Go Back    Forum > Digital Video > Video Project Help > Capture, Record, Transfer

Reply
 
LinkBack Thread Tools
  #1  
06-15-2019, 12:31 AM
Duder_me Duder_me is offline
Free Member
 
Join Date: Apr 2014
Posts: 15
Thanked 0 Times in 0 Posts
Observations:

-First of all, thank you so much for all the advice from both here and VideoHelp. I first tried doing this five years ago and gave up as soon as I started, but after getting myself together and returning with a new perspective, I think I have finally gotten the results I'm looking for. I am very surprised at the quality of the video I'm seeing so far. I'm very impressed. No dropped frames. I also somehow managed to fix this high-pitched audio problem I was having as well. My tapes are mostly PAL, which complicates things. I just made sure all my settings were PAL_B based and the audio was PCM 48000Hz, stereo, 16-bit. I did further research and it turns out many other people have had the same problem. Some suggested to make sure their PC was set to High Performance, so that might also make a difference.

Questions:

-When I deinterlace and double the frame rate of the video after capture, should I use the deinterlace filter with the Yadif deinterlacing algorithm or the bob algorithm? I've read in passing that people use bob, but after reading through more threads, I'm stumped as to what the best method could possibly be. I learned how to deinterlace from this video using the Yadif algorithm, but Huffyuv isn't used for compression. I used Huffyuv, don't worry.

-Also, I need to compress the videos down to where they still look great and can fit on a 15GB cloud drive to send to relatives. I was thinking of buying a small external drive and mailing it over instead, but convenience is also a factor that I need to consider. I'm definitely not going to send them 100GB lossless files, that's for sure. The video I linked above also has a Handbrake tutorial at the end for compressing the video.

-A few of the tapes I have were apparently recorded on a NTSC Sony Handycam from 1997-1999 that may or may not have PAL capability, which I am doubtful of. Some are even labeled as NTSC. I read that you can tell if a tape is NTSC on a PAL camcorder by a thin strip of washed out color on the right side of the video, which I noticed in one of the tapes I captured so far. That's hidden in overscan, so it's normal. The audio is also normal. I haven't deinterlaced yet to smoothen the frame rate. I already captured these tapes using NTSC Playback: On PAL TV on the camcorder, PAL_B, 25 FPS, 720x576. Should I capture these tapes again by changing NTSC Playback to ON NTSC 4.43 on the camcorder and making all the settings on VirtualDub NTSC_M, 29.97 FPS for these tapes?

My Setup:

PC Specs: 2012-2013 Asus R500A-BB71-CB with Windows 7 Home Premium, 2.3 GHz Intel Core i7 3610QM, 8 GB DDR3, 500 GB Samsung 860 EVO SSD, and Intel HD Graphics 4000.

Camcorder: 1998 Sony Hi8 XR CCD-TR840E PAL, TBC on, DNR off, NTSC Playback: On PAL TV, S-Video, Stereo Composite Audio.

Tapes: Vast majority recorded on a 1998 Sony Video8 XR CCD-TRV45E (PAL) and I just learned that a few were recorded on a 1997-1998 Sony NTSC Handycam.

Capture Device: ATI TV Wonder HD 600 USB, Drivers.

Capture Software: VirtualDub 1.9.11 + Filters, HuffYUV-multithreaded.

VirtualDub Capture Settings:

-Video Capture Pin: Video Standard: PAL_B, Frame Rate: 25, Output Size: 720x576
-Video Capture Filter: Video Standard: PAL_B, Video Proc Amp: Sharpness 3
-Video Compression: Huffyuv v2.1.1 - Multi-Threading Patch v.1.0
-Set Custom Format: 720x576 YUY2 YUV 4:2:2 interleaved
-Audio Raw Capture Format: PCM 48000Hz, stereo, 16-bit
-Audio Compression: <No compression (PCM)>
-Capture Settings: Frame rate: 25
-Disk I/O: Chunk size 12MB, Chunks in buffer 24, Unchecked Disable Windows write buffering

VirtualDub Post Settings:

-Video Filters: deinterlace, Interpolate using Yadif algorithm, Double frame rate, top field first
-Video Compression: Huffyuv v2.1.1 - Multi-Threading Patch v.1.0
-Save as AVI
Reply With Quote
Someday, 12:01 PM
admin's Avatar
Ads / Sponsors
 
Join Date: ∞
Posts: 42
Thanks: ∞
Thanked 42 Times in 42 Posts
  #2  
06-16-2019, 01:04 PM
cbehr91 cbehr91 is offline
Free Member
 
Join Date: Sep 2013
Posts: 134
Thanked 20 Times in 18 Posts
Are you planning on streaming these on the web (YouTube, Vimeo, etc.)? If not, then deinterlacing isn't necessary. If yes, the best deinterlacer is Avisynth's QTGMC. I'm deferring to the others about video standard conversion because I am in NTSC land only dealing with NTSC material.

Without seeing some samples it's a good bet your videos are going to need some post-processing like noise reduction and temporal smoothing (every Hi8 tape I've ever digitized has needed both), which you can do in VirtualDub. Also not sure how good the TBC is on the camcorder. Again, need some samples.

