VideoHelp Forum
+ Reply to Thread
Page 3 of 4
FirstFirst 1 2 3 4 LastLast
Results 61 to 90 of 109
Thread
  1. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Originally Posted by avz10 View Post
    .... it did not want to open the avi file and gave me this error message:


    NNEDI3 is used by several popular and essential filters, including big ones like QTGMC. Many Avisynth plugins are multi-function filters that use several other plugins. Previous posts in this thread contain dll's and scripts, and instructions for copying them. Basically, you keep the download package in separate folders, then copy their required files into the Avisynth plugins folder.

    Post #30 has several plugins in the attached "support_files.zip", https://forum.videohelp.com/threads/355843-Virtualdub-filters-For-8mm-Film-That-Was-Cap...=1#post2242610 . The QTGMC folder inside that zip attachment has everything used by QTGMC.

    Others are in "scripts_and_plugins.zip" in post #17 (https://forum.videohelp.com/threads/355843-Virtualdub-filters-For-8mm-Film-That-Was-Cap...=1#post2242610). And there are still more in "more_plugins.zip" in post #54 (https://forum.videohelp.com/threads/355843-Virtualdub-filters-For-8mm-Film-That-Was-Cap...=1#post2245458). I think those .zips have just about everything, and even some duplication of a few dll's. Once you have all this stuff, you have just about everything you'll use in Avisynth for a long time to come. Use the 32-bit versions.

    It does take a while to sort this out, but the posted .zips have everything discussed here and are copies of collections posted elsewhere in this forum. They can be a hassle to locate through Google or some Avisynth filter sites. If you find anything missing, let me know.
    Last edited by sanlyn; 25th Mar 2014 at 14:30.
    Quote Quote  
  2. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    No, I have not installed it previously.

    I downloaded it and put it in the plugins folder; it then gave me an error message at cnr2-this I downloaded, and the script is working.

    A good learning experience.

    Will now start with Step 1!

    Just to confirm- settings now in VD:

    Compression-Lagarith
    Video-"fast recompress"

    Thanks

    Albie
    Quote Quote  
  3. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    Apologies- just missed #61
    Quote Quote  
  4. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    God work. Ye, it's somewhat clumsy at fiurt with all the dll's and such. But once it's set, you have quite an arsenal of cleanup filters.

    Originally Posted by avz10 View Post
    Just to confirm- settings now in VD:

    Compression-Lagarith
    Video-"fast recompress"
    That's correct, buyt VirtualDub can be a bit clumsy as well. I always set "full processing mode" on, then check these in order:
    color depth = YV12 (Or RGB24, depending on which applies for output).
    compression = Lagarith (make sure Lagarith is set for the correct output color: YV12 or RGB(Default). Lagarith's dialog can fool you because it usually defaults to the last settings you used.
    Then set "fast recompress" or "full processing mode" (for VirtualDub RGB filters), as needed.

    So for STEP1 as used here:
    full processing mode = on (so that you can check the color depth)
    color depth = YV12
    compression = Lagarith YV12
    fast recompress = on

    I can't tell you how many times I've had to stop and double check myself. A real pain in the neck.

    For step 2:
    full processing mode = on (so that you can check the color depth)
    color depth = RGB24
    compression = Lagarith RGB (default)
    fast recompress = on

    You'll note that if fast recompress is turned on, you can't check the color depth setting.
    Last edited by sanlyn; 25th Mar 2014 at 14:30. Reason: the usual ridiculous typos.
    Quote Quote  
  5. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    The attached sample "v7b" used the same sample scripts as in STEP1 thru STEP4 above, with one exception: after running the VirtualDub filters in STEP3, I made a "STEP3A" by adding NeatVideo to the results of STEP3. NV was set with noise reduction practically turned off but sharpening higher than usual. The results aren't remarkably sharper, but there's less noise even with denoise set down to 20%.

    That's what I find frustrating. You can't use really strong luma denoisers on this video without destroying everything. I think it's because of those unusually ragged and noisy edges. The filters have brought down chroma noise and flutter a lot, but I've found nothing to tighten those edges despite researching the problem for a week now. Anti-alias, line smoothers, resizing/blur/sharpen, seem to be ineffective. The sharpener in NeatVideo does have a slight effect. But I'll keep at it.....

