Tips and Tricks for Encoding Long Format Content with Windows Media Encoder
Tricia Gill
Microsoft Corporation
August 2003
Applies to:
Microsoft® Windows Media® Encoder 7.x
Windows Media Encoder 9 Series
Introduction
Content produced by the film and video industry includes feature films, commercials, and episodic television. This content is produced with great care in order to meet the highest quality standards.
It can originate from film, videotape, digital videotape, or a mixture of all three. Before being distributed, the content must undergo special processing to correct color imperfections, add
special effects, edit scenes together, synchronize the audio and video tracks, and so on. The final step is to put the content into various master formats. As streaming media becomes more popular,
studios can create Microsoft® Windows Media® files from these masters for distribution.
Feedback
E-mail us with your comments and feedback
about this article.
Abstract
This article discusses tips and tricks for using Windows Media Encoder to encode high-quality masters that will be used for both online streaming and download-and-play scenarios. This article
also includes some steps that you can take before encoding to ensure that the encoding process fits in seamlessly with your existing video production processes.
This article discusses tips and tricks for using Microsoft Windows Media Encoder to encode high-quality masters that will be used for both online streaming and download-and-play scenarios. This
article also includes some steps that you can take before encoding to ensure that the encoding process fits in seamlessly with your existing video production processes.
This article covers the following topics:
Preparing Your Content. Describes some adjustments that you can make to your content before encoding it.
Configuring Your Hardware. Recommends configurations for the encoding and client computers to optimize the encoding and viewing experience.
Encoding Scenarios. Describes common scenarios that are encountered by film and video professionals during the encoding process.
Achieving the Best Quality. Provides techniques that you can use to encode your content most effectively.
Before content ever reaches Windows Media Encoder, you should complete all editing, cleaning, and other processing. The encoder cannot differentiate between artifacts and the actual content
and encodes all of it equally. Encoding these artifacts, such as video noise or film scratches, absorbs encoding resources, takes up precious bandwidth, and may even result in blockiness in
the encoded video. If you remove or reduce these artifacts from your content, you can apply more encoding resources and bandwidth to the actual content details. The result is higher-quality
encoded video; the encoded output looks much better when the input is as crisp and clean as possible.
The type of content you are working with and the ultimate use of the encoded output should dictate which encoding techniques you use. Not all techniques are appropriate for all content. To prepare
your content for encoding, you can apply the following techniques:
Noise filtering. This technique is widely used to reduce noise and increase video quality. Noise filters use spatial-adaptive and temporal techniques to remove imperceptible high-frequency
noise that takes up bandwidth. You can adjust spatial-adaptive filters to soften parts of an image while leaving other parts, particularly high-contrast edges, alone. Temporal filters improve
the picture-quality at the expense of introducing some amount of motion-dependent blur.
Whether you use temporal filters as part of a film scratch removal system or as a stand-alone encoding preprocessor, watch out for the introduction of motion artifacts. Adjust
temporal filters to a lower setting, but do not eliminate the overall noise level.
Brick wall filter and aperture correction. The brick wall filter reduces high-frequency noise by softening all parts of an image equally. An aperture corrector is a parametric equalizer
that can restore the "snap" to a soft image. To preserve a crisp look to your video, set your brick wall filter to introduce a steep low-pass of about 3.2 megahertz (MHz). Use the aperture
corrector to add a broad peak of approximately 2.5 MHz.
Pedestal level. Increasing the pedestal softens the dark blacks to make them less obtrusive. Increase the pedestal level slightly to correct for video graphics adapter (VGA) maladjustments
of the brightness controls on VGA displays.
Grain. You can hide grain in your film by tricking the scratch remover during the telecine process. To do this, select the maximum block size and set the motion threshold to maximum.
Encoding is a resource-intensive process, especially when the content includes a lot of motion or detail. When encoding a live event in real time, the speed of your processor is most important.
When encoding from a file or transcoding, especially when encoding from Audio Video Interleaved (AVI) files, your computer's bus speed is the limiting factor.
