Why is capturing tv so damn hard...???[Update] [Archive]

Rickkins

24th January 2005, 23:10

I've spent the last several days reading, and trying, and reading and trying....

It just doesn't seem possible that this can be so damn difficult.
I bought the asus tv/fm card, with the Philips 7135 chip.
I have an amd 2500+ Barton core chip with 512 ddr (pc2700) ram.
Not the bleeding edge, but not a celery 500 either.
Video card is a fx5200/128 ram.

I've tried a whole bunch of programs.
Some I can get to work...capture nice...but such a freakin huge file...
There must be something I'm missing...and the words on the pages are beginning to blur.

I mean, if it's it really is this difficult to produce decent results
in a reasonable ammount of time....why in the world would any sane person
switch from videocassettes...????

Sergei_Esenin

24th January 2005, 23:58

jggimi

25th January 2005, 00:01

...but such a freakin huge file...

Let's do the math:

If you're capturing in YUY2, each pixel takes 16 bits. Multiply each pixel by the capture resolution, and framerate, and you will quickly determine how much space uncompressed video can take. (Capture in RGB takes 24 bits per pixel, and is not recommended, since our PCI cards capture in YUV/YUY2 anyway.)bits per pixel * resolution * framerate / 8 bits per byte * 60 seconds per minute * 60 minutes per hour / 1024 kilobytes / 1024 megabytes / 1024 gigabytes.
720x480 NTSC = 16*720*480*29.97/8*60*60/1024/1024/1024 = 69.45 GB/Hr
720x576 PAL = 16*720x576*25/8*60*60/1024/1024/1024 = 69.52 GB/Hr

If you're using HuffYUV, a lossless codec, you may expect somwhere in the range of 2:1 - 3:1 compression. This means you can expect full frame HuffYUV captures to consume disk space in the range of 23-35 GB/Hr, based on the compressibility of the content.

There are other codecs that may be suitable for capture, such as one of the commercial MJPEG codecs. These have greater compressibility, but that is because they are lossy, though they can be set to be "nearly lossless" though at that point, the space savings may not make much difference.

One could also capture with a DV codec, though everyone I know of who uses DV for capture do so with a Firewire attached DV camcorder. DV is a lossy codec that consumes 12GB/Hr.

...if it's it really is this difficult to produce decent results...

We'd be happy to lend assistance or provide additional guidance. But in order to help you, you have to help us. What tools are not working for you? What problems were they having? What tools did you get to work? And, what problems with them (besides big files) are you having?

Rickkins

25th January 2005, 03:43

Originally posted by jggimi
[B]...but such a freakin huge file...

720x480 NTSC = 16*720*480*29.97/8*60*60/1024/1024/1024 = 69.45 GB/Hr

If you're using HuffYUV, a lossless codec, you may expect somwhere in the range of 2:1 - 3:1 compression. This means you can expect full frame HuffYUV captures to consume disk space in the range of 23-35 GB/Hr, based on the compressibility of the content.

35 gigs for an hour video...?!?!?!?!

Are you telling me that's what "normal" is...???
Or is that more like the "outter stratesphere" quality wise....??
Is there not some middle ground, that will allow decent copy, along the
quality lines of a good svhs tape for example....with having to buy
a freakin spare 100gig hd...???

My mind boggles...!!!

jggimi

25th January 2005, 04:56

I've already described two middle ground solutions: MJPEG or DV. Here's another: Reduce resolution. If you chop height and width in half, your capture will consume 1/4 the disk space.

There's one other middle ground solution: Hardware MPEG-2 encoders, inboard or outboard. These operate in the same way as an outboard DV camcorder; they encode the video and compress to a lossy format; but MPEG-2 instead of DV.

Then there's the low-end solution that few would recommend: software MPEG compression. Most PCI cards include low end MPEG-2 encoders, and of course, one could compress to an MPEG-4 codec while capturing as well. Most don't recommend it because the CPU must do the compression as well as manage the capture, and MPEG compression is just not as fast as HuffYUV or MJPEG. Its discussed in our Capture FAQ -- but I won't provide the link as you already seem overwhelmed.

Sergei_Esenin

25th January 2005, 07:34

35 gigs for an hour video...?!?!?!?!

You seem to be ignoring the fact that that is an intermediate filesize. After capture, you can encode it to a 400MB/hr XviD file, or whatever you want.

Rickkins

25th January 2005, 19:07

Ya know guys...I'm not raggin' on you here or anything like that.
More, I'm just flabbergasted that despite quite a bit of reading done
beforehand, I had no idea that I would also need to acquire a 120gig
harddrive for my capture card to be any damn good. That little tidbit
of information should be listed somewhere, in huge bold letters.