To save cloud space encoding to H.264 is fine, but definitely keep the original tapes and the lossless captures (hard drives are cheap). Encoding to MPEG-2 using TMPGenc is another option. I am not the biggest fan of Handbrake. Again, deferring to others here. I use Adobe Media Encoder, but that's only because I only ever encode H.264 for web upload, plus I pay for the Adobe suite so I might as well use it!
Reply With Quote
The following users thank cbehr91 for this useful post: Duder_me (06-16-2019)
  #3  
06-17-2019, 02:19 AM
Duder_me Duder_me is offline
Free Member
 
Join Date: Apr 2014
Posts: 15
Thanked 0 Times in 0 Posts
Thanks! I live in NTSC land too , it's just that my father bought our camcorder in a PAL region and it stuck with us until we got a NTSC miniDV camcorder. I don't plan on having them streamed, but I assume the videos will mostly be played back on smartphones and computers with clips being shared on WhatsApp. I also plan on putting them on an iPad. That's still no excuse for skimping on quality. I still prefer big screens though.

I found this YouTube tutorial to deinterlace with with AVISynth+, QTGMC, and FFMPEG. I haven't sat through it yet, but if there's a better method, I'd like to know. Once I'm done with all my tapes, I'll take the the time to deinterlace the ones I'm going to send and deinterlace the others later due to lack of drive space (I sent my 8TB external HDD for a VHS capture job). I'm still going to keep the original .avi files.

I also just noticed that I had DNR set to on. It must have reset after I removed the battery or turned the camcorder off before actually capturing. I've read about color bleeding with this setting, but I didn't really notice anything wrong with the picture. It actually looked better than I thought it would, kind of like miniDV in a way. I haven't gotten too far yet in capturing my important tapes, so should I turn DNR off now and recapture them?

I'll upload an .avi sample tomorrow evening. I think it's Universal Studios and NASA in Florida from 1998. Cool stuff like the Blues Brothers, the Terminator, King Kong, etc. I might put the whole thing on YouTube later.
Reply With Quote
  #4  
06-17-2019, 01:37 PM
traal traal is offline
Free Member
 
Join Date: Jan 2016
Posts: 393
Thanked 74 Times in 67 Posts
Quote:
Originally Posted by cbehr91 View Post
Are you planning on streaming these on the web (YouTube, Vimeo, etc.)? If not, then deinterlacing isn't necessary.
I find deinterlacing necessary even when playing the file locally, otherwise I get motion artifacts and my video players play at 30 frames per second instead of 60 fields per second (NTSC).
Reply With Quote
  #5  
06-17-2019, 02:00 PM
hodgey hodgey is offline
Free Member
 
Join Date: Dec 2017
Location: Norway
Posts: 1,680
Thanked 447 Times in 384 Posts
Quote:
I already captured these tapes using NTSC Playback: On PAL TV on the camcorder, PAL_B, 25 FPS, 720x576. Should I capture these tapes again by changing NTSC Playback to ON NTSC 4.43 on the camcorder and making all the settings on VirtualDub NTSC_M, 29.97 FPS for these tapes?
NTSC on PAL tv setting means PAL-60, i.e 29.97 FPS, 480 lines, so if you captured them with virtualdub set to normal pal maybe the actually were PAL too? The setting only matters for NTSC tapes. On the cameras with LCDs you can see the resolution on the LCD change if playing a NTSC tape, don't remember how it looks in the viewfinder.

I used NTSC-4.43 last time I captured NTSC tapes from our PAL cams, but I didn't really notice a difference between that and NTSC on PAL TV so I don't know if there will be any difference if the capture device supports both. In either case it looked better than using an older native NTSC video8 player. Both format are a variant of NTSC frame rate and line count, but with the colour stored the frequency used in PAL. The difference is whether the colour is in PAL or NTSC format.

Note that the TBC and DNR is not active when playing NTSC tapes, so if those tapes are too wiggly and unstable, passing it through a TBC/DVD recorder or using a NTSC cam with TBC may give a nicer result.

The DNR function is very mild on these cameras, You can easily test yourself with the camera hooked up and switching it off and on to see the differences.
Reply With Quote
  #6  
06-17-2019, 07:53 PM
Duder_me Duder_me is offline
Free Member
 
Join Date: Apr 2014
Posts: 15
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by hodgey View Post
Note that the TBC and DNR is not active when playing NTSC tapes, so if those tapes are too wiggly and unstable, passing it through a TBC/DVD recorder or using a NTSC cam with TBC may give a nicer result.

The DNR function is very mild on these cameras, You can easily test yourself with the camera hooked up and switching it off and on to see the differences.
So I have to get a NTSC Sony camcorder as well if I want TBC applied to NTSC tapes? That's unfortunate. I also don't know where to start when it comes to TBCs or DVD recorders, especially because some are very expensive, so it would be easier for me if I just got another $100 camcorder. At least those are easier to come by and are twice as cheaper as the PAL models. The problem is the headache that comes with choosing a model. Thankfully, I don't have a lot of NTSC tapes compared to the PAL ones.

As for the tapes that I thought were NTSC, they were actually PAL. It was immediately obvious by the cropped image and high pitched audio using NTSC settings. I still have a ways to go until I get to the NTSC ones.