Lab tests conducted at Microsoft used the following hardware configuration:
Dual Intel Pentium III 1-gigahertz (GHz) processor
Pentium III motherboard with built-in U160W SCSI controller and Intel 10/100 BaseT network adapter with AC97-compatible on-board audio and a dual PCI bus
Three 36 gigabyte (GB) Ultra 160 SCSI 10 K hard disks
Accelerated Graphics Port (AGP) 4X 64-megabyte (MB) DDR RAM video controller
Viewcast Osprey-500WM capture card
DVD-ROM internal drive
1.44 MB internal floppy disk
512 MB, 800 MHz ECC RDRAM
21-inch monitor
For high-quality audio, lab computers also used digital audio capture cards that support audio input at 44 and 48 kilohertz (kHz).
Not only do you need a lot of horsepower to encode high-quality video, but your audience needs a powerful computer to enjoy a satisfactory viewing experience. Client computers should have a
fast processor, plenty of memory, and a graphics card that supports the Microsoft DirectX® application programming interface (API).
The following table suggests optimal configurations for viewing content encoded at various resolutions and frame rates.
To view this resolution
Use the following minimum configuration
320 pixels by 240 pixels at 30 frames per second (fps)
300 MHz processor or higher with 64 MB RAM
640 pixels by 480 pixels at 30 fps
700 MHz processor or higher with 128 MB RAM
320 pixels by 240 pixels at 60 fps
700 MHz processor or higher with 64 MB RAM
Note: Before encoding your content, consider your audience. If most of your users do not have the caliber of equipment needed to view
high-resolution video, you might need to sacrifice quality to reach more consumers.
For optimal performance while you are encoding, do the following:
Make sure to defragment your hard disk.
Turn off network or disk sharing to prevent anyone from accessing your hard disk while you are capturing content.
Close any other applications, especially those that access the hard disk.
If you encounter problems, make sure that your hardware can keep up with the data feed, especially when capturing at full resolution (640 pixels by 480 pixels). Also, you might have direct memory
access (DMA) buffer conflicts between the capture and SCSI cards. You can fix this problem by using a dual PCI bus system or by separating the SCSI controllers and the capture card so that they
do not share the same bus.
The following scenarios describe common uses of Windows Media Encoder. To help you encode your content most effectively, each scenario includes encoding recommendations.
Synchronizing the Encoder with Tape Decks
Encoding houses often need to synchronize the process of starting and stopping Windows Media Encoder at exact frame locations in the source. Also, they may need to create several Windows Media
files at different bit rates from the same master. Rather than create these files in several separate processes, encoding houses can create all the masters at once by controlling multiple tape
decks and encoders from a central location. All decks are synchronized to start and stop with the source reel.
Adding computer-based components to the mix, such as a computer running Windows Media Encoder, creates some challenges because hardware and software systems have inherent latency that prevent
absolute synchronization. This latency is usually consistent within a given configuration. In Windows Media Encoder 9 Series, you can directly control the source by creating an edit decision
list (EDL), which identifies the portions of the tape you want to encode. The encoder can then queue the tape automatically and encode the segments defined by the EDL.
In earlier versions of the encoder that do not control the source, you can overcome the latency issue by setting the exact frame at which the encoder will start encoding, thereby compensating
for any delay between the time when the Start command is issued and when encoding actually begins. You will need to experiment with your system to determine its latency in frames. After you
determine the length of that delay in frames, you can configure the encoder to start early by that number of frames. For example, if your computer starts encoding 10 frames after you press Start,
and you want to start encoding at frame 1, then you would start the encoder at -9 (9 frames before you actually need it to start).
Encoding Letterboxed Content
In the letterboxing process, a visual image with a 16:9 aspect ratio is formatted for viewing on a screen with a 4:3 aspect ratio. As a result, black bars are displayed across the top and bottom
of each frame.
When you encode the video, the codecs compress these bars just like any other content, even though no visual data is contained within them. By cropping out the bars but preserving the 16:9 aspect
ratio, you can encode only the film content, reserving precious bandwidth for the details of the encoded video.
Encoding letterboxed content is very similar to encoding any other kind of content, except that you add two steps:
Crop the top and bottom portions of the video to eliminate the black bars.
Set the appropriate output size in the profile.
Creating Archives
You can use the high-quality files that you have already encoded and archived as sources for future encoding sessions. They serve as master input sources that you can encode at various lower
bit rates depending on the bandwidth and connection speeds available to your clients.
Although you can create archives simultaneously while broadcasting, to obtain the highest-quality archive, you should encode directly to a file so you will not be limited by any bandwidth or
network constraints that are often present in broadcast scenarios.