    In VLC player you can see some banding. In the MPC-BE player and VirtualDub playback it's hardly visible.
    Last edited by sanlyn; 25th Mar 2014 at 14:30.
    Quote Quote  
  6. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    This is extremely helpful and you should use this for your guideline- I missed the setting and am now re-doing Step 1.

    So for STEP1 as used here:
    full processing mode = on (so that you can check the color depth)
    color depth = YV12
    compression = Lagarith YV12
    fast recompress = on

    I can't tell you how many times I've had to stop and double check myself. A real pain in the neck.

    For step 2:
    full processing mode = on (so that you can check the color depth)
    color depth = RGB24
    compression = Lagarith RGB (default)
    fast recompress = on
    Something else I did not know, was that the avs script should be in the same folder as the avi file- mine was on different drives and VD did not open it.

    Then I cant thank you enough, but I will comment on my progress.

    Regards

    Albie
    Quote Quote  
  7. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Excellent work. You're really doing well, much faster than when I first stumbled through this stuff (I'm still stumbling, if you haven't guessed).

    I corrected a value in the script for STEP2. I originally posted 8 for sigma4 in the fft3dfilter line. I should have typed "18", not "8". Lower numbers don't subdue the red/blue blotches well enough and will make the color work in STEP3 look too saturated. Will post some tips on the MPEG2 encoding I used, and on how I used DGpulldown.
    Last edited by sanlyn; 25th Mar 2014 at 14:30.
    Quote Quote  
  8. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    Will start with step 2 shortly, but just to double check that my settings are correct. I have changed the color depth to YV12 and it was like that when I opened VD, but in MediaInfo, it gives the color space as YUV. I suppose this is correct?









    I find the walking speed quite natural.
    Image Attached Files
    Quote Quote  
  9. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    "YUV" is correct. YUV is a generic term for several formats that store luma(Y), blue(U) and red(V) video info. YV12 is one of those YUV matrices. Note in MediaInfo that the "Chroma subsampling" is "4:2:0", which is one clue that the YUV data is YV12.

    I see your output file is almost 2 hours run time. Offhand, I'd work with a smaller sample of your original video, such as the AVI sample we've been looking at. That's the best way to analyze the final output. For the entire 2 hours of video, you'll be waiting a very long time to see results at a processing speed of less than 2 FPS or so! Most users would start with a smaller sampling for testing and adjustments. But that's up to you.

    The Walking avi looks much better than the image of that scene which you posted earlier.
    Last edited by sanlyn; 25th Mar 2014 at 14:31.
    Quote Quote  
  10. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    I am struggling to install the fft3w.dll. Apparently one needs to paste it in the system32 folder within the Windows folder, then "run" and insert regsvr32 fft3w.dll and then OK. It does not work. This dll file is also in my Avisynth plugins folder.

    Can you help please?
    Image Attached Images  
    Quote Quote  
  11. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    If you are running the 64-bit version of Windows, put FFTW3.dll in your Windows\SysWoW64 folder. Don't place a copy in System32.
    Last edited by sanlyn; 25th Mar 2014 at 14:31.
    Quote Quote  
  12. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    I'm stuck again! Sorry!

    I managed to load Step 2 and the script worked well.

    STEP3 is run entirely within VirtualDub. Open B_STEP2.avi in VirtualDub (the file is already RGB), set color depth to RGB24, set compression to Lagarith RGB default, and set "Video" to "full processing" mode. Then you can load the C_STEP3_VDub.vcf file into VirtualDub. It will load 4 filters with the settings I used: camcorder color denoise, deflicker, Color Mill, and ColorTools (the last one is disabled). A .vcf file can be opened in Notepad if you want to see what's in there. The .vcf is attached as a .zip file.
    I have done all the above settings, but the C_STEP3_VDub.vcf file does not want to load- see the screenshots, please. I have Windows 7 64 bit.







    Then a second question- is there a way/programme in which I would be able to play for instance the unedited version next to eg Step 3; or is it just possible to compare when I process from eg unedited to step 1?
    Quote Quote  
  13. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    The error message is correct. A .vcf is not a Windows application, nor is it an Avisynth plugin. It's a VirtualDub settings file. Load it in VirtualDub using "File..." -> "Load processing settings...".

