Buffalo Bob
6th July 2012, 15:17
I am preparing to begin a major personal project of converting and cleaning up a large number of short duration videos in my collection for the purpose of creating good quality Standard Definition NTSC DVDs that can be enjoyed and handed down to my children.
The source videos are in various MPEG-2 NTSC and PAL formats. I will be using video software tools and techniques which are utilized on this forum and others.
Given the magnitude and nature of my project, I would very much appreciate the experts here to please review my strategy and to offer any constructive criticism or suggestions. I have several important notes and questions for your review which are listed in the last section at the end.
I have broken down all pertinent information into sections for the ease of reference and comment.
Video Source Types
Each source video clip is comprised of one of the following types of MPEG-2 files:
1. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlaced (telecined from original film source)
2. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlaced (non telecined)
3. NTSC MPEG-2; 720 x 480 encoded as 23.976 fps frame-based (progressive)
4. PAL MPEG-2; 720 x 576 encoded as 25 fps
Quality of Source Video
The quality of the source videos vary greatly depending on when they were created and from what source material. For example, some videos were produced from digitally converted video tape masters made as far back as the early 1980’s. The best quality source videos were produced recently using modern digital recording and processing equipment from beginning to end.
The Desired End Result
I am employing a “Do Not Disturb Unless Necessary” philosophy. Whenever possible I will use the source video just as it is for DVD authoring provided;
1. It does not require noise reduction, and
2. It is already formatted as one of the following MPEG-2 types:a. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlacedb. NTSC MPEG-2; 720 x 480 encoded as 23.976 fps progressive
When necessary, I will process a source video using software tools and techniques to perform a necessary format conversion and/or noise reduction to improve the video quality.
Platform and Tools
All work will be performed with P4 class computers running the XP Professional operating system SP-3, and using the following software tools and applications.
DGIndex v1.5.8 - Creates an index of an MPEG video stream.
DGDecode v1.5.8 - A MPEG-1/2 decoder used in an AviSynth script to open an index file created by DGIndex.
DGPulldown v1.0.11 - Takes a progressive MPEG2 video elementary stream, and applies pulldown flags to change it to a different legal MPEG2 frame rate.
AviSynth v2.58 - A powerful tool for video post-production which relies on a script system that allows advanced non-linear editing. AviSynth scripts are loaded via VirtualDub (and other products).
Lanczos4Resize - A high quality AviSynth filter for resizing video.
TIVTC v1.05 - An AviSynth filter that performs Inverse Telecine on film stored as interlaced video with a telecine pattern.
QTGMC v3.32 - A very high quality deinterlacer for YV12 or YUY2 input used in AviSynth scripts.
VirtualDub v1.9.9 - A video capture/processing utility.
Huffyuv v2.1.1 - Lossless video compression used by VirtualDub 1.9.9.
Neat Video v3.1.0 - A video noise reduction plug-in for VirtualDub (and other products).
Ulead MediaStudio Pro 8.0 or Corel VideoStudio X4 Pro - NLE editor/MPEG-2 Encoder.
Video Processing Methods
There are 4 video processing methods I will be using, one for each Video Source Type:
Method 1 – NTSC 29.97 fps interlaced (telecined) to NTSC 23.976 fps progressive.
Video Source Type 1. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlaced (telecined from original film source).
This video source type shall only be processed when video noise reduction is necessary.
Process steps are:
1. Demux MPEG-2 source file with DGIndex
2. Via an Avisynth script;a. Open MPEG video stream index .d2v file with DGDecode.
b. Inverse Telecine video to 23.976 fps Progressive using TIVTC.
3. Open the Avisynth script with VirtualDub 1.9.9
a. Set Video Compression to Huffyuv v2.1.1
b. Reduce/eliminate video noise with Neat Video 3.1.0
c. Save Segmented AVI file(s).