-- merged --

Scratch everything I said above about NTSC tapes. I just learned that the first camcorder was also a PAL/multi-system one. I'm good to go! I just need to make sure I deinterlace and double the frame rate just right. I think the YouTube tutorial I posted above will suffice for now, but I'm still open to more advice.

One other question. How do I trim/edit a raw AVI capture in VirtualDub without losing any quality? It should seem self explanatory, but to be honest the software makes it too vague to be sure. I let a tape run too long and now it has an extra hour and a half of blue screen on it.

-- merged --

Quote:
Originally Posted by cbehr91 View Post
Are you planning on streaming these on the web (YouTube, Vimeo, etc.)? If not, then deinterlacing isn't necessary. If yes, the best deinterlacer is Avisynth's QTGMC. I'm deferring to the others about video standard conversion because I am in NTSC land only dealing with NTSC material.

Without seeing some samples it's a good bet your videos are going to need some post-processing like noise reduction and temporal smoothing (every Hi8 tape I've ever digitized has needed both), which you can do in VirtualDub. Also not sure how good the TBC is on the camcorder. Again, need some samples.

To save cloud space encoding to H.264 is fine, but definitely keep the original tapes and the lossless captures (hard drives are cheap). Encoding to MPEG-2 using TMPGenc is another option. I am not the biggest fan of Handbrake. Again, deferring to others here. I use Adobe Media Encoder, but that's only because I only ever encode H.264 for web upload, plus I pay for the Adobe suite so I might as well use it!
Here is the sample I was talking about. I used Direct stream copy.

Sony CCD-TR840E (PAL)
-TBC ON
-DNR ON
-NTSC PB: ON PAL TV
-S-Video, Stereo
-Sony 8mm Video Cassette MP120 NTSC8

VirtualDub 1.9.11 + Filters, HuffYUV-multithreaded, Settings left as is except for the following

-Video Capture Pin: Video Standard: PAL_B, Frame Rate: 25, Output Size: 720x576
-Video Capture Filter: Video Standard: PAL_B, Video Proc Amp: Sharpness 3
-Video Compression: Huffyuv v2.1.1 - Multi-Threading Patch v.1.0
-Set Custom Format: 720x576 YUY2 YUV 4:2:2 interleaved
-Audio Raw Capture Format: PCM 48000Hz, stereo, 16-bit
-Audio Compression: <No compression (PCM)>
-Capture Settings: Frame rate: 25
-Disk I/O: Chunk size 12MB, Chunks in buffer 24, Unchecked Disable Windows write buffering


Attached Files
File Type: avi CCD TR840E 8MM Sample.avi (95.99 MB, 24 downloads)
Reply With Quote
  #7  
06-20-2019, 06:11 PM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
Thanks for the sample. Very good color and detail. But poor signal level controls (in fact, no signal level control at all).

Signal levels are illegal with brights and chroma that even exceed RGB 255 -- so there is high end clipping and detail loss, and obvious blooming hot-spots that look like neon. Darks are clipped at y=16 by the capture device, so darker shadow areas are dense and blocked up. The sample has vertical herringbone noise, discussed below. There's also slight chroma bleeding and some ghosting. Most of the time the motion is so violent that it's easy to miss some of this. In slower scenes it will be more obvious, especially after final encoding.

But if most of the video looks as frantic as the sample, viewers will quickly tire of watching anyway. In any case, the motion causes excessive combing that won't go away even with Avisynth cleanup. It's a common headache with much of consumer video, but there are ways to work around it.

Quote:
Originally Posted by Duder_me View Post
-When I deinterlace and double the frame rate of the video after capture, should I use the deinterlace filter with the Yadif deinterlacing algorithm or the bob algorithm?
Neither. Both are inferior quality, especially with the video sampled here. See notes below and images.

Quote:
Originally Posted by Duder_me View Post
I'm definitely not going to send them 100GB lossless files, that's for sure. The video I linked above also has a Handbrake tutorial at the end for compressing the video.
No, don't send lossless. They won't be able to play them anyway.
Why do newbies always head for severely limited apps like Handbrake? Whatever......

Quote:
Originally Posted by Duder_me View Post
Video Compression: Huffyuv v2.1.1 - Multi-Threading Patch v.1.0
HuffYUV-MT has been obsolete since 2004 and Pentium-5. It has no performance advantage today. Newer version 2.1.1 is more efficient. Today's CPU's are multicore. Multithreading and multicore are not the same thing.

Your sample as HuffYUV-MT YUY2: 98.3 mb
Your sample as Lagarith YUY2: 73.5 mb

Quote:
Originally Posted by Duder_me View Post
I found this YouTube tutorial to deinterlace with with AVISynth+, QTGMC, and FFMPEG. I haven't sat through it yet, but if there's a better method, I'd like to know.
Some very useful older filters won't work with Avisynth+ and FFMPEG is rather touchy and specialized (some of its encoding isn't so great). Everything seems to work with ol' reliable Avisynth v2.6 August/2013. You can try whatever you want, but for your noisy sample I'd say QTGMC is your optimum choice (as I'll show, below).