    I see in your posted image that you have "ccd.vdf" and "ccd_sse2.vdf" together in your VirtualDub plugins. Just use ccd_sse2.vdf.

    The .vcf file should be placed in the area with your video project. If you start saving .vcf's in your plugins folder, you'll soon have a very large plugins folder.

    If you made it to STEP3 already, that's fantastic. My first time at this, it would have taken me a month.
    Last edited by sanlyn; 25th Mar 2014 at 14:31.
    Quote Quote  
  14. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    Apologies, but I am struggling with really minor issues.

    I have now processed 2 stages of the original video and want to start with the the third.

    I have also made a "short" copy of the unedited video to compare "before" and "after".

    I am now experiencing problems with both!!

    STEP3 is run entirely within VirtualDub. Open B_STEP2.avi in VirtualDub (the file is already RGB), set color depth to RGB24, set compression to Lagarith RGB default, and set "Video" to "full processing" mode.
    Done that

    Then you can load the C_STEP3_VDub.vcf file into VirtualDub.
    The error message is correct. A .vcf is not a Windows application, nor is it an Avisynth plugin. It's a VirtualDub settings file. Load it in VirtualDub using "File..." -> "Load processing settings...".
    I see in your posted image that you have "ccd.vdf" and "ccd_sse2.vdf" together in your VirtualDub plugins. Just use ccd_sse2.vdf.
    I deleted the ccd.vdf

    If I load it in "load processing settings" either before or after loading "B_STEP2.avi", it gives me this error message:



    I have downloaded the plugin again, tried "ccd.vdf" and "ccd_sse2.vdf" together and each one separately, putting it in the plugin folder, but no luck.

    Help please!

    Furthermore:
    I see your output file is almost 2 hours run time. Offhand, I'd work with a smaller sample of your original video, such as the AVI sample we've been looking at. That's the best way to analyze the final output.
    I made a shorter video as you suggested, but when I try to load it, it gives me this error message (what a steep learning curve!!)

    Folder with the plugins, etc for VD:



    Original script, changed to the source video as "Project unedited"


    Error message:



    I appreciate all your advice!

    Thanks

    Albie
    Quote Quote  
  15. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Well, let's see. You have a folder named VirtualDub. VirtualDub.exe should be located in that folder, not in a different place. It probably explains the error with camcorder denoise.

    There should be only one version of RemoveGrain in your Avisynth folder. It should be RemoveGrainSSE2.dll. There should be no other RemoveGrain dll in the folder. The message is coming from QTGMC.
    Last edited by sanlyn; 25th Mar 2014 at 14:31.
    Quote Quote  
  16. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    My problem trying to produce a short video:

    Error message:

    Your solution:

    There should be only one version of RemoveGrain in your Avisynth folder. It should be RemoveGrainSSE2.dll. The message is coming from QTGMC, which ships with the August 2005 version of RemoveGrainSSE2 and is about 33kb in size. It's odd that QTGMC ran earlier, so something must have been changed since that time.
    This worked! Thanks. I deleted all the other RemoveGrain filters

    I have now completed steps 1 and 2 of the "Short project"

    There must always be a BUT!

    My problem:

    If I load it in "load processing settings" either before or after loading "B_STEP2.avi", it gives me this error message:



    I have downloaded the plugin again, tried "ccd.vdf" and "ccd_sse2.vdf" together and each one separately, putting it in the plugin folder, but no luck.
    Your solution:

    Well, let's see.... You have a folder named VirtualDub in the image above. VirtualDub.exe should be located in that "VirtualDub" folder, not in a different place. It probably explains the error with camcorder denoise. VirtualDub looks for its own "plugins" in the same folder from which VirtualDub.exe is being executed.
    This solved the first error, unfortunately it got stuck again:



    "Filters" show the ccd filter loaded, but nothing else. I could not find a deflicker filter, so looked for that on the web. I downloaded Donald Graft's deflicker filter 1.1, but the error message kept on returning.





    I also tried the 1.3 beta filter, but the same error was displayed. So it seems as if I do not have the correct filter. I went through your attachments again, but could not find a deflicker filter, but, I could have missed it.



    These are all my files in the "Project" folder



    These are the VD plugins.




    Please advise me what I do not have! It seems as if I have the Color Mill and Color Tools filters.