Had I known that vital piece of info, I wouldda held off buying a capture card at this time.

And of course, I'm also pissed cuz' I can't afford to go buy a new 120 gig hd.( I say 120, cuz 80 wouldn't leave any wiggle room)

I could empty my 40 gig...but that would still leave me shy an hour...(at peak settings)

I'm also dismayed at the obvious shortcommings in this technology.
Let's face it...compared to popping in a tape and pressing record, this
process in mindbogglingly inadequate. (although I do accept that the final product is good...I take your word on that)

jggimi

25th January 2005, 19:53

It certainly appears to be "ragging" to me, if you define ragging to mean berate or scold. I could also describe it as whining. It certainly isn't helpful.

Try one of the middle-range solutions that have been suggested to you. Or even, a low-end solution, such as the MPEG-2 capture tool that probably came with your card.

Or, you could simply return your PCI card to the vendor.

Rickkins

25th January 2005, 20:15

Originally posted by jggimi
It certainly appears to be "ragging" to me, if you define ragging to mean berate or scold. I could also describe it as whining. It certainly isn't helpful.

Well no, I'm hardly scolding or berating anyone...it's not like anyone here
is personally responsible for the limitations of the technology.
Whining...oh you could be right there. Sometimes it is helpful, to me,
simply to howl at the wind...get it all out, as it were.:rolleyes: :p
It was, and remains, however, at non-directional whine...:D

Originally posted by jggimi

Try one of the middle-range solutions that have been suggested to you. Or even, a low-end solution, such as the MPEG-2 capture tool that probably came with your card.

Or, you could simply return your PCI card to the vendor.

Yup...gonna try some middle-range solutions, as suggested. Gotta try to find
a midrange solution until such time as I can afford a bigger hd...such
is life, n'est pas...??

Thanks...:)

jggimi

25th January 2005, 20:38

...Sometimes it is helpful, to me, simply to howl at the wind...get it all out, as it were.:rolleyes: :p
It was, and remains, however, at non-directional whine...:D...

Then I'm very glad you got your needed catharsis.

As I'd previously recommended, you might start by trialing an MJPEG codec (PICVideo, Leadtools, etc.) and see what lower quality levels do for you. You might also consider capturing at lower resolution, as it may meet your needs, depending on your final destination format.

Sergei_Esenin

25th January 2005, 23:32

I'm also dismayed at the obvious shortcommings in this technology.

Just to be clear, there are no shortcomings at all. Losslessly or near-losslessly compressed video *always* takes this much space and always will, period. If you want the lower quality and lower diskspace requirements and easy usability of a TiVo, you can get that using a hardware MPEG-2 encoder card and software like SageTV. The quality will never be as good, though just like TiVo's quality it may be adequate for most uses.

Buy the right tool for the right job. There are no shortcomings to a sledgehammer--it's the best thing for particular jobs. If however you use it to drive in a nail, it's just silly to complain it has an alleged shortcoming when the real problem is that you didn't use a regular hammer.

As jggimi suggested, try an MJPEG codec like PICvideo's. Set quality to 18 for lower filesizes, or even lower if lower quality doesn't bother you. If you like it Picvideo might still be running it sale for personal use users.

Rickkins

26th January 2005, 01:04

Originally posted by Sergei_Esenin
Just to be clear, there are no shortcomings at all. Losslessly or near-losslessly compressed video *always* takes this much space and always will, period.

Well, first off Sergi, don't take this as a personal affront, as I'm only
voicing my opinion.
But, just to be clear, the amount effort involved in making a decent
copy is most definitely a shortcoming, no matter how you slice it.
Despite all the jargon, not being able to copy as easily as videotape
absolutely is a shortcoming. All the spin in the world won't change
that.
Personally, I most likely will go out and get a new 160gig harddrive,
because I, (like many here I'm guessing), like to do things right.
And I will need, it seems, at least 160, as I seem to get 1gig per minute of copy..:rolleyes:

Originally posted by Sergei_Esenin

Buy the right tool for the right job. There are no shortcomings to a sledgehammer--it's the best thing for particular jobs. If however you use it to drive in a nail, it's just silly to complain it has an alleged shortcoming when the real problem is that you didn't use a regular hammer.

Cute analogy...but inaccurate.

jggimi

26th January 2005, 01:21

as I seem to get 1gig per minute of copy..

