TMPGEnc Settings - Tips and Detailed Information
The Setting dialog box allows you to alter the encoder parameters. It includes the following tabs:
Select from one of the following:
This option sets the size (resolution) of the video pictures that you are encoding to in pixels across the screen by number of lines.
If you are encoding a VCD then you must set these values to 352 by 240 (for NTSC) or 352 by 288 (for PAL).
For SVCD choose 480 by 480 for NTSC or 480 by 576 for PAL.
This is the ratio of width to height that the encoded video will be. This information is present in the output video stream and used by the decoder to display the video at the correct aspect ratio. The following settings are possible:
This is the number of video frames (complete pictures) that will be presented to the viewer each second. When producing MPEG streams for playback on TV it is important to get these numbers correct. Remember that TVs use interlacing to give twice as many fields per seconds as frames per second, so PAL TV has a frame rate of 25fps, NTSC TV has a frame rate of 29.97.
When playing back on a PC, any frame rate can be used. Lower frame rates will require less bits to encode, but will show jerkier motion.
Rate control mode
This option allows you to choose the type of bitrate control you wish to use. The bitrate control varies the bitrate of the video stream, depending on the options you choose. Different control modes have advantages and disadvantages over one another.
If you are creating an XVCD or SVCD, the bitrate calculator will help you determine the optimum bitrate for encoding your video.
Constant bitrate (CBR)
The video stream bitrate is fixed to a constant value. This means that high motion scenes are allocated the same number of bits as low motion scenes. Either bits are wasted during the encoding of low motion scenes, or block artifacts will be seen during high motion scenes.
Use this setting to create a VCD.
2-Pass variable bitrate (VBR)
Variable bit rate does exactly what it says on the tin - the bit rate is varied. During periods of high motion the number of bits used to encode the video is increased. When the action stops the number of bits used to encode it is reduced. This gives the best video quality for the least average number of bits.
To work out which parts of the video should be allocated the most bits the encoder processes the whole video clip twice (hence the "2-pass" part of the name). During the first pass it keeps a record of the complexity of each frame. Highly complex frames will be allocated more bits. The second pass does the actual encoding.
Use this setting to create high quality SVCD of DVD videos automatically.
Manual VBR (MVBR)
This setting allows you to manually select different bitrates for different scenes of your video. This is done using the "Force picture type setting" option on the "GOP structure" tab. The video bitrate remains constant between your manual settings, so its more like CBR than VBR.
Only use this setting if you have problems encoding short, tricky video sequences and have a lot of time to spend fiddling with the manual settings.
Automatic VBR (CQ_VBR)Single pass encoding is used which automatically adjusts the video bitrate up during high motion sequences and down during low motion sequences.
Constant Quality (CQ)Single pass encoding that varies the bitrate, but attempts to keep the perceived quality of the final video the same.
The "Setting" button opens up the variable bitrate setting dialog box. Depending on the Rate control mode setting, this dialog box has different fields to fill in.
Constant bitrate (CBR)
2-Pass variable bitrate (VBR)
Manual VBR (MVBR):
Automatic VBR (CQ_VBR) and Constant Quality (CQ):
This field is only active when CBR is selected. It allows you to specify the bitrate without opening the Setting dialog box.
VBV buffer size
This value specifies the size of the decoders "Video Buffering Verifier". It represents the amount of coded video data that can be buffered by the decoder. At constant bitrates the buffer allows best use of the MPEG compression techniques. It is filled at a constant rate (that of the video stream) and partially emptied when a frame is decoded. The buffer fills up during sequences that compress well and empties during sequences that do not compress well.
Usual values are 40 for producing an MPEG-1 VCD and 112 for an MPEG-2 SVCD. These are the minimum allowed by the specifications, but if you are encoding for a particular player it may have a larger VBV buffer, so these could be increased (producing XVCD or XSVCD discs).
Profile & level
The MPEG-2 specification defines a number of profiles and a number of levels. The most commonly used combination is Main Profile at Main Level (MP@ML). MP@ML is used for the VCD, SVCD and DVD standards.
Unless you know what you are doing, leave this alone.
Set the video format to that of your source material.
This allows you to select how the video will be decoded on playback. Choose from the following options.
This allows you to specify the precision with which the luminance (brightness information) and chrominance (colour information) signals are encoded. The eye is more (spatially) sensitive to brightness changes than colour changes, therefore fewer chrominance samples than luminance samples are required.