Quote:
Originally Posted by Duder_me View Post
I also just noticed that I had DNR set to on. It must have reset after I removed the battery or turned the camcorder off before actually capturing. I've read about color bleeding with this setting, but I didn't really notice anything wrong with the picture. It actually looked better than I thought it would, kind of like miniDV in a way. I haven't gotten too far yet in capturing my important tapes, so should I turn DNR off now and recapture them?
Be glad it doesn't look like DV, which has an etched-in-plastic look with unreal color, blown-out highlights, and easily visible compression artifacts. Builtin DNR isn't always what it's cracked up to be. There are better denoisers in post-processing, but try a short dnr test and see what happens. Your sample does have chroma noise and some bleeding, and if the player's DNR was on it did nothing for the herringbone.

The image below is a 720p enlargement of part of a frame, showing herringbone noise in the girl's arm, in the faces behind her, and in the red dress at lower right. This is frame 3 deinterlaced with QTGMC, which has removed much of the residual noise but revealed the underlying herringbone.


Below, yadif leaves aliasing and strong noisy grain plus chroma noise, with the herringbone noise lying underneath. Yadif lacks the denoising ability and overall clarifying of QTGMC.


The same frame after deinterlacing and then filtering. The frame shown was deinterlaced with QTGMC and cleaned further with low-pass resiZing and MCTemporalDenoise at "medium" (although "Low" would probably have been good enough).


Chroma noise and ghosting, images cropped from two different frames (girl's neck and chin, back of guy's neck, etc.).


You can deinterlace and denoise using Avisynth, then save the deinterlaced 720x576 50p output file as YV12 using the Lagarith lossless codec. From that file you can take several paths to formatting and encoding for several formats, including square-pixel for streaming to relatives or Facebook, YouTube posting, or DVD.

Code:
AviSource("J:\forum\faq\duder_me\CCD TR840E 8MM Sample.avi")
###--- Use levels() to get legal video levels y=16-235 ---###
Levels(16,1.0,255,16,235,dither=true,coring=false)
AssumeTFF()
ConvertToYV12(interlaced=false,ChromaOutPlacement="MPEG2")
QTGMC(preset="medium",EZDenoise=4,denoiser="dfttest",ChromaMotion=true,\
   border=true,ChromaNoise=true,DenoiseMC=true,GrainRestore=0.3)
vInverse2()
###--- next four statements for low-pass, clean and smooth herringbone ---###
BicubicResize(360,height).SPline36Resize(720,height)
MCTemporalDenoise(settings="medium")
Dither_convert_8_to_16 ()
GradFun3(thr=0.5,mask=0,lsb_in=true,lsb=false,smode=0)

###--- Even-up clean borders and restore 720x576 frame size ---###
Crop(6,8,-10,-12).AddBorders(8,10,8,10)
return last

###### In VirtualDub save as 720x576 YV12 for futher formatting/encoding ######
When I ran the above script in VirtualDub I loaded the VDub ColorCamcorderDenoise filter to tweak the green chroma noise in the right-hand border, and applied it to the script's output. I configured VirtualDub to save the output file as YV12 using the Lagarith codec. I named the file "8MMsample_720x576_50p.avi".

There is a lot you can do with that intermediate saved work file. The first thing I did was to run an Avisynth script that resized the video to square-pixel 4x3/50p for smart TV and streaming, and for web posting. Other square-pixel frame sizes are easily possible:

Code:
AviSource("J:\forum\faq\duder_me\8MMSample_720x576_50p.avi")
ConvertToYUY2(interlaced=false,ChromaInPlacement="MPEG2")
Spline36resize(768,height)
ConvertToYV12(interlaced=false)
return last
I saved the output of that script as Lagarith/YV12 for my h.264 encoder and used a fairly high bitrate of 6000 VBR to encode the video as "8MMsample_4x3_50p.mp4". You'll need a fairly high bitrate to cleanly render all that violent motion. "8MMsample_4x3_50p.mp4" is attached.

The next thing I did with that 720x576/50p work file was to convert it for posting to YouTube. YouTube wants a 4x3 frame inside a 16x9 pillarboxed file as the ideal size (if you don't apply your own pillarbox, YouTube will give you one with some sloppy resizing). The 720x576 video is resized to 4x3 960x720, pillarbox borders are added for the required 16x9 frame, and the result is encoded with 44.1KHz audio for YouTube at a decent bitrate (6500 VBR).

Code:
AviSource("J:\forum\faq\duder_me\8MMSample_720x576_50p.avi")
ConvertToYUY2(interlaced=false,ChromaInPlacement="MPEG2")
nnedi3_rpow2(2,cshift="Spline36Resize",fwidth=960,fheight=720,mpeg2=true)
AddBorders(160,0,160,0)
ColorMatrix(mode="Rec.601->Rec.709")
ConvertToYV12(interlaced=false,matrix="Rec709")
The result of that script was encoded with h.264 and is attached as "8MMsample_720p_for_YouTube.mp4".

Finally, I set up the 720x576 file for DVD. of course you can't make a 50fps DVD. But with progressive video you can discard alternate frames to restore the frame rate to 25fps, and then encode the result with fake interlace flags. Many set top DVD players will play it as interlaced anyway because that's what they're programmed to do. With 50% of the temporal resolution discarded the motion might be a little choppy in comparison, but I doubt most viewers would notice. It's better than the annoying excess combing.