    Thanks again! (Apologies for the language- English is my 2nd language)

    Albie
    Quote Quote  
  17. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    When a .vcf file encounters an error loading something, it simply stops loading. That's not a bug, it's intentional.

    I see that vou have deflick.vdf listed and it appears to be the correct 24KB size, but you might not have version 1.3b1. Remove the deflick.vdf you have now (I think it's v1.1) and use the one in the attached .zip file. The attached .zip contains the v1.3 .vdf and an html Help file. Copy both files into VirtualDub plugins. The deflick you have could be the old 2002 version. I do wish Donald Graft would include more info in his file headers and file names.

    The html file contains some confusing information, but jagabo in another thread suggested the settings I have used. In the help file, ignore the "Window" setting and keep it at zero (0). The value that actually does the work is "Softening". If set too high, you'll get the usual ghost trails that many temporal filters generate. I never set softening higher than 12, and that's for severe cases.

    Donald Graft's VirtualDub filters are here: http://neuron2.net/mine.html#virtualdub
    Image Attached Files
    Last edited by sanlyn; 25th Mar 2014 at 14:32.
    Quote Quote  
  18. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    Thanks, it works!
    Have just done the "Short project" (Step 2 to 3); will start with the full video now (B-STEP2 to C-STEP3).


    With regards to Step 4- are there any specific settings in VD as you indicated for Step 1 and 2?



    So for STEP1 as used here:
    full processing mode = on (so that you can check the color depth)
    color depth = YV12
    compression = Lagarith YV12
    fast recompress = on
    Quote Quote  
  19. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Originally Posted by avz10 View Post
    With regards to Step 4- are there any specific settings in VD as you indicated for Step 1 and 2?

    So for STEP1 as used here:
    full processing mode = on (so that you can check the color depth)
    color depth = YV12
    compression = Lagarith YV12
    fast recompress = on
    Use the same compression and color depth settings shown above.

    https://forum.videohelp.com/threads/355843-Virtualdub-filters-For-8mm-Film-That-Was-Cap...=1#post2245573

    Originally Posted by avz10 View Post
    With regards to Step 4- are there any specific settings in VD as you indicated for Step 1 and 2?

    So for STEP1 as used here:
    full processing mode = on (so that you can check the color depth)
    color depth = YV12
    compression = Lagarith YV12
    fast recompress = on
    Again, good work. The script I used for D_STEP4 is in post #58 (https://forum.videohelp.com/threads/355843-Virtualdub-filters-For-8mm-Film-That-Was-Cap...=1#post2245573). In STEP2, the dither plugin was used to take the video from YUV to RGB. STEP4 uses dither to perform the reverse -- from RGB back to YUV for encoding to MPEG2. It also uses LSFMod for another mild sharpening step. I posted LSFMOD requirements in post #54, in an attached more_plugins.zip. Likely you already have most of LSFMod's requirements in your plugins for QTGMC but there might be one or two that you haven't copied yet. I also tried some VirtualDub sharpeners, but the NeatVideo trial was the only one that didn't give noisy edges.

    STEP4 uses the statement "AssumeFPS(25)", which will deliver a 25FPS video for encoding at that frame rate. So the video will run "fast" until it's fixed with DGDecode later. I don't know of an MPEG2 encoder that would encode MPEG2 correctly at 17.5 FPS; TMPGEnc won't do it except at 25FPS.

    The encoding step depends on what you're using for MPEG encoding. I tested STEP4 output in HCenc and in TMPGenc Plus 2.5. Both worked and the two encodes look alike, with HCenc having just a slight edge in smooth play. Let's see if STEP4 goes well. It used the same dither() plugins used in STEP2.
    Last edited by sanlyn; 25th Mar 2014 at 14:32.
    Quote Quote  
  20. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    Well, I'm done with this part of the project!!

    Thanks to you, I have excellent video material of this precious part of my family's history. Just a pity that we could not have worked on the original 8 mm film that got lost.

    This is another story- I have my father's films and was initially very involved in this thread:

    https://forum.videohelp.com/threads/279996-Having-Trouble-with-DIY-Telecine-%288mm%29-S...ighlight=avz10,

    whereby every frame is captured, with the trigger of a mouse click. I struggled too much and got too busy- this project is for my retirement!!


