What is best capture device

[oln]

I can't speak for the HXD-890 specifically, but I've used the earlier HX750 and related pioneer dvr-440/540 and they can do a pretty decent job at stabilization

For the Sony RDR-HX750 and newer the sony dvd-recorders are like 95% the same as the pioneer variants, using the same chipsets. Earlier Sonys are completely different, for pioneers afaik is the DVR-x3x and newer that work well. There are a handful of Toshibas that were based on the same chipset like the RD-XS24 which I have and act very similarly though the user interface is different and no NTSC support. Most Toshibas used different chipsets however.

They are a bit more prone to frame drops (seems they might have a shorter buffer?) and the horizontal stabilization is not quite as solid on very wiggly tapes, but on the other hand they let you adjust levels, contrast and some agc stuff so you don't get the brigthness clipping issue of the PAL Panasonics. There are also a few additional parameters one can adjust with a service remote though I haven't tested that much. The PAL/International variants also support PAL-60 and NTSC 4.43 which is convenient for doing NTSC tapes without needing a multi-system deck, while the panasonic variants only supports standard NTSC.

Of course with the HX750 and later models like the HXD890, you can use HDMI out and bypass the digital->analog conversion entirely (needs a splitter to evade HDCP no matter the content). At least on the HX750 this also gives you the full 720 pixels without blanking 8 px on the left/right edge (Same with all other dvd-recorders I've tested HDMI out on including panasonics so it's not specific to these per se). Component out is also an option.

The 3 pioneer 540/440's I have I've have this odd issue with a noise on the audio input every few seconds which is a bit annoying, no idea what causes it so have to bypass audio on them. Not an issue on the HX750 though so maybe I've just been unlucky with those, though 2 of them I got for free anyway so can't complain.

[lollo]

Thanks oln, always very useful! Pros and cons on every device.

Of course with the HX750 and later models like the HXD890, you can use HDMI out and bypass the digital->analog conversion entirely

Alwin, this is a good point. If you are forced to use this machine because your VCR has no lineTBC; you may consider capturing its HDMI output. The signal is digitized inside the Sony for processing; converting it back to analog and captuing it with a card means an additional lossy D/A conversion. You just need a HDMI card to dump that stream on your PC. Several examples in the forum are given.

[DB83]

15 hours later.....

8 views. No comments. No problem. Must assume that all was fine.


But maybe not.


BTW. No filtering. No TBC. No Y/C either

[lordsmurf]

The preference for capture according to The Lord is VDub 1.9.11, which I am using.

What did you do after the capture?
Is that the raw capture file, or did you edit anything in VirtualDub?

[Alwyn]

What did you do after the capture?
Is that the raw capture file, or did you edit anything in VirtualDub?

I did a Direct Stream Copy to trim the file size, but otherwise nothing. I have checked the master capture files, and they are RGB as well. All the LAGS files are YUV.

[dellsam34]

If you are saving to a lossless compressor you should choose full processing mode.

[Alwyn]

If you are saving to a lossless compressor you should choose full processing mode.

I don't know what you are on about. I captured HUFFYUV and wanted to post some of the video here, so I used Direct Stream Copy. Surely using Full Processing is going to result in an export that is different to the original captured video? What are you suggesting I do to create a postable size, exactly replicating my capture?

[Sharc]

15 hours later.....

8 views. No comments. No problem. Must assume that all was fine.


But maybe not.


BTW. No filtering. No TBC. No Y/C either

So here my few comments:
The video levels are perfectly within limits.
The video is progressive, or PsF, rather than interlaced. I guess it was like this on the tape.
On top you see some flagging and wiggle. It's not so eye-catching because the picture is relatively dark. A "line TBC" would fix this.
It's possible that the composite source had little spectral overlap of luma and chroma (the picture looks relatively soft/lowpassed), hence the requirements for the separating filter was low, means the weakness of the USB-live2 didn't materialize (just my speculation ....)

Not bad, after all.

[dellsam34]

If you are saving to a lossless compressor you should choose full processing mode.

I don't know what you are on about. I captured HUFFYUV and wanted to post some of the video here, so I used Direct Stream Copy. Surely using Full Processing is going to result in an export that is different to the original captured video? What are you suggesting I do to create a postable size, exactly replicating my capture?

