color pixels.

[FulciLives]

I just figured out how to get my ADS Instant DVD 2.0 (external USB 2.0 hardware MPEG encoding device) to capture at 15,000kbps CBR with "I" Frame only encoding.

Now I need to find a digital cable channel that is CLEAR (a trick for Comcast Cable) with a cooking show so I can try to capture my own nice close-up of food for comparison!

So far I am most impressed with this capture device. Wish I had a DataVideo DAC-100 or Canopus ADVC-100/110 to compare it to though.

- John "FulciLives" Coleman

[edDV]

I just figured out how to get my ADS Instant DVD 2.0 (external USB 2.0 hardware MPEG encoding device) to capture at 15,000kbps CBR with "I" Frame only encoding.

Now I need to find a digital cable channel that is CLEAR (a trick for Comcast Cable) with a cooking show so I can try to capture my own nice close-up of food for comparison!

So far I am most impressed with this capture device. Wish I had a DataVideo DAC-100 or Canopus ADVC-100/110 to compare it to though.

- John "FulciLives" Coleman

Cool on the I frames only. That rocks!
It would be great to do side by side comparisons or work off the same file. Hard to share a calibrated S-Video analog source. Cable systems and cable boxes vary considerably.

[davideck]

Hey FulciLives -

Do you have access to a Mini-DV Recorder/Camcorder with Analog In?
A great comparison would be between a Mini-DV Dub and a DVD Dub.

[vhelp]

Evening guys.. it's Friday hehe

See fulci, I knew you didn't need the .reg tweek, just needed to
do it in capwiz or whatever sw came w/ your box.

-vhelp 3544

[vhelp]

@ edDV

Hay, I read your responses above. Great. If I have some things
to add/comment, I'll post. Right now, I'm just setteling in after
a hard days worth. You know how it goes.

-vhelp 3545

[FulciLives]

Cool on the I frames. That rocks!

I had to do a registry hack to get the "I" Frame only encoding. Did that by the seat of my pants. In other words had no clue what I was doing but with enough tinkering I figured it out LOL

It would be great to do side by side comparisons or work off the same file. Hard to share a calibrated S-Video analog source. Cable systems and cable boxes vary considerably.

Hell my cable box doesn't even have a S-Video output. There is a "hole" with a metal plate covering it that says S-Video output LOL ... so I am using composite outputs. It's a Motorola box.

Do you have access to a Mini-DV Recorder/Camcorder with Analog In?
A great comparison would be between a Mini-DV Dub and a DVD Dub.

When I was at Circuit City (where I bought the ADS Instant DVD 2.0) they had a Canon ZR-100 DV cam for (after rebate) something like $299 or $309 or something like that ... it has analog in and I've heard good things about the analog in on Canon camcorders ... I thought about it ... but I do PAL VHS stuff sometimes so if I go the DV route I really need something like the Canopus ADVC-100/110 or the DataVideo DAC-100

I only have an old Sony Hi8 cam and not a one of my group of friends have a digital cam of any sort (one has VHS-C and another has a Sony Video 8 not even Hi8).

I suppose I could buy the Canon only to test the analog DV input to computer method then return it to get my money back but what a hassle plus that is not 100% ethical LOL

- John "FulciLives" Coleman

P.S.
I just did a 121 minute capture (15,000kbps CBR "I" frame only with 384kbps audio) and the MPEG file was only 13.7GB which is about what DV is for 60 minutes ... of course DV does 1536kbps PCM WAV audio.

@vhelp
I took so long to type this up ... then had to attend to my pizza delivery (mmm good) that I missed your two posts just now LOL

[vhelp]

@ fulci

Well, I finished my (small) pizza pie too - am stuffed.. topping it
off with coffee. It's my winedown. Anyways.


@ edDV

Testing equipment theories and syncing up testings ...

I think we can sync up together, our theories and tests with a DVD
as a reference. I know just the movie to test. "The Incredibles".
TONS of RED 's, and BLUE 's to test various capture equipments, such
as DV devices; hardware mpeg boxes; and Analog Capture cards.

Also, my laserdisc player is Pioneer CLD-V2600 (composite/s-video)

** When I tested my LD, (again) I tested w/ RCA and S-Video, and I
have concluded that RCA was the best connection to produce the closest
pixel information. It would seem that w/ LD players, the composite
signal *is* recorded onto the source (though separated Y/C)
The theory here, IMO, is that the RCA connections brings the detail
closest to the original source vs. s-video (using my ADVC-100 box)
I haven't tested my ADS Xpress box yet. That will be my next test.
Its hard to test because laserdisc content is flooded w/ dot-crawl,
something we blaimed on DV and other nonsense in the past. But, w/
RCA connections, (and because of the way the signal is recorded onto
the laserdisc [composite]) the detail is cleaner, less noisey than
an s-video.. again, using my ADVC-100 box and RCA connections.

How about starting an new thread topic on testing scenarios, so's
not to contaminate this one ??

-vhelp 3546

[edDV]

@ edDV

Testing equipment theories and syncing up testings ...

I think we can sync up together, our theories and tests with a DVD
as a reference. I know just the movie to test. "The Incredibles".
TONS of RED 's, and BLUE 's to test various capture equipments, such
as DV devices; hardware mpeg boxes; and Analog Capture cards.

I can rent that for $1 on Mondays and Wednesdays at the grocery

Also, my laserdisc player is Pioneer CLD-V2600 (composite/s-video)

** When I tested my LD, (again) I tested w/ RCA and S-Video, and I
have concluded that RCA was the best connection to produce the closest
pixel information. It would seem that w/ LD players, the composite
signal *is* recorded onto the source (though separated Y/C)
The theory here, IMO, is that the RCA connections brings the detail
closest to the original source vs. s-video (using my ADVC-100 box)
I haven't tested my ADS Xpress box yet.