Always begin with the best possible source. All preprocessing and editing of your content should be complete, and all artifacts and scratches removed, if possible. Next, create a profile, and
specify a bit rate that is higher than the bit rates you intend to stream. Be advised that Microsoft Windows Media Player cannot play content encoded at very high bit rates, such as 6 megabits
per second (Mbps) or higher, but you will be using this file as a high-quality master and not for streaming.
After your profile is completed, set up an encoding session, and then encode your content by using the profile you just created. When you are ready to use that archive to encode content at a
lower bit rate, set up an encoding session. If your clients will be connecting at a variety of speeds, you can select a multiple-bit-rate (MBR) profile, which will cause the encoder to encode
the content at several different bit rates. After you have set up your encoding session, encode the content. The resulting files are transcoded from the archive to the lower bit rate format.
Encoding in Real Time and Capturing to AVI
When deciding whether to encode in real time, you should consider the quality of your source and the intended use of the encoded video.
Using Windows Media Encoder to encode content in real time has the advantage of being a one-step process. There are no intermediate files to create or store, and you can capture and encode the
audio and video in a single pass of the original source. You can also encode from multiple sources and switch between them during the encoding session. Real-time encoding also results in fewer
RGB/YUV color space conversions, which can cause various color and encoding artifacts on the monitor.
But real-time encoding does require a powerful computer. The following minimum configuration is recommended:
Dual Pentium III 733-MHz processor or faster (640 pixel by 480 pixel image size may require dual 1-GHz processors)
256 MB RAM (512 MB recommended)
Dual PCI bus
The video capture card is also an important component. Use a video capture card that allows complete digital real-time capture directly to Windows Media Format. For example, the Viewcast Osprey-500WM
enables you to use high-quality digital video (DV), SDI video, and digital audio input sources. In addition, the Osprey-500WM provides hardware video processing, including a temporal/spatial
deinterlacing filter, a digital video decoder, and video scaling.
When encoding in real time, you should continually monitor your system to ensure it is keeping up with the data. The encoding process uses many components of your encoding computer, including
the capture card, drivers, processor, hard disk, and memory. CPU usage should never exceed 80 percent because the processor must be able to handle other system processes while encoding is in
progress.
Capturing in AVI format enables Windows Media Encoder to take all the time it needs to convert the AVI file into Windows Media Format. By taking extra time, the encoder can determine the best
combination of bandwidth, bit rate, quality, frame rate, and buffer size to produce the best quality encoded video. You can also edit an AVI file before encoding it to rearrange segments in
the file, reduce noise, adjust contrast and brightness, and so on. Editing an encoded Windows Media file properly is not possible because of buffering issues. Therefore, if you have editing
to do, capturing in AVI format might be the best solution.
Capturing in AVI format does have some disadvantages. It is a two-step process because you capture the data first, and then encode it. Because AVI is an uncompressed format, your files could
be very large. Moving these files from the hard disk to memory also puts higher demands on your system bus. The following minimum configuration for capturing AVI at 640-pixel by 480-pixel resolution
is recommended:
133 MHz system bus
64-bit U2 or Ultra 160 SCSI controller
SCSI U2 or Ultra 160 hard disks running 10,000 to 15,000 RPM
RAID 0 stripe
Video capture card that allows complete digital real-time capture
If you decide to capture in AVI format before encoding, you need to be aware of the 2 GB file limitation. The original AVI format specification required that hardware and software tools built
to that specification support AVI files that are smaller than 2 GB. The specification was changed to allow files that are larger than 2 GB, but some of the hardware and software tools that support
the AVI format do not yet support this updated specification. This AVI format is often called AVI Type II or extended AVI.
Windows Media Encoder supports the AVI Type II format and will encode AVI sources that are larger than 2 GB. If you plan to edit your AVI files before encoding them and expect the files to exceed
2 GB, make sure your editing tools support the AVI Type II format.
Encoding Content That Is Near-DVD Quality
Producing encoded content at near-DVD quality requires a powerful computer. Because encoding this content is resource-intensive, you will achieve the best results by capturing in AVI format
first. Capturing in AVI format enables you to create an uncompressed file that you can edit before encoding. It also means that the encoder can take more time to compress the content without
the burden of encoding it in real time. This results in higher-quality output because no frames are dropped as a consequence of trying to keep up with the flow of data.