Are you capturing with HuffYUV in YUY2? Because that is close to an uncompressed YUY2 capture rate.

Rickkins

26th January 2005, 01:46

Originally posted by jggimi
as I seem to get 1gig per minute of copy..

Are you capturing with HuffYUV in YUY2? Because that is close to an uncompressed YUY2 capture rate.

Uh...yea I was...no good...??? :o

jggimi

26th January 2005, 14:53

What does GSpot (http://www.doom9.org/Soft21/Editing/gspot221.exe) say about your captures?

Rickkins

26th January 2005, 15:13

{from a test capture, vd & huffy)

This is a valid AVI file.

The filesize is 2,628 MB (or 2,691,491 KB or 2,756,087,776 bytes).
Of that, roughly 2,607 MB (99%) is video data; the remaining 20.9 MB ( 1%) is audio. This file has 134 KB (or 138,216 bytes) of extra "garbage" at the end that is not part of the data yet is not marked as "junk" either. This is not usually a serious problem, however, and is unlikely to cause a problem.

WildCelt

26th January 2005, 15:46

I hate the shortcomings of my PC. I mean, if I thought I had to type everything, I wouldn't have bought it! I was expecting it to read my brainwaves. (A bit of friendly sarcasm :) )

What shortcomings? It is what it is. Uncompressed (or gently compressed via huffyuv) video produces large files, end of story. I started off with a 40GB hard drive, then bought a 100GB hard drive, then another 120 GB hard drive. I now have the 120GB one, a 160GB one and a 200GB one, and I still run out of space!

Welcome to the obsession...

BTW, who told you capturing and encoding digital video was "easy"? It is not as simple as pressing "record," and, for truly superb results, never will be.

jggimi

26th January 2005, 15:47

I meant, what codec does it show in the .avi?

If HuffYUV was used for compression, it should show a FourCC code of "hfyu" and describe it as "Huffman Lossless"

If you don't see that (upper right corner of GUI), then HuffYUV compression is not being used.

If you do see it, then perhaps you are capturing in RGB, and, as has been described, that takes 50% more disk space.

You can check by creating a 2-line AviSynth script (.avs file) and opening it with VirtualDub:AviSource("c:\your\path\to\your_test_capture_file.avi")
Info()Create a 2-line .avs file like that, and open it in Vdub. Which colorspace is being used?

Rickkins

26th January 2005, 17:28

UYVY - UYVY (packed 4:2:2)

jggimi

26th January 2005, 19:03

I stated above that 23-35 GB/Hr should be expected, and the Capture Guide you found too complex states 20-40 GB/Hr, in its discussion of HuffYUV and MJPEG in section 1.2.

If you're seeing 60 GB/Hr, that is a result beyond reasonable capacity expectations.

Wilbert

26th January 2005, 21:33

Personally, I most likely will go out and get a new 160gig harddrive,
because I, (like many here I'm guessing), like to do things right.
If you don't want to buy a new hdd there are several options:

1) get mjpeg (as others said) and cap at full size
2) get ffvfw (now included in ffdshow) and cap at full size:

http://www.doom9.org/index.html?/capture/capturing_vdub.html (see FFVFW MPEG4 compression for settings)

3) mjpeg + vertical (and horizontal) reduction

4) ffvfw + vertical (and horizontal) reduction

option (4) will save you the most space.

Rickkins

27th January 2005, 16:49

Uh...you're not gonna believe this...but now my capped images are upside down.
Quite clear...but upside down.:rolleyes:

It's almost time for the sledgehammer :p

jggimi

27th January 2005, 18:08

I'll repeat the comment I made in my very first reply to this thread.

We'd be HAPPY to help you, Rikkins. But, you're going to have to give us more information than that.

Rickkins

27th January 2005, 18:34

Trying mjpeg and ffvfw, I'm not getting caps that playback literally
upside down. (capping with vdub...)

Wilbert

27th January 2005, 23:29

If you open your mjpeg/ffvfw in Vdub it shows up upside down?

Rickkins

28th January 2005, 02:10

Yup...regardless of what I opened them up in, they turned out upside down.
Of course they're gone now.
I tried a capture with windvd recorder...just to see what happened.
It gave me a decent cap, at about 7 gigs. I am now in the process
of turning it into an xvid....I guess I'll soon see.
I'm not expecting much. :D

28th January 2005, 09:57

@ Rickkins ,
You can use ffdshow MPEG-4 decoder, in MISCELLANEOUS you will find FLIP check box.

eb

Wilbert

28th January 2005, 10:15

Yup...regardless of what I opened them up in, they turned out upside down. Of course they're gone now.
Then a buggy decoder is decoding them. If you look under file -> file info (in VDub) you can find the decoder.

