Into another round of work on old family VHS tapes.
I appreciate the material is not of great quality ......... but have several digitized tapes which have same 'problem'
Would like to 'understand' how to read the clip .. to know what is wrong and then best approach to resolve ............ These are from PAL VHS-c tapes circa mid 90's
Don't want to spend a huge amount of effort, but be good to know if there is some simple fixes I could apply ... especially to fix what appears to be ghost image.
+ Reply to Thread
Results 1 to 26 of 26
Last edited by Tafflad; 20th Jan 2016 at 16:08.
IMHO There is no easy answer to your question.
Video is a learning curve and every situation will be different.
Possibly, in this case, there is some bad interlacing artifacts. It might help if you were to describe your capture work-flow - capture card etc. I thought you could only capture DV.
I see the clip is high bitrate lagarith and that may also influence what you see i.e. a bad capture. Any dropped frames etc. ?
There is absolutely no point in making it larger via Lagarith. The DV capture is "lossless" in the sense that it is bit-for-bit identical to what is coming out of the ADVC300, since it outputs DV. I see absolutely no point in capturing using Lagarith, because it adds nothing except for making the file sizes enormously larger.
I looked at the video. It is being reported as upper field first, but it is actually bottom field first. So, it is a simple field reversal problem.
Make that change (to lower/bottom field) in your editing program (or in AVISynth) and all should be well.
It is actually a remarkably good-looking capture for VHS source material. Nice job.
^^ Could that 'unneccesary' conversion have caused that field switch ?
Methinks the OP has been confused by several suggestions on here that lossy sources should be converted to lossless for editing purposes. However DV sits in the middle - neither lossy or truly lossless. But if the ultimate target is lossy eg mpeg2(DVD) I see no point in additional conversion and opportunity to 'screw' things up.
Guys .... apologies for my confusion over Lagarith
I had followed advice here that when editing in VD save the output as Lagarith to make it easier when using it in Vegas. I should have loaded original DV. (to be honest I thought I had - mistakenly I had cut sample from wrong file)
Anyway cut out the same section this time original DV capture ............. welcome advice and especially on how 'read' the problem.
The original does report as 'Bottom Field First'
Since you made your original transfer with DV stay with it. The Lagarith route is most useful if you're going to do heavy, heavy processing.
DV has a lot of advantages. As an i-frame only format it is very smooth to edit. If all of your edits are straight cuts -- no added effects -- there is no generational loss in editing. Vegas, Vdub and most other half-decent software will simply pass long the unmolested data stream in your revised order.
Bottom field first is standard for DV. Nothing wrong with it.
OK .... thanks for that explanation .... back to original Q ... how can I improve this ?
So is this 'ghosting' you see the alternative lines on the edges of the rider ?
If the answer is yes then that IMO is just normal interlacing.
DB83 has the best advice. I've looked at your DV version of the capture and, like the Lagarith version (once the field reversal was corrected), it is a really good-quality VHS capture. I see absolutely no major problems with it, and not a hint of ghosting of any kind. The only minor improvement would be to remove some of the chroma noise (those unwanted colors you see on the water).
So, I too think you are simply seeing interlacing artifacts which are not a problem with your video, but instead are a problem with a software video player that isn't set to deinterlace. If you play this on any TV set, or if you change the deinterlace setting on VLC as DB83 suggests, it will look great.
DV is very lossy, especially for NTSC.
P.S. Here's a link to that ancient (2003) multi-generation (100 generations!) test done by a member of the Sony (then Sonic Foundry) forum:
Multi-generation DV render test
OK .. Pint owed
It is better with Deinterlace ON
Not sure what option .... Mean or BOB seems to give best results ...
So if I want to convert these tapes to digital files ........ either run off MPEG 2 DVD or as MP4 for on-line ........ what should I do to the DV capture, do I need to run it through some DeInterlace tools
I will be doing as usual my final editing in Vegas .... just want to be sure I am putting files in 'optimally' pre-treated.
I know that many posts advise against de-interlacing.
If it's better to fix first - tips welcome, I'd also look at fixing Chroma at same time ... in the past one of your posts advised using CNR plugin in VD, this still the best option.
No. ALWAYS keep an interlaced source interlaced.
Already explained, and dvd is no different, that the tv will apply its own deinterlacing at playback time. You can prepare a deinterlaced - also known as progressive - source but that is going to ruin your footage.
The gurus on here will always rec that you do all filtering in avisynth. I am no guru so am afraid I can not help you there except to say that you should always take care with any noise reduction since it tends to soften the image.
To create a DVD in Vegas, simply put the DV video on the Vegas timeline, and then use the "DVD Architect" MPEG-2 template. You do NOT want the widescreen template, because VHS tape is 4:3, not 16:9. Then set the Average bitrate in the Vegas MPEG-2 encoder to whatever bitrate you need in order to fit the amount of video you are trying to put on a DVD. Use a bitrate calculator to determine this setting. Use variable bitrate, and check the "2-pass" option if you are going to use a bitrate below 6,500,000 bps. Make sure the MPEG-2 field order matches the field order of your original (bottom field first). The newer versions of Vegas now set the default to upper field first, and while this shouldn't make any difference, Vegas has a flaw where it actually reduces the render quality when it changes field order. So, it is important in Vegas for these to match.
I just tried CNR2 on your DV sample, and it didn't do a thing for the chroma issues. I also tried my usual MDegrain2 script and found that it destroyed a lot of detail. Therefore, if I were you, I'd be tempted to not do anything, and simply edit and enjoy. As I've said twice before, these is na uncommonly good VHS capture.
You have an ADVC 300 and rumour has it that it has several setting which enhance/improve the transfer/conversion process.
Problem is that Canopus do not give any help in their manual as to how one should set these. Their logic, methinks, is that a professional, who such a unit is aimed at, already knows what to do so does not need to be reminded. They did not appreciate that mere amateurs, myself included, would flex the cc and get hold of one.
At one time I read a most interesting article that fully explained every setting. Jeez. I wish now I had saved that.
So. To echo above. Let sleeping dogs lie.
OK .... understand what is being said and the logic behind it for creation of a DVD for TV playback.
However what about files being shared via internet sharing sites?........ would I be better converting to Progressive for these ? .... still use DV footage into Vegas, edit and then during render set to progressive ... or leave as interlaced.
Just thinking - as the test above proved using standard computer video player .. picture is poor unless deinterlaced .... (blended or mean rather than discard) do I therefore have to prepare files as progressive to prevent such issues ?
Or can I stay with interlaced as these sharing sites (Vimeo, youtube. facebook) process teh video anyway.
or am I being stupid again.
The picture is not poor. It is your own perception of it.
Create a dvd, as interlaced, as play it in a dvd player on a tv and then consider.
Internet playback is a different beast, So you do indeed compile as progressive. But do not let that influence your decision. Different floats for different boats.