When encoding content that is near-DVD quality, create or edit a profile to include the settings identified in the following table.
Setting
Value
Video size
640 x 480
Frame rate
24 fps (inverse telecined) or 30 fps (video)
Bit rate (content dependent)
At least 750 Kbps, and up to 1-2 Mbps or more
Audio format
64 Kbps, 44 kHz, stereo or above
Switching Sources While Encoding
Windows Media Encoder enables you to define and switch between multiple sources, such as cameras, files, and other devices, during an encoding session. Each source, with the exception of files
or screen captures, must be connected to a separate capture card.
Windows Media Encoder can also work in conjunction with other providers' hardware or software switchers. When using a hardware switcher, connect the various sources to the switcher, and then
connect the switcher to a capture card that is installed on your encoding computer. The output of the switcher should be a National Television Standards Committee (NTSC) signal. Then set up
your encoder to accept input from the capture card.
When using a software switcher, connect your devices to a capture card or cards that are installed on your computer. You can control the input from each device from the other provider's software.
Many applications allow you to encode within their environment and programmatically call the Windows Media Encoder.
Achieving high-quality encoded video is not an exact science. There are a number of adjustments that you can make to your content, hardware, and software depending on the results you are trying
to achieve. This section addresses the fine-tuning that you can do to encode your content most effectively. Each of the following recommendations assumes that you are encoding at a constant
bit rate.
Adjusting Stream Settings
For constant-bit-rate streams, the image quality, buffer, bit rate, and key frame interval settings are closely tied to one another. Adjusting any one of these settings affects the others. Finding
the right balance among these four settings takes trial and error, and how you set them depends on the effect you are trying to achieve. For example, you might be willing to drop a few frames
in order to set your image quality level to high and achieve the clearest images in your video. You must decide which of the settings is most important and adjust the others accordingly. Some
guidelines are provided in the following topics.
Image Quality
The image quality (video smoothness) setting allows you to choose between smoother motion and clearer images. The clearer the images, the more bandwidth is required to stream the video. If the
image-quality level is too high relative to the available bandwidth or buffer size, the codecs must drop frames to maintain the desired quality level. When selecting the level of quality to
apply to your content, keep the following tips in mind:
Set the image-quality level between 80 and 100 for content that does not contain a lot of motion or intense detail.
Set the image-quality level between 50 and 70 for content that contains high motion or a lot of detail.
Experiment with your content to find the highest quality setting that you can use without dropping frames. The quality setting you choose will be the lowest threshold allowable for the quality
of the output. To determine whether Windows Media Encoder 9 Series is dropping frames, click the Statistics tab on the Video panel. You can review the total number of dropped frames as
well as detailed information about the number of frames dropped before compression, during compression, and after compression. To ensure the highest quality video, the number of dropped frames
should remain at zero.
Buffer
The buffer is a temporary storage location on the client computer where content is stored before Windows Media Player renders it. The buffer is used so that the client does not have to wait
for content to travel from the server. This improves the viewing experience because it cuts down on gaps or breaks in the video. But keeping content in the buffer does require memory. The bigger
the buffer, the more memory is required to store the information. Increasing the buffer size, especially for high-resolution video, can overload the computer's memory and affect the computer's
ability to encode information.
Specifying a larger buffer size allows the Player to provide better video image quality. However, users must wait longer before the video starts because the video cannot begin playing until
the buffer is full. Typically, this buffer delay matches the buffer size that you specify in the encoder. You should set the buffer size based on whether you want your users to view streaming
content or play back the content locally. If you are streaming content, set the buffer size as low as possible. A setting of less than 7 seconds is recommended, and a setting of 3 to 5 seconds
is optimal. If users will play back the content locally, a larger buffer size will increase the video quality. In this case, a setting of 7 to 15 seconds is recommended.
The one exception to buffer size is with content that you stream from a server running Windows Media Services 9 Series to a client running Windows Media Player 9 Series. In this scenario, you
can set a larger buffer size in the encoder. The Fast Start feature in Windows Media Services 9 Series enables the Player to fill its buffer faster than real time if sufficient bandwidth is
available. This means that the initial buffer delay may be much shorter. For example, when a user attempts to play back content encoded at 56-Kbps by using a DSL or cable modem, the start-up
delay may be only a second or two, even if the buffer size specified during encoding was larger. For more information about Fast Start, see Windows Media Services Help.
