Hi everybody. I'm about to start digitizing our family's tape collection. The earliest tape is from approximately 1983. I plan to use a Panasonic NV-HS1000 and a USB capture device.
I'd like to ask for your help, but am not sure how to do so; what information to include, how to encode the samples, where to upload and similar. I have some basic knowledge of Virtualdub and Avisynth meaning that I know how to perform any action seen elsewhere (e.g. copy a script, download and set up any missing plugins for Avisynth or anything similar). I've seen great results on a topic about the wedding tape so I'm hopeful that you would be willing to help me as well.
I already have one of the tapes captured with a simpler VHS deck, and saved with Lagarith. I will be getting the Panasonic deck over this weekend, so I can provide samples with that device the other week.
Thank you very much!
+ Reply to Thread
Results 1 to 17 of 17
ipavlic, in the future please use a more descriptive subject title in your posts to allow others to search for similar topics. I will change yours this time. From our rules:Try to choose a subject that describes your topic.
Please do not use topic subjects like Help me!!! or Problems.
And welcome to our forums.
EDIT: Basic information: What are your computer's capabilities/ specifications? Describe your present capture setup. Describe your present work flow. Which USB capture device? Are the tapes in good condition? Are you in NTSC or PAL format? Probably a lot more. Stand by.
Last edited by redwudz; 20th Feb 2014 at 20:26.
You can upload samples to this web site. See the "Upload files/Manage attachments" button below the quick reply box near the bottom of each thread. I think the current limit is 500 MB per file.
To clarify Baldrick's and jagabo's suggestions: details about how you captured will allow readers to give more specific tips. Also, a short 10 seconds or so of an unprocessed sample of your capture is worth a thousand questions and guesses, especially scenes that have people moving or the camera in motion. If you captured to AVI, you can cut a short sample in VirtualDub and save the clip using "direct stream copy" (see VDub's "Video" top menu item). "Direct stream copy" will avoid colorsapace changes and other alterations.Our inventions are wont to be pretty toys, which distract our attention from serious things. They are but improved means to an unimproved end. -- Henry David Thoreau
- Macbook with 64-bit Windows7 on a Bootcamp partition. The processor should be i7-2620M, 8GB of ram. Additional external USB 3.0 WD drive to help store the captured files.
- USB grabber is a basic type, I believe it is called just "SilverCrest USB 2.0 Video Grabber". By the manufacture date, it should be using a Conexant CX23103-11Z chipset (earlier versions used a em2863 USB 2.0 Video device from EETI (former eMPIA Technology, Inc))
- I don't know how to judge the condition of the tapes. They have not be played often and they seem to be in good condition. Given that they are old, that may not be true though.
- PAL format is used.
- pressing play on the video
- waiting for it to autotrack
- connecting the USB grabber to the video
- going to Virtualdub and choosing Lagarith as an output codec with no additional configuration
- choosing save as AVI
I'll be uploading the samples as soon as possible. They are huge.
True. Apparently ipavlic didn't configure Lagarith, which defaults to RGB. Even at that, 10 seconds of PAL 720x576 or 640x480 video compressed with Lagarith RGB would be less than 150 MB. The O.P. might be looking at uncompressed RGB, which would be 3 times that size. Or it could have been captured to a larger frame size.Our inventions are wont to be pretty toys, which distract our attention from serious things. They are but improved means to an unimproved end. -- Henry David Thoreau
Hi everybody. I apologize for the delay. I've managed to upload the samples from my initial capture:
Sample 1 - shows movement and color
Sample 2 - shows dynamic range and some artifacts:
Sample 3 - shows colors and some artifacts:
Sample 4 - shows an outdoor shot
Thank you for the samples. I'll be looking at them and will report later today. I expect others will report as well.
Briefly, your samples appear to be lossless YUY2, which is preferred for VHS capture. However they also appear to be uncompressed. I recompressed all 4 samples with Lagarith YUY2 using "fast recompress" in Virtualdub. These are the files sizes after using Lagarith:
sample posted Lagarith ------ ------ -------- Sample 1 205.2 MB 104.2 MB Sample 2 246.1 MB 105.3 MB Sample 3 115.3 MB 56.8 MB Sample 4 164.3 MB 82.9 MBOur inventions are wont to be pretty toys, which distract our attention from serious things. They are but improved means to an unimproved end. -- Henry David Thoreau
Those are some pretty sick looking caps. Hopefully the HS1000 will clean up the time base. And using an s-video cable instead of composite should remove all that herringbone noise in the colored areas. Your caps also have a lot of blown out brights. Be sure to adjust the capture devices proc amp to bring those down to normal levels. There are a lot of auto-gain problems too. If they are on the tape (at least some of them probably are) there's not much you can do while capturing. But be sure the capture device isn't exacerbating them -- turn off any auto brightness, contrast, etc. settings in the device.