    As you predicted, the speed in this video is very fast. (B-STEP2 32 GB 1h21min; C-STEP3 44 GB 1h21min; D-STEP4 17.5 GB 56min)

    So the video will run "fast" until it's fixed with DGDecode later.
    After we have fixed the speed, I want to save the videos both as avi and mpeg 2. I have enough space on my external hard drive.

    I have one more request- some parts of the video are quite dark. I would like to improve on that, but need your advice on the filter(s) to use. I suppose one can cut a relevant piece in VD, improve on it, but does VD allow us to replace the old(darker) frames or that part of the video back into the original video?
    I attach a few seconds of video. I do realize that with more light, one often gets more grain.

    I wanted to post you two clips- the "dark" clip and a clip from my "short project".
    Dark clip- 95 frames 20 mb and clip from short project- 74 frames, 2 seconds, 88mb!
    I must have a setting wrong.

    I am attaching the "dark clip". (Please do not spend a lot of time on this!!!)

    Thanks again
    Albie
    Image Attached Files
    Quote Quote  
  21. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Yes, that's really a dark clip. It's underexposed, and film and camcorders are notoriously noisy with low light levels. We've seen worse, to the point where the detail was all noise, period. This clip would take experimentation to see how much real data could be retrieved. Keep in mind that where noise is the major element of the "detail", it's impossible to retain detail without keeping a good portion of the noise.

    I bookmarked the long thread you referred to and will browse it later.

    The lower part of Post #17 (https://forum.videohelp.com/threads/355843-Virtualdub-filters-For-8mm-Film-That-Was-Cap...=1#post2242610) shows how the last scene in your earlier sample was filtered separately from the others. With VHS, and especially with home video, luma levels and color balance change drastically. It's even true of retail tapes. While this often doesn't look so horrible with analog playback on a CRT, digital processing is more sensitive to those variations and makes no corrections. So it is not unusual to take sections of source video, process them separately, and re-join them later.

    There are several ways to do this. To avoid colorspace changes and other hassles, it's better to cut those sections in Avisynth. The easiest way is to isolate part of a video with the Trim() utility. Trim() is documented in Avisynth Help. Trim() works in YUY2, YV12, and RGB. The parameters you need are the numbers of the start and end frames. If you have a section that begins at frame 10,000 and ends at frame 10,499 (that's a total of 500 frames), the command would be:

    Code:
    AviSource(path & filename.avi)
    Trim(10000,10499)
    It's a good idea to include 4 or 5 extra frames at each end. Otherwise, temporal filters won't process the first and last frames.

    There are several ways to re-join the cut section. One simple way is with the Splice commands ("AlignedSplice" and "UnalignedSplice"). The "align" has to do with how you want audio sync to occur. I always use UnalignedSplice whether there is audio or not. A shortcut for UnalignedSplice is the "+" symbol, which can also be used to join strings of characters.

    In the code below, "Vid1" is used to name the first 9,999 frames of the original source of many thousands of frames. "Vid2" is used to name the separated source of 500 frames, and includes all frames. "Vid3" is a name created for the main source from frame 10,500 to the end. "Vid4" is a name created for the joined pieces and final output:

    Code:
    Vid1=AviSource("path & mainfile.avi").Trim(0,9999)   #<- the first 0 means to start at the first frame.
    Vid2=AviSource("path & CutClip.avi")
    Vid3=AViSource("path & mainfile.avi").Trim(10500,0)  #<- the second 0 means go to the end.
    Vid4=Vid1+Vid2+Vid3
    return Vid4
    If "Vid2" included 5 extra frames at the beginning and end, you would have 510 frames instead of 500. Because the first frame is frame 0, the actual frame numbers in a clip of 510 frames would be 0 to 509. The frames you want to remove are the first 5 frames (0 to 4) and the last 5 frames (505 to 509). The frames you want to keep are the 500 frames between 5 and 504. You would trim the extra frames from the start and the end by including the frame numbers you want to keep:

    Code:
    Vid2=AviSource("path & name of CutClip.avi").Trim(5,504)  # <- a total of 500 frames.
    There are Avisynth plugins that can replace frames in one clip with frames from another clip. You can also join clips and pieces in VirtualDub.

    The names "Vid1", "Vid2", etc., are simply names you can invent. They are called user-defined variables. You could just as well call them v1, v2, a1, a2, a, b, c, etc., as long as the names aren't the same as commands, functions, or plugins.