If you set up an encoding session that uses peak bit rate-based variable bit rate (VBR) encoding for audio, follow these recommendations:
If you are using constant bit rate (CBR) encoding for video, specify a peak buffer size for audio that is lower than the buffer size for video.
If you are using peak bit rate-based encoding for video, specify a peak buffer size for audio that is lower than the peak buffer size for video.
Buffering and Variable Bit Rates
All content varies in the amount of motion or detail it contains in any particular frame. Streaming more motion or intense detail requires more bandwidth to achieve the same viewing quality
that is achieved when streaming content with less motion or detail.
If you are encoding for delivery over a fixed connection, all content is encoded at a constant bit rate. If that bit rate is high, more bandwidth will be required to stream the content. You
can set a lower bit rate and achieve some variability between the frames by using a larger buffer. The key is to find the right combination of bit rate and buffer size.
The optimal size of the buffer depends on the kind of content you are encoding. For download-and-play scenarios, you can set a higher buffer value without suffering the penalty of a long loading
time because the content will be played locally after it is downloaded. A larger buffer allows for more variability in the bit rate during playback. This variability closely approximates the
variability in the original audio or video. However, when users progressively download files from a Web server or stream digital media from a streaming media server, the buffer value directly
impacts the amount of time the Player is in a buffering state. In this case, reducing the buffer value can improve the users' experience.
The Player buffers files by using the following formula:
buffer time = Max([encoder buffer value + 2 seconds], player's network buffer size)
By default the buffer setting for the Player is 5 seconds. Therefore, an encoder buffer value of less than 3 seconds does not change the amount of buffering but it will decrease the content
quality. If you decrease the Player buffer to less than 5 seconds, and you set the encoder buffer value to less than 3 seconds, you can decrease buffering time. However, setting the Player buffer
to 2 seconds or less can decrease performance and cause additional rebuffering.
The following guidelines are based on distribution method:
For download-and-play at 320-pixel by 240-pixel resolution, use a 20-second buffer. Depending on the amount of memory that you expect your clients to have, you might need to use a smaller
buffer, because the entire buffer needs to be cached in memory before playback.
For streaming, limit your buffer to 20 seconds or less to minimize the amount of time clients need to wait for the stream. Setting the buffer close to 20 seconds will give you better encoding
of high-bit-rate segments but will result in a longer wait on the client side. Setting the buffer below 20 seconds improves the wait and reduces the latency in client-side seeking. The recommended
value for streaming is 3 to 5 seconds.
Note: In Windows Media Player 9 Series and Windows Media Services 9 Series, the Fast Start feature enabled. If the Player has enough
bandwidth, approximately 10 seconds of buffer is sent to it as quickly as possible. The more bandwidth available to the Player, the faster it can receive the content. Therefore, if Fast Start
is enabled and enough bandwidth is available, you can increase the encoder buffer value to 5 or 7 seconds without significantly impacting the buffering time that users experience.
Key Frames
Video is composed of key and delta frames. Key frames contain an entire image, and delta frames contain only the part of the image that changed from the last key frame. Delta frames require
much less bandwidth than key frames, so it is best to maximize them to reserve more bits for the image quality in each frame. For example, using more key frames when streaming a constant-bit-rate
file, results in fewer bits being devoted to image quality because those bits are needed to generate each key frame. On the other hand, if a delta frame changes very little from the previous
frame, few bits are needed to create that frame, and these bits can be applied toward generating better video quality.
Generally, high-motion content requires more key frames because the scenes are changing more frequently. If you increase the time between key frames, the size of the video (in bytes) gets
smaller because more delta frames are used. Similarly, if you decrease the time between key frames, more key frames are used and the file size of the video gets larger. Fewer seconds between
key frames can enable better seeking within the file, but at the cost of bandwidth.
When you configure your encoding session, you can set the interval at which key frames are generated. The key frame interval is part of the profile and controls the number of seconds between
key frames. The optimal interval you choose depends on the type of content that you are encoding and the buffer, image quality, and bit rate settings you selected.