You cut your scenes, use full processing mode and save as HuffYUV again with the exact same parameters, Full processing mode with a lossless compression doesn't mean you are re-encoding.

[lollo]

15 hours later.....

8 views. No comments. No problem. Must assume that all was fine.


But maybe not.


BTW. No filtering. No TBC. No Y/C either

Sharc already said everything, in his excellent way as usual.

I can just add, that your capture is not bad, a lineTBC is needed (you do not have straight vertical lines to highligth the problems).

USB-Live 2 is marginal when working with unstable signals. It completely fails on Alwin's tape because his tape provide a worse signal to the card than yours. It's difficult to generalize. Same for Y and C separation, problems are more evident in some cases than others.

Once more, the more important elements for a good capture is the VCR and the conditions of the tapes. For dealing with a wide range of tapes is often recommended to have more than one VCR, i.e. a JVC and a Panasonic, because they react/fix differently at specific problems.

The alternative is to leave the "bad" signal as it is and fixing it inside the capture card as for dellsam34 approach, with a high-end capture card like BrightEye 75 and/or Snell&Wilcox TBS800. The (small) advantage in this case is that you have less A/D/A conversions in the chain, but these devices are hard to find and very expensive.

[Alwyn]

Dellsham, a full reprocess completely changes the file:


[Attachment 65552 - Click to enlarge]

[lollo]

Alwin, is probably better to start from scratch:

• do not download any prepared or pre-assembled package, you have less control on what is inside, and ofthen they are bad

• remove or temporary hide any VirtualDub directory (it does not "install" so you can use a new dir and execute from there)

• do everything for 32-bit version

• uninstall HuffYUV and manually remove all related files in windows systems directories

• dowload VirtualDub 1.9.11 from official site https://sourceforge.net/projects/virtualdub/files/virtualdub-win/1.9.11.32842/

• download original HuffYUV 2.1.1 from https://www.videohelp.com/software/HuffYUV (for example) and install it

Configure VirtualDub:

VirtualDub Settings						
Video		X	Overlay			
			Video source >		X	Video Svideo
			Capture pin…			PAL-G
							25000
							YUY2
							720x576
			Capture filter…			: Standard Video PAL_G
			Crossbar…			1: Video Svideo In
			Cropping…			0 0 0 0
			Compression			Huffyuv

			Set custom format		720 576
							YUY2
							YUV 4:2:2 interleaved
Audio						
		X	Enable audio capture			
			Raw capture format…		PCM, 48000Hz, stereo, 16-bit
			Compression…			<No Compression (PCM)>
			Audio source		X	Audio line
		X	0 capture device			
Capture						
			Settings…			25000
			Timing		X 		Do not resync between audio and video streams
					X 		Automatically disable resync when integrated audio/video capture is detected
					X Automatic	
							Number of audio blocks to use at start 30
			Disk I/O…			512K
							2
					X		Disable Windows write buffering

Now you should have YUV 4:2:2 capture

[Sharc]

@Alwyn: The colorspace is now YUV 4:2:2, and the file size became lower accordingly.
Still wondering why your original captures were in RGB .....

Edit: lollo was faster

[Alwyn]

Me too, Sharc! As I said before, I normally don't use HUFF; I use LAGS or MagicYUV and both come out in YUV. I only did these in HUFF because my perception is that that is the most common codec used amoungst you aces.

Anyway, I've just done some tests using Lollo's settings (but without re-installing HUFF, as it is a nightmare to do so). I have a total of 8 versions of VDub here.

I started at the Sourceforge version, then tried Lordsmurf's, 1.9.11AMD64, 1.10.4, 1.10.44gbpatch. All gave me RGB.

Then I tried VirtualdubMPEG2 and it gave me YUV (as does Virtualdub2, as I think I previously noted).

My hunch is that there is a bug in HUFF with those early versions of VDub and PAL captures, but I will re-install HUFF later and re-check.

[DB83]

Thanks guys.

It was probably a 'biased' sample and I agree that results will vary accordingly.


As for the interlaced/progressive comment, that does puzzle me. I, typically, just take mediainfo's word for it although even thpugh Lagarith is in variance since it does not even report that.

[lollo]

Many (most) of my captures from TV shows are not interlaced but "progressive", in the sense that the 2 fields are from the same moment in time.