    I suggest that your permanent archive would be the captured AVI original. I often keep intermediate AVI work files, but if you keep the scripts you can always recreate those files from the original capture. The advantage in keeping the AVi original is that methods for deinterlace, inverse telecine, unblend, etc., are always being improved. QTGMC itself is a new plugin.
    Last edited by sanlyn; 25th Mar 2014 at 14:32.
    Quote Quote  
  22. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    To encode the results of STEP4 you can use the encoder of your choice. I used HCenc and TMPGENc Plus 2.5, which give cleaner results and are more flexible than most free and expensive one-button applications or NLE's. I strongly suggest that you use a two-pass encode, with a target bitrate of 6500 and a max bitrate of 9000. A soft video with noise and the strong probability of banding and macroblocks requires high bitrates. High bitrates would not allow you to fit a two hour MPEG2 on a single DVD disc, but it would easily fit on a dual-layer DVD. Also, some space has been saved by removing duplicate and blended frames and by using pulldown to fill in the gaps. Pulldown doesn't create new frames; it flags the video and tells your playback equipment to repeat periodic frames or fields. Thus, the actual sizes of the encoded MPEG and the pulldown MPEG will be the same.

    The STEP4 clip I posted was encoded with the free HCenc encoder. HCenc ships with a user interface. They need no installer. Just unzip the HCenc package into a folder that you create for it, and click on the encoder's GUI "exe" file. HCenc comes with a user manual, but most of the time you can use default settings and change just a few of them. The attached .rtf file can be opened in Microsoft Wordpad, which comes free with Windows. It has screen captures of all the panels and settings I used for the STEP4 encode, marked with blue arrows to indicate the items to be checked (as I said, most of those settings are defaults anyway).

    HCenc requires that you open a video AVI using an avs script. You can open the STEP4 file with D_STEP4.avs, which is the same script that was used to create the STEP4 output file. If that's a bigger file than your short sample, it will take a long time to open. So if you have a finished AVi that you want to encode, just make a simple script that opens your AVI, like this:

    Code:
    AviSource("drive:\path to folder\video.avi")
    That's all you need to open a YV12 AVI in HCenc. HCenc will encode the AVi as a PAL 25FPS progressive .m2v.

    The last step would be to apply pulldown to the .m2v using DGpulldown.
    Last edited by sanlyn; 25th Mar 2014 at 14:33.
    Quote Quote  
  23. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    DGpulldown is rather straightforward, and it works quickly. Start by opening the .m2v “Source” file, then check the file location and name in the “Dest ES” output section. The check boxes below are all you need, along with the “17.5” and “25” in “Custom”.You will have to turn on “Top field first” and “Set timecodes”.Then click “Convert” at the bottom. You’ll see a completion message when the pulldown is finished.

    You will recall from the HCenc encoder images that a GOP (Group of Pictures) size of 10 was requested. This GOP was requested because DGpulldown will add pulldown frames to adjust the frame rate from 17.5 FPS to 25 FPS. DGpulldown will do this by adding pulldown flags that adds 5 pulldown frames to every group of 10 original frames, for a final output GOP of 15 – that’s the maximum GOP size for PAL DVD.

    Image
    [Attachment 18301 - Click to enlarge]
    Last edited by sanlyn; 25th Mar 2014 at 14:33.
    Quote Quote  
  24. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    Thanks for all the info, will start tomorrow depending on what I need to do at work!
    Quote Quote  
  25. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Yes, work always gets in my way, as well (as if fixing damaged video is not work? ? ? )

    And wives tend to impose delays. They can't help it, they just don't understand this stuff.
    Last edited by sanlyn; 25th Mar 2014 at 14:33.
    Quote Quote  
  26. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    TThe easiest way is to isolate part of a video with the Trim() utility. Trim() is documented in Avisynth Help. Trim() works in YUY2, YV12, and RGB. The parameters you need are the numbers of the start and end frames. If you have a section that begins at frame 10,000 and ends at frame 10,499 (that's a total of 500 frames), the command would be:

    Code:
    AviSource(path & filename.avi) Trim(10000,10499)
    I understand the reasoning behind the extra frames and the numbering.

    Should I therefore cut the video in three sections, convert the middle clip and then join them again?