Code:
AviSource("J:\forum\faq\duder_me\8MMSample_720x576_50p.avi")
SelectEven()
The output from that script was saved as Lagarith YV12. It was then configured for PAL 4:3 DVD spec and a fairly high motion-rendering target bitrate of 7000 VBR/9000 max with Dolby AC3 audio. The result is attached as "8MMsample_25p_DVD_fake_interlace.mpg".

All of the above are suggestions, but they demonstrate the possibilities. Overall, a pretty good job for initial captures.


Attached Images
File Type: jpg qtgmc herringbone original.jpg (114.6 KB, 175 downloads)
File Type: jpg yadif mixed noise original.jpg (132.2 KB, 171 downloads)
File Type: jpg noise filtered.jpg (94.9 KB, 174 downloads)
File Type: jpg chroma noise and ghosting.jpg (141.1 KB, 176 downloads)
Attached Files
File Type: mp4 8MMsample_4x3_50p.mp4 (7.58 MB, 10 downloads)
File Type: mp4 8MMsample_720p_for_YouTube.mp4 (7.88 MB, 9 downloads)
File Type: mpg 8MMsample_25p_DVD_fake_interlace.mpg (8.06 MB, 7 downloads)
Reply With Quote
The following users thank sanlyn for this useful post: captainvic (06-21-2019), Duder_me (06-24-2019), Murf (11-16-2020)
  #8  
06-24-2019, 07:44 PM
Duder_me Duder_me is offline
Free Member
 
Join Date: Apr 2014
Posts: 15
Thanked 0 Times in 0 Posts
Wow! Thank you so much! This is all new to me and it's a little overwhelming to be honest. Just when I think I'm done, something else comes up. I'm not a professional, so I don't really understand video correction too well. I've never used AVISynth before. I think I'll start with using those scripts you provided for my original sample and once I get the hang of it, I'll do it for real with the other captures. As for Handbrake, I never liked using it, especially for DVD rips. I could never figure out how to match the quality of a VOB file.

I made the sample from one tape by splicing together clips with the most variety in VirtualDub. I had to make sure it was just under the file size limit, so that's why it's a lot more frantic than it really is.

I'm more of a video tutorial, step by step guy, especially because this is all new to me and I want to do everything exactly right without spending more time than necessary. However, I understand that every scenario is different and needs different solutions, and it's beneficial to take the time to learn how all this works.

I think some of the blown out colors are due to the tinkering with the exposure button on the camcorder during the original recording. Many of my tapes have moments where the colors are really blown out one moment and then the exposure is instantly adjusted with the press of a button to make it look fine.

Quote:
Originally Posted by sanlyn View Post
HuffYUV-MT has been obsolete since 2004 and Pentium-5. It has no performance advantage today. Newer version 2.1.1 is more efficient. Today's CPU's are multicore. Multithreading and multicore are not the same thing.

Your sample as HuffYUV-MT YUY2: 98.3 mb
Your sample as Lagarith YUY2: 73.5 mb
Thanks for clarifying that. HuffYUV seems to be the recommended compression filter around here for 8mm/Hi-8 capture, and now that I have a lot more drive space to work with, file size is no longer a big issue. Is Lagarith something I should use from here on out? It's a fork of HuffYUV, but I would prefer to stick with the original if that's what's recommended.

Quote:
Originally Posted by sanlyn View Post
Be glad it doesn't look like DV, which has an etched-in-plastic look with unreal color, blown-out highlights, and easily visible compression artifacts. Builtin DNR isn't always what it's cracked up to be. There are better denoisers in post-processing, but try a short dnr test and see what happens. Your sample does have chroma noise and some bleeding, and if the player's DNR was on it did nothing for the herringbone.
I also have miniDV tapes to capture too, but I'll worry about that later. Would it be beneficial to capture my tapes again with DNR off? I was going to do that originally, but once I realized it was on in the settings, I didn't really notice any problems with the video, so I left it that way. I have 17 tapes. Some vary in length, but most are over an hour.

Quote:
Originally Posted by sanlyn View Post
The image below is a 720p enlargement of part of a frame, showing herringbone noise in the girl's arm, in the faces behind her, and in the red dress at lower right. This is frame 3 deinterlaced with QTGMC, which has removed much of the residual noise but revealed the underlying herringbone.

Below, yadif leaves aliasing and strong noisy grain plus chroma noise, with the herringbone noise lying underneath. Yadif lacks the denoising ability and overall clarifying of QTGMC.

The same frame after deinterlacing and then filtering. The frame shown was deinterlaced with QTGMC and cleaned further with low-pass resiZing and MCTemporalDenoise at "medium" (although "Low" would probably have been good enough).
Thanks for the comparison. I've been switching back and forth between them and while at first I thought the first one was superior, the third one wins out in the end. I hope the third method doesn't lose too much detail. Otherwise, I'd rather stick with the first one or something that is the best of both worlds.

Quote:
Originally Posted by sanlyn View Post
You can deinterlace and denoise using Avisynth, then save the deinterlaced 720x576 50p output file as YV12 using the Lagarith lossless codec. From that file you can take several paths to formatting and encoding for several formats, including square-pixel for streaming to relatives or Facebook, YouTube posting, or DVD.
Is there a specific installation method I need to follow for Avisynth or do I just download Avisynth v2.6 August/2013? I'm having trouble finding the specific version you recommended. Could you share a link? Do I need to download anything extra as well? Aside from the version of VirtualDub I downloaded and HuffYUV, my PC is pretty vanilla at the moment.

