This post is a follow-up to my recent attempt to optimize digitizing a collection of Hi8 home videos. See https://forum.videohelp.com/threads/376934-Requesting-Feedback-on-3-Hi8-Video-Capture-Samples. I decided to make a comparison between lossless and DV AVI capture. For lossless capture, I am using a Hauppauge USB-Live 2. For DV AVI, I have a Sony MiniDV (DCR-TRV38) being used in pass through mode and a Canopus ADVC100. Based on the recommendations made to responses from my previous post, I acquired a Hi8 camcorder with S-video out and line TBC (Sony CCD-TRV608). I am using VirtualDub for the capture software. For lossless, I am using the HuffyYUV codec. I have attached a short video clip making a comparison of lossless with TBC Off, lossless with TBC On, DV AVI MiniDV pass through and DV AVI Canopus ADVC100. The Sony CCD-TRV608 has a DNR setting which I left off. The TBC was also turned off for DV AVI capture.
Lossless TBC Off:
Lossless TBC On:
MiniDV pass through:
In my opinion, the lossless capture has the most detail and the best color, but I do not appreciate any difference with line TBC. In particular, I find the horizontal lines with motion much more visible with the lossless capture. Line TBC did not seem to improve this which I was most trying to accomplish. Any suggestions?
+ Reply to Thread
Results 1 to 21 of 21
Perhaps the CCD-TRV608 doesn't really have a line TBC. I note that the manual says the function of the TBC is "to correct jitter," implying that it's more of a frame stabilizer.
Not sure what to do now. I'm not one to easily call it quits, but I feel the need to start making some progress on this project.
I was thinking about trying out the professional service offered on this site. Maybe try seeing what conversion of one tape looks like before I commit to the whole batch. At least I now know a bit more about what to look for.
Having said that, the CCD-TRV608 is oversharpening. Its output here looks a lot less natural than your CCD-TRV32 captures. Knightmessenger in your other thread pointed out that your other camcorder is actually Video8, not Hi8. Were all your tapes recorded using this Video8 camcorder?
I have found that my Digital8 model oversharpens my Video8 recordings, too.
In particular, I find the horizontal lines with motion much more visible with the lossless capture. Line TBC did not seem to improve this which I was most trying to accomplish.
I am actually viewing my video on a Samsung plasma set but streaming through a HTPC I recently built and using a program called OpenELEC. I am just now starting to learn the settings on this program. I assumed it was automatically de-interlacing but discovered I was wrong. When I set the de-interlace setting from AUTO to ON, I immediately appreciated the big difference with TBC that you demonstrate in your captured pic. I have now fixed my interlace artefact problem and my line TBC seems to be working.
Yesterday I sat my family down and showed them extended clips of the three capture approaches I am analyzing (now properly de-interlaced). So far, everyone (including myself), picked the Sony MiniDV pass through capture approach. When I asked why, most felt that the lossless capture looked too "grainy". Exploring this at higher zoom, my lossless capture seems to definitely have more "noise". A lot of this "noise" seems to be due to rapidly oscillating vertical wavy lines. Although this is best appreciated in real time, I have tried to demonstrate it in the picture below.
By comparison, my Sony miniDV seems to be doing a pretty good job of suppressing a lot of this noise. Any suggestions on how I can improve on this with my lossless capture?
Last edited by VideoDoc; 25th Feb 2016 at 22:15.
It's a small gripe, but I would prefer if you didn't quote entire posts, particularly those that include images.
For what it's worth, the spatial field placement of the second Sony capture is "wrong". One line is cropped from the top, and a bottom black line has been added. To maintain bottom-field-first, the temporal placement has also been changed (each frame starts one field "earlier"). This inconsistency seems to be common across Sony's Digital8 line, and I guess it happens with their MiniDV products too.
most felt that the lossless capture looked too "grainy". Exploring this at higher zoom, my lossless capture seems to definitely have more "noise". A lot of this "noise" seems to be due to rapidly oscillating vertical wavy lines.
Any suggestions on how I can improve on this with my lossless capture?
AVISource("Sony CCD-TRV608 to Hauppauge TBC-ON DNR-OFF.avi") LanczosResize(400,height).LanczosResize(width,height) # Arbitrarily-chosen downsize width. A different resizer may be preferable.
You could also try lowering the Sharpness control of the USB-Live2. I have no idea whether it's at neutral by default. I wouldn't necessarily suggest using it to add blur, though. It probably isn't limited to the horizontal direction.
Last edited by Brad; 25th Feb 2016 at 22:37. Reason: Ninja edit to remove typo removed by VideoDoc's edit!