Do not rely on information form any software, but check yourself.

Summary of http://neuron2.net/faq.html (not available, except in web.archive.org):

How do I figure out the nature of my source video?

It's important to know whether your video is progressive, interlaced, telecined, field-blended, etc., when making decisions about how to process it.

Here's a reliable method for analyzing your video:

The first step is to create an Avisynth script that sources your video and separates the fields. If you have MPEG video, you can use DGIndex to make a project with "Honor pulldown flags".

Then make a script to step through the fields:

MPEG2Source("file.d2v")
AssumeTFF() # or AssumeBFF(). See the next FAQ.
SeparateFields()

If you have a different video source type, just use an appropriate source filter, such as AVISource().

Now serve the script into VirtualDub and step through the fields in a high motion section. We will label a new picture with a new letter below. If you see a repeating field sequence like this:

... a a b b c c ...
then it is progressive video. If you see this:

... a b c d e f ...

then it is pure interlaced video. If you see this:

... a a a b b c c c d d ...

Then it is 3:2 pulled-down progressive video.

Make sure that you always apply this test in a scene with continuous movement, e.g., an object moving steadily across the screen. There is no more reliable way. It also allows you to spot field blending very easily. If you see any blended pictures, then you have field blending.

[lollo]

but without re-installing HUFF

big mistake

as it is a nightmare to do so

1- unzip huffyuv-2.1.1.zip to C:\

2- Copy huffyuv.dll and huffyuv.inf in C:\Windows\SysWOW64

3- Open up a command prompt (DOS box) as administrator (in the keyboard Windows + X -> Windows PowerShell (administrator))

4- In the command prompt (DOS box) go to C:\Windows\SysWOW64

5- Type rundll32.exe C:\Windows\SysWOW64\setupapi.dll,InstallHinfSection DefaultInstall 0 C:\huffyuv.inf

edit formatted instructions as code otherwise lines are cut

edit 2 also be sure that the codec is properly configured:


[Attachment 65554 - Click to enlarge]

[Sharc]

As for the interlaced/progressive comment, that does puzzle me.

PsF = Progressive segmented Frame is further explained here. It's used to satisfy playback scenarios which expect an interlaced format, but the source is progressive and has no native field structure.
PsF is also called pseudo-interlaced etc. You find it on DVDs as well.
https://en.wikipedia.org/wiki/Progressive_segmented_frame#:~:text=Progressive%20segmen...ther%20segment.

For producing true interlaced video from progressive footage one would have to synthesize the missing field by interpolation.

[DB83]

I have a wild 'theory' of why my sample is progressive even though I would have expected a commercial VHS source to be interlaced.

Since the usb-live2 defeats mv could it be that it has a 'TBC-like' feature to process a complete frame rather than alternating fields ?

[lollo]

No, it captures what's on the source, i.e. the fields, without caring of what they represent

[Alwyn]

Lollo, All done, same result. Good thing I remembered my DOS "cd"

I can't understand why VDub MPEG 2 and VDub 2 capture in YUV, as set in the Custom Video Format box.

Another interesting thing is that, with the GV, the only "data format" I can choose is YUY2. For 16bit RGB, 24bit RGB, 32bit ARGB and UYVY, YUYV I get a message saying those codecs aren't supported by the GV.

Here are the two files I have now got installed:


[Attachment 65555 - Click to enlarge]


[Attachment 65556 - Click to enlarge]

[lollo]

The size of your huffyuv.inf files is different than mine. Uninstall and manually remove later all huffyuv files, then install huffyuv from scratch.

[Sharc]

From the filesize it looks like Alwyn's huffyuv.inf belongs to the 64bit version.

[lollo]

He's maybe mixing stuff. That'a why I suggested to remove everthing and start from scratch. 32-bit only first. Then we'll see...

[Alwyn]

Lollo, what are your file sizes?

[lollo]

here my HuffYUV file: huffyuv-2.1.1.zip

[DB83]

@Alwyn

Where do your files come from ? I see 2020 as the creation date yet even the unofficial 2.20 version is some 15 years earlier.


The last offical version is 2.1.1 and that is dated 2000. I also have a 64bit version dated 2005. Even so there is a vast gap between 2020 and 2005 and my own 32 bit installation would have later than either of these.