4. Open saved Segmented AVI file(s) with NLE editor/MPEG-2 Encoder and produce NTSC MPEG-2; 720 x 480 file encoded as 23.976 fps Progressive
Method 2 - NTSC 29.97 fps interlaced to NTSC 29.97 fps progressive.
Video Source Type 2. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlaced (non telecined).
This video source type shall only be processed when video noise reduction is necessary.
Process steps are:
1. Demux MPEG-2 source file with DGIndex
2. Via an Avisynth script;
a. Open MPEG video stream index .d2v file with DGDecode.
b. Deinterlace video to Progressive using QTGMC.3. Open Avisynth script with VirtualDub 1.9.9
a. Set Video Compression to Huffyuv v2.1.1
b. Reduce/eliminate video noise with Neat Video 3.1.0
c. Save Segmented AVI file(s).
4. Open saved Segmented AVI file(s) with NLE editor/MPEG-2 Encoder and produce NTSC MPEG-2; 720 x 480 file encoded as 29.97 fps Progressive
Method 3 - NTSC 23.976 fps progressive to NTSC 23.976 fps progressive.
Video Source Type 3. NTSC MPEG-2; 720 x 480 encoded as 23.976 fps Progressive.
This video source type shall only be processed when video noise reduction is necessary.
Process steps are:
1. Demux MPEG-2 source file with DGIndex
2. Via an Avisynth script;
a. Open MPEG video stream index .d2v file with DGDecode.
3. Open Avisynth script with VirtualDub 1.9.9
a. Set Video Compression to Huffyuv v2.1.1
b. Reduce/eliminate video noise with Neat Video 3.1.0
c. Save Segmented AVI file(s).
4. Open saved Segmented AVI file(s) with NLE editor/MPEG-2 Encoder and produce
NTSC MPEG-2; 720 x 480 file encoded as 23.976 fps Progressive (film).
Method 4 – PAL 25 fps to NTSC 23.976 fps progressive (film).
Video Source Type 4. PAL MPEG-2; 720 x 576 encoded as 25 fps.
This video source type shall always be processed to create NTSC video encoded as Progressive.
Process steps are:
1. Demux MPEG-2 source file with DGIndex
2. Via an Avisynth script;
a. Open MPEG video stream index .d2v file with DGDecode.
b. If Interlaced, deinterlace video to Progressive using QTGMC.
c. Resize video to 720 x 480 using Lanczos4Resize. 3. Open Avisynth script with VirtualDub 1.9.9
a. Set Video Compression to Huffyuv v2.1.1
b. If necessary, reduce/eliminate video noise with Neat Video 3.1.0
c. Save Segmented AVI file(s).
4. Open saved Segmented AVI file(s) with NLE editor/MPEG-2 Encoder and produce
NTSC MPEG-2; 720 x 480 file encoded as 25 fps Progressive.
5. Run DGPulldown on the video MPEG-2 stream to flag it up to 29.97fps
(using the 25fps --> 29.97fps conversion).
Important Notes and Questions
1. An important part of my strategy lies in the belief that producing frame-based (progressive) video is more advantageous than interlaced-based video for a variety of reasons:
a. Current video processing and display equipment (DVD players & TV’s) capable of displaying frame-based video will display it with better clarity than video that is interlaced-based.
b. Interlaced-based video is a dying technology that will eventually be replaced by frame-based video.
c. When frame-based video has to be processed and displayed on an interlaced display, the DVD player or the TV monitor will add any necessary interlacing as needed. Can anyone present a valid argument against this part of my strategy?
2. Method 1 produces a denoised inverse telecined NTSC 23.976 fps progressive video from a source which is NTSC 29.97 fps interlaced (telecined). When I need to denoise this particular video type, it makes sense to also return it to its original Film format (23.976 fps progressive).
Can anyone present a valid argument not returning telecined material to its original Film format?
3. Method 1 and Method 2 both produce a denoised frame-based (progressive) video from an interlaced source.
Can anyone present a valid argument not producing a frame-based (progressive) output video when possible, and instead producing an interlaced output video?