Rickkins

28th January 2005, 21:02

First off, I'd just like to say thanks to those folks who are choosing to help...much appreciated.

Now, ok....I got that squared away.
I've done a small test cap, using this method 1) get mjpeg (as others said) and cap at full size

Gave me a manageable size .avi file.
I've read the tut on post processing...forget it. While I have no doubt
that it is the bellweather of guides, it is far to complicated for the
average joe to figure out....and I can only aspire to average joe status.:p

I have managed to create from this, a dvd ready to be burnt....first usingy tempenc to change the avi to an mpegII file. Then doing a simple
demux to separate audio and video...then using dvdlab to create the dvd
itself...(not gonna burn it, as it is just crapp off cnn :rolleyes: .

The result was reasonable when played back on the computer..audio and video in sync, picture not bad.(etc etc)

This process was quite easy to do...but is there a process as easy that will also
produce equally good or better results...???

Also, I want to figure out how to turn the .avi that was produced originally into
an xvid file. I tried running autogknot with the mpeg I mage from tmpenc...no dice...got an error msg.

And, most importantantly, I need a method that won't make my brain bleed...:D :p :D :rolleyes:

Thanks again...:)

leoncheong

1st February 2005, 03:58

Have you tried capture directly to Xvid at 0.25PAL? You CPU should be able to handle this.

Rickkins

1st February 2005, 15:53

Originally posted by leoncheong
Have you tried capture directly to Xvid at 0.25PAL? You CPU should be able to handle this.

I've tried to capture directly to xvid...the results were not acceptable.
Note also that I am in NTSC territory.

I think that I am just gonna save up and get me one of these...
http://www.hauppauge.com/pages/products/data_pvr150.html

From what I hear, they are the bomb-diggiddy...:D

Boulder

1st February 2005, 21:39

That still doesn't save you from processing the capture:D The quality isn't that good if you just capture files that can be burned on DVD without any postprocessing and re-encoding.

Rickkins

2nd February 2005, 15:56

Well...I finally got a dvd made.
Capped with iuVcr, huffy codec@320x240...made approx 10 gig file.
Used tmpenc express to create mpegII (11hrs)& tmpenc dvd author to create the dvd.

Two things.

1) The picture looks way better on my tv(32" Toshiba)
2) The sound gradually goes outta sync.

Boulder

2nd February 2005, 16:16

I strongly suggest you use either VirtualVCR or the latest VirtualDub for capturing. VirtualVCR has a dynamic audio resampling option which keeps the a/v sync, VirtualDub drops a frame here and there (=unnoticably) to keep it unless Avery has added the resampling feature too since the last time I used it for capturing.

You're also much better off capturing 480 lines. If you want a n x 240 resolution, just deinterlace (if necessary) and then resize. You are throwing away half of the available information if you capture directly to n x 240.

Rickkins

13th February 2005, 20:41

Well...finally got one freakin' movie successfully recorded to dvd.
Ended up going with VirtualVcr, pic mjpeg, 640x480. Gave me a file of
24gig for just over 100mins of movie.
Used tempenc xpress to recode it to mpegII format. This took over 11hrs.,
not on max quality settings.....(I like to find something quality but
faster)
The final stumbling block seemed to be the final process. Tried a half
a dozen times to create a dvd, using a rewritable disc. All attempts
were unacceptable....till I finally decided it must be the media itself.
It was....burnt in onto a dvd-r (ritek)...very decent results.

Some tweaking will be attempted on future projects, no doubt.

Thanks to all who helped out by answering noobular questions.:D

Next project....capping some tv epps, and cutting out commercials...

Watch this space for more noobular questions....:p

Boulder

13th February 2005, 21:25

Try encoding to MPEG2 with QuEnc and 2-pass, you might be pleasantly surprised. The quality should be equal or better than TMPGEnc (which then again excels in MPEG1). Just keep the "Extreme&Slow" disabled and it won't be slow at all.