And you hardly need a DOS dosk to install. Right-click on the .info from Windows Explorer (I assune that Win10 works the same) and select "install"


And RGB, IIRC, is the default in vdub. Some while back I stood aghast and the filesizes that these lossless codecs were producing. Then I realised that one could change the compression to YUY2 (YUV)

[Alwyn]

Folks, nothing I do will give me a YUV capture with those early versions of VDub. Here's what I did:

Deleted all HUFFYUV.DLL and INFs from Syswow64 and System32. Ran the uninstallers in Programs and Features (there were two there). Ran through the registry and deleted all HUFF relevant entries (there were a few). I left a few that related to PDFs and MS Office. Rebooted (actually a few times) and confirmed with GSpot and Virtual Dub that HUFF was not installed.

I then followed Lollo's instructions to the letter. And then got an RGB HUFF AVI with VDub 1.9.11. I then did another uninstall/Reg wipe and tried Hofmand's installer but that resulted in another RGB file. Interestingly, his HUFF entry in VDub is a little different, it's called "Huffyuv 2.1.1-CCESP Patch v0.2.2" (and his MSI didn't work; I had to run that in the command prompt as well due admin rights).

One thing I noticed was that after running Lollo's command line, the huffyuv.dll file in Syswow64 got changed to HUFFYUV.DLL.

I see the end of the command line refers to C:\huffyuv.inf. Is that correct?

After all that, I've had enough. I'll keep posting my clips as HUFF/RGB for standardisation but I'm moving on with LAGS and MagicYUV.

@DB83, the dates are coming from when I download them and when I extract them ie recently.

And you hardly need a DOS dosk to install. Right-click on the .info from Windows Explorer (I assune that Win10 works the same) and select "install"

I am aware of the procedure for installing via infs. I am following Lollo's procedure, which requires a CD at the DOS prompt to change the directory. I assume Lollo is sending me down the command-line path for a reason. FYI, "CD" is DOS for "Change directory".

And RGB, IIRC, is the default in vdub. Some while back I stood aghast and the filesizes that these lossless codecs were producing. Then I realised that one could change the compression to YUY2 (YUV)

Please read my post #117. FYI, Virtual Dub will remain set on what you had for the previous capture. Only the Export codec flips back/defaults to uncompressed on every occasion. And HUFF RGB is not "uncompressed".

[lollo]

Huffyuv 2.1.1-CCESP Patch v0.2.2 was a hack to the original codec to feed Cinema Craft Encoder, a MPEG-2 encoder that many were using 20 years ago (but it was quite expensive, so lot of people were using warez with all its negative consequences of instability and legal aspects). It should not be used for capturing.

The procedure for install original HuffYUV in post #107 has some redondancy but is safe, and you followed. I do not understand why you are not able to capture in YUV.

In system32 you should have no HuffYUV files.

[Sharc]

In system32 you should have no HuffYUV files.

Hmmm, are you really sure? I do have the .dll there as well. I don't remeber how I installed it in W10 few years ago .... but I think it was basically the same as your proposed procedure.

From my notes:

Download the 32 and 64 bit Huffyuv 2.1.1 from https://www.videohelp.com/software/HuffYUV.
Once you have downloaded them, extract to a folder. Next you will need to open a command prompt as Administrator. Then run the following commands.
(In this example I have extracted to the P:\Program Files Video\Huffyuv directory)

32bit huffyuv:
rundll32 C:\Windows\SysWOW64\setupapi.dll,InstallHinfSectio n DefaultInstall 0 p:\Program Files Video\Huffyuv\huffyuv.inf

64bit huffyuv:
rundll32 C:\Windows\System32\setupapi.dll,InstallHinfSectio n DefaultInstall 0 p:\Program Files Video\Huffyuv\64bit\huffyuv.inf

And I have for some reason bookmarked the link below, which explicitly requests

NOTE!
On Windows operating systems with 64 Bit architecture, the dll file must be in both the "sysWOW64" folder as well as the "System32" folder. In other words, you must copy the "Huffyuv.dll" file into both folders.

https://www.dlldownloader.com/huffyuv-dll/#Operating-Systems-Compatible-with-the-Huffyuvdll-File