I thank you for taking the time to read this post and for any comments/suggestions you may contribute.
The source videos are in various MPEG-2 NTSC and PAL formats. I will be using video software tools and techniques which are utilized on this forum and others.
Given the magnitude and nature of my project, I would very much appreciate the experts here to please review my strategy and to offer any constructive criticism or suggestions. I have several important notes and questions for your review which are listed in the last section at the end.
I have broken down all pertinent information into sections for the ease of reference and comment.
Video Source Types
Each source video clip is comprised of one of the following types of MPEG-2 files:
1. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlaced (telecined from original film source)
2. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlaced (non telecined)
3. NTSC MPEG-2; 720 x 480 encoded as 23.976 fps frame-based (progressive)
4. PAL MPEG-2; 720 x 576 encoded as 25 fps
Quality of Source Video
The quality of the source videos vary greatly depending on when they were created and from what source material. For example, some videos were produced from digitally converted video tape masters made as far back as the early 1980’s. The best quality source videos were produced recently using modern digital recording and processing equipment from beginning to end.
The Desired End Result
I am employing a “Do Not Disturb Unless Necessary” philosophy. Whenever possible I will use the source video just as it is for DVD authoring provided;
1. It does not require noise reduction, and
2. It is already formatted as one of the following MPEG-2 types:a. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlacedb. NTSC MPEG-2; 720 x 480 encoded as 23.976 fps progressive
When necessary, I will process a source video using software tools and techniques to perform a necessary format conversion and/or noise reduction to improve the video quality.
Platform and Tools
All work will be performed with P4 class computers running the XP Professional operating system SP-3, and using the following software tools and applications.
DGIndex v1.5.8 - Creates an index of an MPEG video stream.
DGDecode v1.5.8 - A MPEG-1/2 decoder used in an AviSynth script to open an index file created by DGIndex.
DGPulldown v1.0.11 - Takes a progressive MPEG2 video elementary stream, and applies pulldown flags to change it to a different legal MPEG2 frame rate.
AviSynth v2.58 - A powerful tool for video post-production which relies on a script system that allows advanced non-linear editing. AviSynth scripts are loaded via VirtualDub (and other products).
Lanczos4Resize - A high quality AviSynth filter for resizing video.
TIVTC v1.05 - An AviSynth filter that performs Inverse Telecine on film stored as interlaced video with a telecine pattern.
QTGMC v3.32 - A very high quality deinterlacer for YV12 or YUY2 input used in AviSynth scripts.
VirtualDub v1.9.9 - A video capture/processing utility.
Huffyuv v2.1.1 - Lossless video compression used by VirtualDub 1.9.9.
Neat Video v3.1.0 - A video noise reduction plug-in for VirtualDub (and other products).
Ulead MediaStudio Pro 8.0 or Corel VideoStudio X4 Pro - NLE editor/MPEG-2 Encoder.
Video Processing Methods
There are 4 video processing methods I will be using, one for each Video Source Type:
Method 1 – NTSC 29.97 fps interlaced (telecined) to NTSC 23.976 fps progressive.
Video Source Type 1. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlaced (telecined from original film source).
This video source type shall only be processed when video noise reduction is necessary.
Process steps are:
1. Demux MPEG-2 source file with DGIndex
2. Via an Avisynth script;a. Open MPEG video stream index .d2v file with DGDecode.
b. Inverse Telecine video to 23.976 fps Progressive using TIVTC.
3. Open the Avisynth script with VirtualDub 1.9.9
a. Set Video Compression to Huffyuv v2.1.1
b. Reduce/eliminate video noise with Neat Video 3.1.0
c. Save Segmented AVI file(s).
4. Open saved Segmented AVI file(s) with NLE editor/MPEG-2 Encoder and produce NTSC MPEG-2; 720 x 480 file encoded as 23.976 fps Progressive
Method 2 - NTSC 29.97 fps interlaced to NTSC 29.97 fps progressive.