Quote:
Originally Posted by sanlyn View Post
All of the above are suggestions, but they demonstrate the possibilities. Overall, a pretty good job for initial captures.
Thanks for sharing those! I'm sure they'll be useful to others as well. I'm not planning on streaming at the moment, but instead I plan on hosting them on something like Google Drive (I might need multiple accounts), then the videos can be downloaded from there. I will also put them on a 128GB iPad Air 2.
Reply With Quote
  #9  
06-25-2019, 07:43 AM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
Herringbone noise is an obvious defect that's difficult to eliminate. It doesn't appear in normal video. It will look worse in calmer scenes without the frantic motion and in the final encodes.

Avisynth 2.6 is the most trouble-free version, especially for beginners. Geeks like to try all the new stuff, but at this stage you really don't need to contend with updates to updates to revisions of updates, etc., etc. The August version I mentioned has been taken down from the listings for some reason, but the May 2015 version is still a standby at https://www.videohelp.com/download/AviSynth_260.exe.

Avisynth is a simple install. All it does is create the Avisynth program folder, tells the Windows registry where it's located, copies a couple of basic builtin utilities into the plugins folder, and copies Avisynth.dll to the Windows system folder. When the installer asks questions just take the defaults, although you would want to check the box that associates .avs script files with Notepad so that when you double-click an .avs file it automatically opens in Notepad. An Avisynth script is just a plain text file with an .avs file ending.

The usual default install location for Avisynth is in "Program Files" in 32-bit Windows or in "Program Files (x86)" in 64-bit Windows. If you download any plugins or add-ins later, don't download them directly into your Avisynth plugins folder. Most plugins come as .zip files or other packages that contain multiple files and documentation -- all that stuff doesn't belong in your plugins folder and will soon create a chaotic mess. First, create a master folder for all Avisynth and VirtualDub plugins. Call the folder "Filters" or any name you want. Then for each new filter create a subfolder with that filter's name. Download the package into that subfolder and unzip it if required. Then just copy only the filter itself into the app's plugins folder. Plugins in Avisynth belong in "Avisynth/plugins" and in Virtualdub in "Virtualdub/plugins".

VirtualDub plugins have a .vdf file ending. Avisynth plugins come in three file endings: .dll, .avsi, and .avs. A dll or avsi file loads automatically when Avisynth calls for it. An .avs has to be imported explicitly using a builtin Import() function, which simply copies the text of the .avs filter into your own script at runtime. There are many reasons why some filters are released as .avs files rather than as other types, but trust the designer and use what's given. Usually an .avs allows you to keep umpteen different versions of similar functions but without confusing Avisynth internally.

One thing to prepare for is the data compression formats for all these filters. Usually it's .zip for PKZip, which is common enough to decompress. Rar for WinRAR is sometimes used. PkZip and RAR have dozens of free apps. One compressor that the geeks really like is 7Zip (.7z). Fortunatelty 7zip has a handy free interface at https://www.7-zip.org/.

VirtualDub has an executable base program file called VirtualDub.exe. What people find confusing about Avisynth is that it doesn't have an ".exe". Never fear. You run Avisynth with an app that can act as a file server client -- in this case, VirtualDub is the favorite client and pseudo-GUI. When you tell VirtualDub to open a video you can point it to an .avs file, because VirtualDub sees an .avs file as a video file. You can also apply VirtualDub filters to Avisynth's output at the same time, if you want.

In the workflow, capture, and restoration areas of the forum there are many threads that go into tons of detail about scripts and plugins with a lot of download links and references. Browsing such discussions is the way most of us learned about Avisynth and VirtualDub. There is some older stuff on Youtube but you have to take YouTube with a grain of salt; much of it isn't really so correct, and some of it is just nutty. The favorite resource docs are in the Avisynth wiki pages which you can access in Google by entering "Avisynth" followed by a filter name. For example, if you enter "Avisynth LSFmod" you'll get the wiki page at http://avisynth.nl/index.php/LSFmod. It contains a download link for the basic filter as "LSFmod v1.9 avsi". It also lists additional support plugins required, and then gets into all the subfunctions and parameters you can play with. It's not necessary to understand 100% of all this cryptic minutiae, but it does give you a general idea of what's going on. Most of us execute a sharpening filter like LSFmod with the bald statement "LSFMod()" and take the defaults. You learn such fine points by seeing how others use those filters in scripts and in Virtualdub.

A major source for current Avisynth external filters is http://avisynth.nl/index.php/External_filters, where dozens of plugins are listed by category. For VirtualDub plugins there is http://www.infognition.com/VirtualDubFilters/, although the Avisynth list is more up to date than some of the old links at infognition. There are other archives sprinkled all over the internet.

Then there are innumerable links in digitalfaq threads. Some of these get a little old, so be mindful that some things change over time. For example there are download links to two versions of Virtualdub's popular CCD filter (ColorCamcorderDenoise). Fortunatelty in this case the .vdf names are different, so you can have both versions in your VDub plugins at the same time.