    In the code below, "Vid1" is used to name the first 9,999 frames of the original source of many thousands of frames. "Vid2" is used to name the separated source of 500 frames, and includes all frames. "Vid3" is a name created for the main source from frame 10,500 to the end. "Vid4" is a name created for the joined pieces and final output:

    Code:
    Vid1=AviSource("path & mainfile.avi").Trim(0,9999) #<- the first 0 means to start at the first frame. Vid2=AviSource("path & CutClip.avi") Vid3=AViSource("path & mainfile.avi").Trim(10500,0) #<- the second 0 means go to the end. Vid4=Vid1+Vid2+Vid3 return Vid4
    With regards to a AviSynth plugin/script to add some light- I would definitely not overdo it, what do you propose?

    Encoding

    To encode the results of STEP4 you can use the encoder of your choice. I used HCenc and TMPGENc Plus 2.5, which give cleaner results and are more flexible than most free and expensive one-button applications or NLE's. I strongly suggest that you use a two-pass encode 2.5, which give cleaner results and are more flexible than most free and expensive one-button applications or NLE's. I strongly suggest that you use a two-pass encode
    I just want to double check- So, I encode first with HCenc and the resulting video from that encoding, I encode in TMPGENc?

    DPpulldown sounds easy and I suppose the speed will now portray as normal.

    Have a good day

    Albie
    Quote Quote  
  27. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Originally Posted by avz10 View Post
    I just want to double check- So, I encode first with HCenc and the resulting video from that encoding, I encode in TMPGENc?
    No, once you have your .m2v encoded, don't re-encode. I mentioned HCenc and TMPGenc because I made two test versions of the video, but you only need one or the other. HCenc looked very slightly "better", but that's not absolute -- given the same bitrate and other parameters, it would be very difficult to see a difference between the two. Both are good encoders. But you need only one encode.

    Originally Posted by avz10 View Post
    DPpulldown sounds easy and I suppose the speed will now portray as normal.
    AFter pulldown is applied, the video will play at a normal pace. Neither encoder will work properly at 17.5FPS, because it's not a standard frame rate that they recognize. So "25FPS" is a ploy to trick the encoders into behaving normally. DGpulldown applies the proper flags to get playback as close as possible to the requested frame rate.

    Your question about adjusting brightness requires a long answer that I can address later today. Basically, because you are using VirtualDub filters to work with some noise and color, you can adjust brightness with any of several filters. I'd suggest Color Mill because it can adjust color and brightness in specific areas. But note that video will usually look brighter on TV than on your computer. You'll also see a difference with various media players. These players have image adjustments, but I leave them unadjusted: it's difficult to judge their effects by eyesight alone, and they aren't linear. I find that Media Player Classic or Media Player Classic-BE (MPC-BE) gives a more accurate idea of how levels and colors will fall into place. The MPC-BE download link on this page in the Tools section (https://www.videohelp.com/tools/MPC-BE) defaults to the newer 64-bit version of MPC-BE.

    Brightness and contrast filters work in different ways, depending on whose filter you use. ColorMill and gradation curves work in a way that is similar to those in higher-end video applications. The term "brightness" usually refers to adjusting black levels. Adjusting brightness will darken or brighten dark areas more than other parts of the image. "Contrast" will darken or brighten the brightest areas more than darker ones. In practise, both adjustments affect each other to a visible extent, so you have to tweak back and forth between the two to make things correct. The adjustment that affects the entire image but is focussed mostly in the middle areas is usually called "gamma".

    I'll post more info about color and levels tools later today.
    Last edited by sanlyn; 25th Mar 2014 at 14:34.
    Quote Quote  
  28. Member
    Join Date
    Jan 2006
    Location
    South Africa
    Search Comp PM
    I am now at the encoding and pulldown stages.

    As I stated earlier, I want to save the videos both as avi and as mpeg. I feel that I should first do pulldown to get the speed right and then one can basically "store" this file. Then I can encode to mpeg.