Video Source Type 2. NTSC MPEG-2; 720 x 480 encoded as 29.97 fps interlaced (non telecined).
This video source type shall only be processed when video noise reduction is necessary.
Process steps are:
1. Demux MPEG-2 source file with DGIndex
2. Via an Avisynth script;
a. Open MPEG video stream index .d2v file with DGDecode.
b. Deinterlace video to Progressive using QTGMC.3. Open Avisynth script with VirtualDub 1.9.9
a. Set Video Compression to Huffyuv v2.1.1
b. Reduce/eliminate video noise with Neat Video 3.1.0
c. Save Segmented AVI file(s).
4. Open saved Segmented AVI file(s) with NLE editor/MPEG-2 Encoder and produce NTSC MPEG-2; 720 x 480 file encoded as 29.97 fps Progressive
Method 3 - NTSC 23.976 fps progressive to NTSC 23.976 fps progressive.
Video Source Type 3. NTSC MPEG-2; 720 x 480 encoded as 23.976 fps Progressive.
This video source type shall only be processed when video noise reduction is necessary.
Process steps are:
1. Demux MPEG-2 source file with DGIndex
2. Via an Avisynth script;
a. Open MPEG video stream index .d2v file with DGDecode.
3. Open Avisynth script with VirtualDub 1.9.9
a. Set Video Compression to Huffyuv v2.1.1
b. Reduce/eliminate video noise with Neat Video 3.1.0
c. Save Segmented AVI file(s).
4. Open saved Segmented AVI file(s) with NLE editor/MPEG-2 Encoder and produce
NTSC MPEG-2; 720 x 480 file encoded as 23.976 fps Progressive (film).
Method 4 – PAL 25 fps to NTSC 23.976 fps progressive (film).
Video Source Type 4. PAL MPEG-2; 720 x 576 encoded as 25 fps.
This video source type shall always be processed to create NTSC video encoded as Progressive.
Process steps are:
1. Demux MPEG-2 source file with DGIndex
2. Via an Avisynth script;
a. Open MPEG video stream index .d2v file with DGDecode.
b. If Interlaced, deinterlace video to Progressive using QTGMC.
c. Resize video to 720 x 480 using Lanczos4Resize. 3. Open Avisynth script with VirtualDub 1.9.9
a. Set Video Compression to Huffyuv v2.1.1
b. If necessary, reduce/eliminate video noise with Neat Video 3.1.0
c. Save Segmented AVI file(s).
4. Open saved Segmented AVI file(s) with NLE editor/MPEG-2 Encoder and produce
NTSC MPEG-2; 720 x 480 file encoded as 25 fps Progressive.
5. Run DGPulldown on the video MPEG-2 stream to flag it up to 29.97fps
(using the 25fps --> 29.97fps conversion).
Important Notes and Questions
1. An important part of my strategy lies in the belief that producing frame-based (progressive) video is more advantageous than interlaced-based video for a variety of reasons:
a. Current video processing and display equipment (DVD players & TV’s) capable of displaying frame-based video will display it with better clarity than video that is interlaced-based.
b. Interlaced-based video is a dying technology that will eventually be replaced by frame-based video.
c. When frame-based video has to be processed and displayed on an interlaced display, the DVD player or the TV monitor will add any necessary interlacing as needed. Can anyone present a valid argument against this part of my strategy?
2. Method 1 produces a denoised inverse telecined NTSC 23.976 fps progressive video from a source which is NTSC 29.97 fps interlaced (telecined). When I need to denoise this particular video type, it makes sense to also return it to its original Film format (23.976 fps progressive).
Can anyone present a valid argument not returning telecined material to its original Film format?
3. Method 1 and Method 2 both produce a denoised frame-based (progressive) video from an interlaced source.
Can anyone present a valid argument not producing a frame-based (progressive) output video when possible, and instead producing an interlaced output video?
I thank you for taking the time to read this post and for any comments/suggestions you may contribute.