Occasionally forum corrections are necessary. For example one of the earlier links for the entire QTGMC plugins package and support files accidentally omitted one of the requirements. But it was corrected with a newer link at http://www.digitalfaq.com/forum/atta...kagenov2017zip in post #6 of the same thread. In fact that thread can serve as an intro to plugins and scripting. The thread is http://www.digitalfaq.com/forum/vide...html#post57994. Script details and pictures are in post #3, with a bunch of download links in post #2, which happens to contain the the erroneous QTGMC link that was corrected in post #6.

There is also a helpful fix-it link for Windows7 and later for possible trouble with popular plugins such as RemoveDirtMC, http://www.digitalfaq.com/forum/vide...s-running.html.

Then there are tons of scripts, details, links, and images for working with some troublesome captures on page 3 at http://www.digitalfaq.com/forum/vide...per-vhs-3.html and on page 4 of the same thread at http://www.digitalfaq.com/forum/vide...per-vhs-4.html.

Much earlier threads sometimes have a few outdated links but they're still useful and informative. Such a thread is one that demonstrates and illustrates everything from initial script and analysis functions to using popular VirtualDub image controls that mimic those in expensive apps. The details begin with "Information Overload #1" in post #16 and continue in subsequent posts of that thread with Information Overload #2, #3, #4, and #5.

The "main page" for Avisynth isn't exactly that easy to find for some reason, but it's at http://avisynth.nl/index.php/Main_Page. An older user guide might be more helpful at doom9: http://foro.doom9.org/index.html?/ca..._avisynth.html. There isn't a new "Main page" for VirtualDub that would be very helpful but there's the basic Doom9 guide at http://foro.doom9.org/index.html?/ca...sing_vdub.html.
Reply With Quote
The following users thank sanlyn for this useful post: captainvic (06-25-2019), ehbowen (06-25-2019)
  #10  
06-25-2019, 10:41 AM
hodgey hodgey is offline
Free Member
 
Join Date: Dec 2017
Location: Norway
Posts: 1,680
Thanked 447 Times in 384 Posts
Quote:
Your sample does have chroma noise and some bleeding, and if the player's DNR was on it did nothing for the herringbone.
That herringbone noise may have been caused by something in the capture chain. It looks a bit too high-frequency for a video8 tape. Could be a bad cable or something else.
Reply With Quote
  #11  
06-25-2019, 12:00 PM
Duder_me Duder_me is offline
Free Member
 
Join Date: Apr 2014
Posts: 15
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by hodgey View Post
That herringbone noise may have been caused by something in the capture chain. It looks a bit too high-frequency for a video8 tape. Could be a bad cable or something else.
Could it be the S-Video cable? The cable is really long, but it came with the camcorder I ordered, so I thought it was more convenient to use that one.
Reply With Quote
  #12  
06-25-2019, 11:10 PM
captainvic captainvic is offline
Premium Member
 
Join Date: Sep 2018
Posts: 47
Thanked 6 Times in 5 Posts
sanlyn, thanks for clarifying which versions of HuffYUV and Avisynth to use. I can't seem to find which version of Lagarith is "best" as I search around DFAQ. Do you recommend Lagarith version 1.3.27 available here?

https://www.videohelp.com/software/Lagarith-Lossless-Video-Codec
Reply With Quote
  #13  
06-26-2019, 08:55 AM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
On most download sites only the latest 1.3.27 is available. Lagarith's home page has a handy installer: https://lags.leetcode.net/codec.html
Reply With Quote
The following users thank sanlyn for this useful post: captainvic (06-26-2019)
  #14  
06-26-2019, 09:04 AM
captainvic captainvic is offline
Premium Member
 
Join Date: Sep 2018
Posts: 47
Thanked 6 Times in 5 Posts
Quote:
Originally Posted by sanlyn View Post
On most download sites only the latest 1.3.27 is available. Lagarith's home page has a handy installer: https://lags.leetcode.net/codec.html
Thanks. To clarify, that VH link I posted does have older versions of Lagarith. Is there any advantage to using an older version, or should I stick with 1.3.27?

I almost made the mistake of grabbing VDub 1.10.x but thanks to this site realized the older version 1.9.11 was the one to get.
Reply With Quote
  #15  
06-26-2019, 09:18 AM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
The version of Lagarith on Lagarith's page link is the latest. Earlier versions are supposed to be Ok with early Windows such as W2000 and W98, and some older ones are manual install only. The new installer will work on 32-bit and 64-bit Windows. If you only have 32- bit, then the installer only installs 32-bit. If you have 64-bit, the installer will install both versions to work with 32-bit and 64-bit apps. I last used that installer in 2017 on a Win10 laptop.
Reply With Quote
The following users thank sanlyn for this useful post: captainvic (06-26-2019)
  #16  
06-26-2019, 10:07 AM
lordsmurf's Avatar
lordsmurf lordsmurf is online now
Site Staff | Video
 
Join Date: Dec 2002
Posts: 13,508
Thanked 2,449 Times in 2,081 Posts
Quote:
Originally Posted by sanlyn View Post
The new installer will work on 32-bit and 64-bit Windows. If you only have 32- bit, then the installer only installs 32-bit. If you have 64-bit, the installer will install both versions to work with 32-bit and 64-bit apps. I last used that installer in 2017 on a Win10 laptop.
Do this.