    A little later about brightness and contrast- I do quite a lot of work on Photoshop and has used the VD filters in the past, so I am not too new with levels, etc
    Quote Quote  
  29. Originally Posted by avz10 View Post
    As I stated earlier, I want to save the videos both as avi and as mpeg.
    When saying AVI, you mean XviD or DivX AVI, right? AVIs don't use pulldown. You can keep the framerate at 17.5 for AVI. It's only for DVD that you have to encode at a 'compliant' framerate (25fps in your case) followed by 17.5->25fps pulldown with DGPulldown.
    I feel that I should first do pulldown to get the speed right and then one can basically "store" this file.
    And if you do mean XviD or DivX AVI, I doubt very much I'd like my archival copy to be one of them. If you mean DV or lossless AVI then, again, no pulldown and encode for 17.5fps.
    Quote Quote  
  30. Banned
    Join Date
    Oct 2004
    Location
    New York, US
    Search Comp PM
    Originally Posted by avz10 View Post
    A little later about brightness and contrast- I do quite a lot of work on Photoshop and has used the VD filters in the past, so I am not too new with levels, etc
    Having some experience with Photoshop will be very useful. Video works in a similar way, but it has its special considerations. In any case, it's difficult to judge graphics characteristics using eyeballs alone. Histograms are not just helpful, they're essential. They show where the problems are.

    Avisynth has a histogram, a vectorscope, and other tools that tell you what the YUV data is doing. One of the simplest is a favorite tool: the "Levels" histogram. This histogram works only in YV12. A script displays the levels histogram using this code:

    Code:
    ConvertToYV12(interlaced=false)  # <- convert if necessary
    Histogram(mode=Levels")
    The image below is a frame from the original, unprocessed sample. In the image is a YUV histogram (midddle) and an RGB histogram (right). The YUV histogram has three sections: Y or Luma/brightness (top), U (middle), and V (bottom). The histogram shows that luma is barely within the acceptable RGB 16-240 range; the "trouble" range is indicated by the different colors of the histogram's left and right edge borders. Brightness tapers off quickly across the middle; note that the red portion of the V channel at bottom has very little red response. The U and V portions of this histogram aren't all that accurate, but it's accurate enough to show that the red deficit, so the image has too much cyan (blue+green), especially at the bright end. VirtualDub's ColorTools RGB histogram shows the color problem more precisely: red tapers off above the midtones (middle portion) and green and blue are outside the RGB 235 indicators at the right -- meaning that there is no bright red. Blue and green brights exceed desirable levels. At the left side, all color and luma have a small peak below RGB16, so the image is dim overall, with "thin" color data and little in the midrange, and with "hot spots" in the bright blanket. (Click the image to enlarge)
    Image
    [Attachment 18309 - Click to enlarge]


    Below is the same frame after processing in STEP1. SmoothLevels, ColorYUV and Sgradation were used to draw luma and chroma "into" the histoghram, and to lighten the shadow portions. One could make the shadows brigfhter, but the only detail in those dark areas is noise. YUV shows only a very slight increase in red. The RGB histograms shows the enlarged peaks of the brighter shadow areas, and bright green and blue have been pul;led inisde the histogram borders. There is still a too little red; because you can't "create" red data in an image, RGB filters will be used to literally "stretch" the red that's present. (Click the image to enlarge)
    Image
    [Attachment 18310 - Click to enlarge]


    The same frame after STEP3 and ColorMill. The brighter shadow areas have been tamed somewhat to give the image some depth, bright reds are extended, bright green and blue are reigned inward, saturation and middle reds are increased, and the point known as the "middle point" has been raised to get some contrast without blowing away highlights. Note that even though the "hot spots" have a lower brightness value than the original, those areas still look "hot" because there is no detail in those highlights.
    Image
    [Attachment 18311 - Click to enlarge]


    The scene that precedes this shot was made in bright, direct sunlight and is brighter. The scene shown was photographed in subdued light (it looks like overcast). The following scene was shot in very hazy, cloudy sunlight. The last scene was shot in shadows with random areas of bright sun, and is more heavily cyan in the darks than the other scenes. If one wants more "perfect" corrections, gradation curves and other filters would be added to "tweak" each camera shot. The same, exact settings can't be used to perfection for every scene. The settings used in the scripts and .vcf were a compromise to make the filters work for suitable for all the scenes.

    The shadow areas in these images will look more dense than they do in the video, due to the way some viewers and browsers display those extremes.
    Last edited by sanlyn; 25th Mar 2014 at 14:34.
    Quote Quote  



Similar Threads

Visit our sponsor! Try DVDFab and backup Blu-rays!