For Laserdisc, composite NTSC (or PAL) is directly recorded to the disc as a composite analog signal. It is very similar to the way 1" Type C VTR's record and nothing like VHS. As such, composite NTSC is the native format unlike Y/C for VHS and S-VHS.

The S-Video output on a Laserdisc player is created by an internal Y/C separator. Those were poor on that vintage player so composite output is the one to use.The Y/C separator in the ADVC or ADS Xpress is probably superior. A 3D comb filter would be even better.

That will be my next test.
Its hard to test because laserdisc content is flooded w/ dot-crawl,
something we blaimed on DV and other nonsense in the past. But, w/
RCA connections, (and because of the way the signal is recorded onto
the laserdisc [composite]) the detail is cleaner, less noisey than
an s-video.. again, using my ADVC-100 box and RCA connections.

Elimination of dot crawl and cross color while maintaining high frequency response is the goal of the Y/C separator (see link below).
http://members.aol.com/ajaynejr/vidcomb.htm

How about starting an new thread topic on testing scenarios, so's
not to contaminate this one ??

-vhelp 3546

Sure, I'll give it some thought.

[slacker]

A related question (I THINK!)...

If it is correct to say that 15mb I-frame ONLY VHS transfers through a PC capture card are comparable to the roughly 25mb DV format used by a digital camcorder, then from a quality and practical standpoint could you just transfer your VHS using the analog/digital passthru feature of a DV camcorder such as the SONY HC90 directly to your PC using firewire rather than mess with the capture card? Would this be a better archive format than the 15mb I-frame ONLY MPEG2 format cited above?

[MOVIEGEEK]

If anything, Blu-Ray has LOTS of support from over 50-100 companies (I don't know exactly how much, cos they seem to be getting new members every day as it seems), and HD-DVD doesn't even as many backers; if anything, it's likely going to be in trouble.

That, and a recent consumer survey found more than 50% of people would support Blu-Ray over HD-DVD.

If you want more info on who supports it (assuming you haven't checked so), go to http://www.blu-ray.com

More than 50% of people prefer a car that gets more than 50mpg(21kpl) but that doesn't mean we will get one soon.Nobody knows which format will win but they better be cheap with lots of titles or HD will go the way of DVD-Audio/SACD.

On Topic:I would capture in DV-AVI via firewire.

[The_Doman]

If it is correct to say that 15mb I-frame ONLY VHS transfers through a PC capture card are comparable to the roughly 25mb DV format used by a digital camcorder, then from a quality and practical standpoint could you just transfer your VHS using the analog/digital passthru feature of a DV camcorder such as the SONY HC90 directly to your PC using firewire rather than mess with the capture card?

If you have that option available you really should try that out.
I prefer to use my SONYD8 camcorder (PAL) with passthough for that when capturing VHS instead of my other capture cards. (PVR150,Bt848)
The real good thing is that the passthrough kind of acts like a TBC in that the analog to digital converter (or passthrough feature) stabilizes the video.
I REALLY can see the difference compared with my "normal" tv/capture cards.
Ofcourse a nice stabilized capture really helps when converting to mpeg (or other formats).

[jzmax]

I wonder if something like this would help improve the video quality of VHS captures in HD:

http://cgi.ebay.com/HDTV-Video-Enhancer-Converter-DVD-TV-Plasma-LCD-Monitor_W0QQitemZ5...QQcmdZViewItem

[edDV]

A related question (I THINK!)...

If it is correct to say that 15mb I-frame ONLY VHS transfers through a PC capture card are comparable to the roughly 25mb DV format used by a digital camcorder, then from a quality and practical standpoint could you just transfer your VHS using the analog/digital passthru feature of a DV camcorder such as the SONY HC90 directly to your PC using firewire rather than mess with the capture card? Would this be a better archive format than the 15mb I-frame ONLY MPEG2 format cited above?

DV at 25 Mb/s compresses 5x at 4:1:1 (or 4:2:0). 15Mb/s 4:2:2 compresses ~12x (uncompressed 4:2:2 is ~180Mb/s).

I frame only MPeg2 compression at higher rates would be similar to DV. DV optimizes for luminance quality and trades off some chroma sampling resolution using 4:1:1 or 4:2:0. MPeg cards usually use capture sections that sample at 4:2:2. Other things being equal*, to get equivalent luminance compression to DV (i.e. 5x, I frame only) with a 4:2:2 capture you would need 25% more data rate or 25 x 1.25 = 31 Mb/sec or double the 15 Mb/s. This assumes both compress ~5x. All of the higher luminance quality passes to the MPeg2 encoder and is used to compress to DVD rates. Chroma gets converted to 4:2:0 during the MPeg2 encode.

At those rates, MPeg2 capture would have equal luminance and chroma compression but twice the chroma sample rate at capture. This higher chroma sampling gets reduced to 4:2:0 during encoding. That leaves the issue open whether higher chroma sampling has any benefit for low quality NTSC and VHS sources. The case made above is that 4:1:1 is more than adequate for consumer video sources.

* Caveat is we don't know exactly how this increased chroma sample rate is used in the hardware MPeg2 codec.

[magelord]

I encode with H.264 to later put on blu-ray. 480 is supported by blu-ray not just 720p 1080i and 1080P. and next gen dvd is needed not just for the better video, but also so tv shows can all be fit on one disc and not 7 or 8 dvds for one season.

[lordsmurf]

I wonder if something like this would help improve the video quality of VHS captures in HD:

http://cgi.ebay.com/HDTV-Video-Enhancer-Converter-DVD-TV-Plasma-LCD-Monitor_W0QQitemZ5...QQcmdZViewItem

That looks like some off brand thing.
FAROUJDA is the name brand in upscalers and res boosters.