I also suggest you learn some basic Avisynth stuff;)

Rickkins

13th February 2005, 21:46

Originally posted by Boulder
T

I also suggest you learn some basic Avisynth stuff;)

But will it hurt...:p

joshyg2

14th February 2005, 05:11

just putting my 2 cents,

now days in our digital world, its so easy to capture from TV (digital that is), because all you are doing is transfering the original broadcast data on to your hard drive. then proccessing later, thats if you choose too.

but in the analogue world, its much hard to capture video.

i live in Australia, where digital came very late, and still isnt very mainstream, most people still use analogue.

so i learn the hard way how to capture.

i started off with the cheapest capture card you could imagine, its was like $60 AUS. I used the software that came with it, and captured to DivX on the fly.

i swear it was the crapiest looking video you have ever seen.

but after reading, and reading on the net, i discovered new programs, etc.

it took me about 3 years to perfect analogue capture.

i bought a new Hauppauge Stereo Analogue TV Tuner.

i used all the right software, i bought new coaxial cables etc, captured to HuffYUV full res: 720x576. bought new hard drives, 4 - 250 GB drives - RAID. and i was away.

then it took me a year to learn AVISYNTH and i am still learning / experimenting with different filters.

but yeah, it takes a lot of time and effort and $$$ to learn how to capture analogue video properly.

and i dont like people that complain about hard drive space, never try and reduce your filesize. Maximize your hard drive space (MORE DRIVES). the only way to go.

anyway, i plan to soon get a DVB capture card, so i dont have to do as much work.

josh

Boulder

14th February 2005, 09:34

Originally posted by Rickkins
But will it hurt...:p
Nah.. you don't need any of the fancy stuff for basic processing.

1)Load the source
2)Crop
3)Denoise -- RemoveGrain().RemoveDirt() works wonders
4)Resize

That's it;) You can use VirtualDub to determine the area to be cropped (the black borders) and FitCD will do the resizing math for you.

JimmyBarnes

17th February 2005, 10:18

Originally posted by Rickkins
I've tried to capture directly to xvid...the results were not acceptable.
Note also that I am in NTSC territory.

I think that I am just gonna save up and get me one of these...
http://www.hauppauge.com/pages/products/data_pvr150.html

From what I hear, they are the bomb-diggiddy...:D

I presume that this is a so-called hardware capture/converter device (and anathema to purists). But 2 GB/hr for "DVD-like" quality sounds a lot more practical than 65 GB/hr for that "perfect capture"...

I'm still using VHS tape and haven't seen what one of those DVD recorders with hard-disk can do...

I've had a PC for some time in the loungeroom which I use for playing most any format, and have always wondered what it would take to upgrade it to a TV recorder. Am beginning to get a clue...

JB

emazur

5th March 2005, 20:11

"2) get ffvfw (now included in ffdshow) and cap at full size:

http://www.doom9.org/index.html?/ca...uring_vdub.html (see FFVFW MPEG4 compression for settings)"

I tried this but when I played back the movie it was upside down and reversed (picvideo mjpeg caps doesn't do this). How can I fix it?

Also that link on doom9 points to a version from 2003, so I tried a newer version instead from:
http://www.afterdawn.com/software/download_splash.cfm
but I don't see any "MPEG-4 Codec" under the codec list in my capture soft, so I uninstalled and used the 2003 version which works. What's the problem there? I did try that registry hint on the doom9 link just in case but no luck

PhillipWyllie

28th June 2005, 01:57

Firstly if you get the "bug" and chose to get "rid" of your VCR for recording you're going to need at least 2 100Gb HDDs.
I'd use VirtualVCR to capture @ full resolution( compresing with huffyuv in the yuv2 colour space). Learn and use Avisynth to do the post processing. Get rid of TMPEGEnc and get CCE( it's much much faster) to encode to MPEG2. Like most things, the more you practice the better you get.
Lastly never ever capture in MPEG( or any lossy format for that matter).

Fluffbutt

8th July 2005, 16:21

This is very simple. If you want best quality, capture first using huffyuv or MJPEG and then encode afterwards to a smaller format like XviD or Windows Media 9, preferably using 2-pass mode. The "huge files" are an intermediate step and can be deleted after encoding.

If you want a one-step encoding process, you should get a hardware encoder--but the quality will not be as good.

Excuse the busting into someone else's question, but i'm also a noob for captures and I'd like to ask what's wrong with hardware encoders?

I'd have thought that a hardware based encoder like hauppage 150/250 cards would be the beesballs for encoding...?

PhillipWyllie

8th July 2005, 19:28

I'm assuming you mean the Hauppage WinTV-PVR-150 and WinTV-PVR-250 which are capture cards not harware mpeg2 encoders. They capture in the MPEG-2 format. My advice is to capture in AVI( with little compression as possible [huffyuv]) using Virtual VCR(free), edit in avisynth(free), encode to MPEG-2 with CCE and author with scenarist.
If you wish to encode using hardware check this out: http://dv411.com/encoders.html