Until you have some better caps there's not much point in discussing filtering.
@ipavlik, if by "artifacts" you refer to rips, streaks, spots, comets, mangled and hopping frames, etc., those are not digitally created. They either exist in the original tapes or they result from poor tracking and tape damage. Herringbone, chroma upsampling errors, and dot crawl can be avoided using s-video rather than composite from the VCR. And I feel that you might be able to find a better, not much more expensive, capture device.
The image below demonstrates real "artifacts" -- herringbone, dot crawl, and ragged, sawtooth edges. The image is from Sample 3 and has been blown up 2X, although you really don't have to magnify the image to see these problems. It's possible to smooth this disturbance with processing, but the filters required to fix noise of this kind will seriously degrade the image.
The image below is from Sample 3. It is a 2X enlargement of part of the right-hand border. Note the bends in the border edge, which indicates time base errors on playback. during playback these bent verticals and other edges will "wiggle". There is no way to repair this in software; a line-level tbc is required to prevent these effects. A tbc can also avoid many of the bent and torn frames in your capture. Hopefully the better VCR will have this feature. If it doesn't, there are ways of working around it. The ragged edge on the border is caused by dot crawl and herringbone.
The image below is a before/after of frame 71 in Sample 4. This has been captured at black levels that are too high. If the shots were overexposed in the camera, you can partially compensate by adjusting input brightness and contrast levels during capture. I would not attempt color correction during capture -- doing so during capture would be an exercise in futility, because analog tape changes color balance from scene to scene. The color here was fixed in software, using the original YUY2 colorspace to first restore reasonable black levels and reduce green, then mild tweaking with RGB color filters. Note that the color fix for this scene won't work with the scene that follows it, which has different levels and color balance altogether.
Take another look at your Sample 4 .avi and notice the right-hand border. Near the top of the frame, the right hand border leans to the left. A tbc can probably fix that.
Last edited by sanlyn; 23rd Feb 2014 at 03:55.
avisynth. The results were not bad. Should I instead fix each scene manually?
Last edited by ipavlic; 23rd Feb 2014 at 05:32. Reason: added a question about the capture device
Autolevels (otherwise known as autogain or auto gain control = AGC). These give unpredictable results and, specifically for the Sample 4 video, you'll get brightness and chroma "pumping" whenever bright or dark objects enter the frame, or when the lighting changes in a scene. The problems and severe luma fluctuations in your other three samples are the kinds of damage AGC can cause. The demo scene I worked with required basically offsetting luma levels downward by about 25%, which was not difficult.
An example of autogain problems: a video features a man in a dark navy blue suit who is instructing a class and writing on a grayish-green blackboard. While he writes on the blackboard, his back is to the camera. When he turns to face the camera, his white shirt causes the picture to darken. When he turns around and starts writing on the blackboard, the picture brightens again. Then he turns to the camera (picture darkens), turns halfway to the blackboard (picture brightens), turns to the class again (picture darkens), and so on. Soon a female student wearing a bright red dress enters the picture: it's a very bright dress (picture darkens). When she leaves the frame, the picture darkens again.
A better VCR will improve tracking, assuming the machine is in working condition. If you find a good VCR but it has no tbc circuitry, a used DVD recorder can be hooked up as a pass-thru tbc device: you "play" the video through the recorder and into the capture device, rather than record to lossy MPEG (which would be a bad idea with damaged tapes). A thread on using pass-thru devices is here: http://forum.videohelp.com/threads/319420-Who-uses-a-DVD-recorder-as-a-line-TBC-and-what-do-you-use.
Two popular capture USB devices are the Diamond VC500 series and the ATI 600 USB. Both are available used and accept s-video connections. I'd advise that you find a better VCR for starters; you might get better results with your Grabber using a higher grade player. But I haven't seen any rave reviews concerning the Grabber or other similarly low-cost devices. The EZcap was another favorite -- but beware of cheap imitations, some of which even use the EZcap name. Because you are in the UK, some of the better devices might have different brand names than are used in the USA.
Last edited by sanlyn; 23rd Feb 2014 at 22:46.