I've been setting up new systems recently, both for myself and some members here, and this is what I used.

- Did my advice help you? Then become a Premium Member and support this site.
- For sale in the marketplace: TBCs, workflows, capture cards, VCRs
Reply With Quote
The following users thank lordsmurf for this useful post: captainvic (06-26-2019), Duder_me (06-26-2019)
  #17  
06-26-2019, 02:11 PM
Duder_me Duder_me is offline
Free Member
 
Join Date: Apr 2014
Posts: 15
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by lordsmurf View Post
Do this.

I've been setting up new systems recently, both for myself and some members here, and this is what I used.
Thank you for confirming that you use Lagarith as well. It looks like I need to update my software workflow and redo my captures. My impatient side is telling me that as long as the video looks like how it's played on a late 90s Sony Trinitron, I'm satisfied. However, I'll hold out a bit longer to fix my herringbone issue, clean up the video with Avisynth, and hopefully wrap up this project by July.
Reply With Quote
  #18  
06-26-2019, 03:14 PM
sanlyn sanlyn is offline
Premium Member
 
Join Date: Aug 2009
Location: N. Carolina and NY, USA
Posts: 3,648
Thanked 1,307 Times in 982 Posts
Perfection isn't possible (I keep telling myself so), but I'm glad you see that a major defect should be addressed as well as you can manage. Otherwise you should have a nice looking video. I've gone through quite a lot of video more than once myself.


I wish we could determine where the problem originates. If you have access to another player, or if other tapes have the same problem, that could help provide an answer.
Reply With Quote
  #19  
06-26-2019, 10:22 PM
Duder_me Duder_me is offline
Free Member
 
Join Date: Apr 2014
Posts: 15
Thanked 0 Times in 0 Posts
Quote:
Originally Posted by sanlyn View Post
Perfection isn't possible (I keep telling myself so), but I'm glad you see that a major defect should be addressed as well as you can manage. Otherwise you should have a nice looking video. I've gone through quite a lot of video more than once myself.


I wish we could determine where the problem originates. If you have access to another player, or if other tapes have the same problem, that could help provide an answer.
The tape was a Sony 8mm Video Cassette MP120 NTSC8 recorded on a Sony Handycam Video8 XR CCD-TRV45E (PAL). That camcorder only has composite output. I still have it, but instead I went with a Sony Handycam Hi-8 XR CCD-TR840E (PAL) from eBay to capture via S-Video with internal TBC. I believe this is the highest end model in its class, except it doesn't have a pop-out screen like my TRV45E. I just found a TRV87E, also the highest end in it's class, with a pop-out screen. Are the TR models inferior to the TRV models in playback?

The S-Video cable is longer than it needs to be and appears to be well built and shielded. It came with the camcorder I ordered, but I don't know where it's from. My work space probably wasn't the most stable either, with cables running down a small foldable table. which may have resulted in some interference. I captured the tapes in my living room close to the windows, mostly in the evening. My ATI TV Wonder HD 600 USB was hanging down from the USB port due to the weight of the cables. However, I never had dropped frames. Sometimes, I would hear the Windows USB eject sound go off when not capturing.

I've read on some older threads on how radio interference can affect tape captures, something that I didn't realize was so sensitive and wish I was made aware of sooner. Should I buy some RFI chokes? Would a new S-Video cable, like from Blue Jeans, help?

I had my MacBook Pro, my iPhone 5, and two external hard drives, one WD My Passport and one Elements about two feet away from the camcorder too. The camcorder is running with a UK wall plug adapter and the original Sony power supply. The room temperature was around 73-76 degrees Fahrenheit. The tapes were stored in a variety of conditions over the years, mostly in a camcorder bag. Most recently, they were in a plastic container at the bottom of a shelf upstairs, which is sometimes a little warmer and dustier than downstairs where I captured the tapes.

I will try to upload samples from other tapes later this week.
Reply With Quote
  #20  
06-28-2019, 08:57 PM
cbehr91 cbehr91 is offline
Free Member
 
Join Date: Sep 2013
Posts: 134
Thanked 20 Times in 18 Posts
Re: the herringbone noise. Are you using the OEM Sony AC adpater for the camera? I had a generic China Pride brand one for a Digital8 camera that put out herringbone noise on S-Video out. Chokes did not help but it doesn't hurt to try. Spending the extra bucks for a genuine Sony adapter cleared up the issue for me.
Reply With Quote
Reply




Similar Threads
Thread Thread Starter Forum Replies Last Post
Two VirtualDub questions: cropping artifact, file splitting start/end points ashworth080142 Restore, Filter, Improve Quality 5 07-23-2014 07:05 AM
Convert NTSC to PAL and back to NTSC ? admin Capture, Record, Transfer 0 07-24-2010 08:29 PM
Ask Questions in Forums, not PMs admin General Discussion 0 01-07-2009 02:49 AM
Blu-Ray questions.. SavageAmusement Blank Media 2 07-18-2006 07:01 PM
Hd questions manthing Videography: Cameras, TVs and Players 2 03-12-2006 11:12 PM

Thread Tools



 
All times are GMT -5. The time now is 03:22 AM