I'd like to see a capture from an actual Hi8 recording, played back using the CCD-TRV608. Just to see whether it oversharpens Hi8 material, too.
Noise present in the original recording, and noise caused by the playback system, is exacerbated by the sharpening.
It's at least partially due to the DV compression. High-frequency "details" (noise in this case) are the first to get discarded.
I thought I would take the opportunity to provide some closure on my Video8 transfer project. After spending a considerable amount of time exploring my three options for capture, I have decided to go with DV AVI via MiniDV pass-through. Despite my best efforts with lossless capture, I still feel that using my Sony Mini-DV in pass-through mode produces the best results. I also have the advantage of a work-flow that is relatively simple and an output which is relatively easy to manage. Let me be clear, I was willing to put up with extra work associated with lossless capture if I was able to produce better results. In my hands, that was not the case. With close critical inspection, I can see why lossless capture should produce a better result. Careful frame analysis clearly demonstrates better color and detail. Unfortunately, the overall gestalt of a video watching experience does not always parallel the conclusion of microscopic examination. I am certain that those with more experience and better equipment could have done better. I stress that I am certainly in no position to weigh in on the on-going battle of which is the better technique. As a matter of fact, I still plan to have one of my tapes professionally converted just to appease my academic interest. I would like to thank everyone that took the time to review what I submitted and generously provided their advice. Special thanks to Brad (vaporeon800) who provided 90% of that advice. I hope that someone will likewise find the information I provided useful.
It does seem to be a good option, provided you lower the black level when re-encoding to your viewing format.
In the playback menu for Hi8 camcorders, mine have 3 settings I would check for.
EDIT, TBC, DNR
So edit is used when dubbing to vhs, it is supposed to make the copy not have as much noticeable generation loss. By testing direct captures from a Hi8 camcorder, I have concluded that it looks better for direct captures as well with edit turned On.
TBC is time base corrector as you probably know. That probably makes sense to keep On.
DNR is digital noise reduction, as someone mentioned that the TRV-608 is over sharpening, maybe it's because this setting is turned on. I usually leave this off because I figure if I really need to do noise reduction, it would be better accomplished on a computer than a basic camcorder setting.
TBC is time base corrector as you probably know. That probably makes sense to keep On.
DNR is digital noise reduction, as someone mentioned that the TRV-608 is over sharpening, maybe it's because this setting is turned on. I usually leave this off because I figure if I really need to do noise reduction, it would be better accomplished on a computer than a basic camcorder setting
I would turn off the Hi8 cam's TBC when capturing via the MiniDV. The latter seems to have more powerful TBC.
The Hi8 TBC should make the MiniDV one ineffective. I dunno whether there would be any further interactions.
Capturing video can be so subjective, you can drive yourself mad picking which one seems best.
Most aspects are not subjective: chroma noise, improper color, tracking, etc. Only a tiny fraction is.
The biggest issue is experience. Having more makes it all easier. Having worked with video since the analog-only days, the main issue for me is always hardware failures.
So this is really interesting. Old thread and you are probably done your project by now but am attempting the same thing with my Sony TRV608.
So unfortunately when you share files through the internet it's a lossy transfer obviously on my end so you get noticeable pixelation. I usually get way better result if the capture was made in high definition as this is at 480 resolution so I have more to play with. However believe it or not, the video you are looking at is the Sony CCD-TRV608 to Hauppauge TBC-ON DNR-OFF. For some reason when I put the file into UniConverter and convert to an MP4, the lines completely disappear. I then changed the aspect ratio to standard 3:4, and brought up the saturation a lot, and fiddled around with the brightness and contrast until it looked good to my eyes but of course that is also subjective and that could be changed further in post process to your liking.
It would have looked a lot better if the original file was higher resolution as I upscaled it to 1080P for editing purposes which pixelates it if it's not the same resolution. So to make an accurate representation of what it could look like shrink the window size down really small. Also as you can tell the exposure on the original video was too dark so it's obviously a dark video. I didn't want to try and correct that because it starts to blow out the whites as you can see. Also ignore that screen shot is pixelated. Anyways just thought it was an interesting observation. My favorite is the raw s-video capture as I can visualize more detail. Honestly that is why I am a big believer in using the source that captures the most detail and not worrying about gamma, contrast, and ratio which can be altered later in post process.
Sony CCD-TRV608 to Hauppauge TBC-ON DNR-OFF Edited.mov
Last edited by BenKlesc; 15th Apr 2021 at 18:21.