Regardless of the key frame interval you choose, Windows Media Encoder will automatically generate a key frame any time the level of change between frames reaches a certain threshold. For
example, if 50 percent or more of the image is different, the encoder generates a key frame. Thus, for rapid changes in the video, the encoder will generate additional key frames at each transition
in addition to generating key frames at your specified interval. One approach to optimizing key frames is to set a large key frame interval, such as 20 seconds, and allow the codecs to fill
in additional key frames when necessary.
Encoding MBR Streams
In Windows Media Encoder 9 Series, each video stream in an MBR file can have a different video resolution, and each audio stream in an MBR file can have a different audio format. For example,
you can include multiple languages in a single MBR file. In earlier encoder versions, though, you cannot include multiple audio formats or video resolutions in a single MBR file. You
can achieve a similar result by creating several MBR files, which are based on the connection speeds of the clients you are targeting. For example, you can have a single MBR file targeted
at high-bandwidth connections (300, 500, and 700 Kbps) and another MBR file targeted at low-bandwidth connections (28.8 and 56 Kbps). Each MBR stream contains one audio track and multiple
video tracks. You can encode these streams simultaneously by opening two instances of the encoder on a single computer or by using two computers. You can also encode them consecutively. When
encoding simultaneously on a single computer, remember that your capture card can only be used by one process at a time, so you might want to have multiple capture cards.
Encoding content at multiple bit rates can place a heavy burden on your processor, especially when encoding in real time. If you are archiving your MBR stream, the file will be larger than
when encoding a single stream because of the multiple streams stored in the single file. If you are archiving for later, on-demand use, be advised that MBR files take a long time to download
and play.
Applying Filters
Taking video that was originally intended for television or movie theater screens and displaying it on computer monitors can create unique challenges. For example, displaying interlaced video
on a progressive screen results in combing, and flickering or streaking can result if the monitor's refresh rate is not synchronized with the video rate (60 Hz or 120 Hz for 30 fps video).
You can also see artifacts on monitors as a result of the telecine and editing processes. Freshly telecined content has an interlaced, coherent 3:2 pattern. Any video editing that is performed
after the film has been telecined can break the coherent 3:2 pattern, resulting in an incoherent source. Inverse telecine and deinterlacing filters are available to correct these problems.
You can apply filters before or during the encoding process.
The filter you apply should be determined by the source of the original content. For example, if the content was shot on 24 fps film and then telecined, apply the inverse telecine filter.
If the content was originally shot on video, apply the deinterlacing filter—inverse telecine will not be useful because this content does not have a 3:2 pulldown pattern. If you are
unsure about the source of your content, use a deinterlacing filter.
Inverse Telecine Filter
When a telecined source is displayed on a progressive screen, artifacts such as picture jerk or motion blockiness can result. To avoid or minimize these artifacts, you can apply an inverse
telecine filter to the content to remove the effects of the telecine process.
Inverse telecine solutions are available in both hardware and software forms. Hardware preprocessors and encoders produce very clean results. They are expensive, though, and cost may be prohibitive
for some users. Software-based inverse telecine filters are far less expensive than their hardware counterparts and generally work well for coherent sources. These filters do not handle 3:2
pattern incoherence as effectively.
In the hardware solution, you run the source through an inverse telecine preprocessor to produce a clean, coherent source. You can then capture the source to AVI or feed it directly into the
encoder. Keep in mind that video capture cards often support only NTSC 30-fps interlaced video. If you are using a hardware preprocessor, create a coherent telecined source for the capture
card, and then turn on the inverse telecine software filter in Windows Media Encoder. This filter should handle this coherent content easily. Alternatively, you could convert previously mastered
MPEG-1 files to Windows Media Format.
Inverse telecine filters are also available in some other providers' tools and in Windows Media Encoder. The Windows Media Encoder filter enables you to convert 30 fps video to 24 fps, with
the following qualifications:
You are applying the filter to an NTSC source that was previously telecined at 30 fps.
You are not capturing a live, uncompressed source.
You are not converting a Windows Media file with a .wmv file name extension to an .asf extension.
Your content does not have time compression applied.
You have captured both fields, 480 lines vertically, even for half-resolution video. (To create 320-pixel by 240-pixel content, capture at 320 pixels by 480 pixels. You can then set your
output video size in the encoder to 320 by 240.)
Applying this filter removes the extra frames that were added, returning the video to its original 24-fps format and reducing the picture jerk and motion blockiness that are artifacts of the
telecine process. You should apply this filter only to content that originated on film.
Deinterlacing Filter
Any content encoded from sources such as digital video cameras, VHS tapes, 8mm film, and cable television is usually interlaced at 30 fps, which creates artifacts on a progressive display. You
can use Windows Media Encoder to deinterlace the content under the following conditions:
Your source was previously interlaced.
You are not capturing a live, uncompressed source.
You are not converting a Windows Media file with a .wmv file name extension to an .asf extension.
Your content does not have time compression applied.
You have captured both fields, 480 lines vertically, even for half-resolution video. (To create 320-pixel by 240-pixel content, capture at 320 pixels by 480 pixels. You can then set your
output video size in the encoder to 320 by 240.)
For content that originated on videotape, apply the deinterlacing filter to achieve 30-fps full resolution (640 pixels by 480 pixels) or 60-fps half resolution (320 pixels by 240 pixels). Deinterlacing
blends the odd and even fields into a single frame at full resolution and creates one frame per field at half resolution.
For PAL video, simply apply the delinterlacing filter.
Mixed Content and 60 fps
Sometimes content is shot on film and edited in video, resulting in multiple original sources that can cause both telecine and interlacing artifacts on computer screens. Applying only one of
the filters (inverse telecine or deinterlacing) will eliminate some artifacts, but not all of them. To reduce the artifacts when encoding mixed content, you can encode the content at 60 fps.
Encoding interlaced content at 60 fps enables you to bypass any deinterlacing filter, but the resolution will be half the size of the original full-frame resolution.
During the encoding process, you produce two half-height, 30 fps frames that are joined together into a single frame and played at 60 fps. Because you are compressing twice as many frames per
second, your encoding computer must have enough memory and a fast enough processor to handle the load.
When encoding at 60 fps, your target bit rate should be at least 300 Kbps. To achieve this bit rate most effectively, make the following changes to your profile:
Increase the buffer size to more easily accommodate variable bit rates.
Increase the smoothness of your image quality.
Increase your key frame interval so key frames are generated only as needed.
When setting up your encoding session to encode mixed content at 60 fps, you can do one of two things:
Use a high-motion profile with an output size of 320 pixels by 240 pixels (or 640 pixels by 240 pixels) and 60 fps. You must capture at 480 vertical lines. Make sure the deinterlacing
filter is off. This solution addresses both the telecine and interlacing artifacts.
Use a profile with an output size of 640 pixels by 480 pixels and 30 fps. Make sure the deinterlacing filter is on. You may see some artifacts from the telecine process, such as picture
jerk, but the fields will be blended to eliminate interlacing artifacts.
Audio and Video Levels
The audio and video devices that you are encoding from and the audio and video settings on your computer should be set properly to minimize artifacts in encoded audio and video streams.
When setting up your audio source and the audio settings on your computer, keep the following tips in mind:
If the audio is going from a balanced output to an unbalanced input on a sound card, make sure you use an audio adapter to isolate and match the input and output.
Adjust the line input audio level in your system to avoid distortion. Some audio cards include an audio meter, which can show you when your recording level is too high or too low. If the
level is too high, digital distortion can result. If the level is too low, noise in the audio system will become more audible. To determine whether you have an audio meter, double-click
the Volume icon in the notification area to display your audio mixer properties.
Normalize the audio levels on your source by using another provider's application such as Sound Forge by Sonic Foundry, Inc.
When you set up your video source, do the following:
Adjust your video monitor using SMPTE color bars.
Adjust your computer monitor to match your video monitor using a high-resolution bitmap of the SMPTE bars.
Adjust the brightness, contrast, saturation, and hue on your capture card so that it matches those on the video monitor.
For more information about using Windows Media Encoder, see Windows Media Encoder Help. For more information about Windows Media technologies, see the Windows
Media page.
This software is based in part on the work of the Independent JPEG Group.
GIF decompression code, copyright 1990, David Koblas. Permission to use, copy, modify, and distribute this software and its documentation for any purpose and without fee is hereby granted,
provided that the above copyright notice appear in all copies and that both that copyright notice and this permission notice appear in supporting documentation. This software is provided "as
is" without express or implied warranty.
Back to the top
Back to the top