MP@ML only allows the use of 4:2:0. Unless you know what you are doing leave this set to 4:2:0.
DC component precision
The DC component represents the average brightness of a block (8x8 pixel region). In MPEG-1 video this is fixed to 8 bits, but MPEG-2 allows higher precision. It is important that the DC component is accurately represented, so a setting of 10 is recommended for MPEG-2 sequences, unless you are encoding using a low bitrate.
Motion search precision
Select the precision with which the encoder tracks movement between consecutive video frames. The better the search, the longer it will take. I tend to choose either "High Quality". Many people report that "Highest Quality" provides no visible improvement, but dramatically increases the encode times. Use low quality settings only for test runs or if you have a very slow PC.
Video source type
Choose the appropriate setting to match the source material:
When interlaced source material is used this option selects which field of each video frame is presented first. Choose from:
If your source material comes from a DV camcorder, set this to "Bottom field first". If in doubt, try using "bottom field first". If you find your video flickers on playback it may be that you have selected the wrong option here.
It is possible to determine the field order before encoding by selecting one and then opening the deinterlace filter window. Once there move the slider across to a part of the source with motion in it and press the scroll button. If the image is jerky then the wrong field order is selected.
Source aspect ratio
This sets the width to height ratio of the source material. If your source is actually larger or smaller that the setting it will be cropped or padded with black pixels to fit. Use one of the following options:
If you are creating an SVCD from 704x480 or 704x576 source video you'll need to ensure that an 8 pixel padding region is added each side of the video before encoding. The easiest way of ensuring this is to set the source aspect ratio to 720x480 or 720x576. This is required to maintain the aspect ratio of the video when scaled to 480 pixel wide resolution. The 480 pixel wide video is scaled back to 720 pixels by the decoder.
Video arrange method
Various filters are available that have an effect on the encoding process.
This filter allows you to specify a start and end video frame for encoding. Also a time delay offset to the audio track can be added, allowing lip synchronisation to be achieved.
This setting undoes 3:2 pulldown. If you have NTSC format video (30 fps) from a film source This will allow the original film framerate (24 fps) to be recovered.
Ghosting is an artifact introduced on analogue signals when a reflection or duplicate transmission path is present. It produces a "ghost" image shifted slightly left or right of the main image. Enable this filter to reduce the ghost image.
Analogue video signals are subject to noise. If you tune your TV to a weak station you'll see a lot of noise on the picture. If you have source material with visible noise, try this filter out.
Noisy source material not only looks bad, but the encoder will spend precious bits faithfully encoding the noise! Using this filter should help reduce the bit requirements of the encoder.
If you have a noise free picture this filter may introduce some blurring in the video, so disable it.
If your source material is blurred or out of focus, this filter may help to bring some better definition to the image by sharpening the horizontal and vertical edges. Don't expect too much though.
This filter can also be used to soften edges by specifying negative filter values.
Simple colour correction
This filter can apply simple colour correction to you video sequence, allowing you to specify brightness, contrast, gamma, red and blue adjustments.
Custom colour correction
This filter allows complex colour adjustments. Several different filters can be applied in sequence. Try this filter if you cannot achieve the desired results using the simple colour correction filter.
This filter allows interlace to be removed from the video sequence.
This filter allows a clipping region to be specified on the original source video. As a result only a portion of the video will be encoded with a fixed border round the edge.
Converts from film framerate (24fps) to NTSC video framerate (30fps).
Do not (do) frame rate conversion
This prevents frame rate conversion. If the source material framerate does not match one of the MPEG frame rates, the video and audio will not be synchronised.
Allows audio to be adjusted
GOP is an MPEG term meaning Group Of Pictures. It is a collection of consecutive frames of video. Usually between 0.5 and 1 second of video will be held in 1 GOP. Each picture within the GOP can be 1 of 3 types:
A GOP always starts with an I picture.
Number of I, P and B picture(s) in a GOP
These 3 options control how many of each frame type are normally inserted into each GOP. Normal values would be:
Output interval of sequence header
Sets how often the video sequence header is sent. Normally you would send this once per GOP to enable decoding to resume at any point. If you are using low bitrates you could increase this value so that the sequence header is only sent say once every 5 GOPs.
MAX number of frames in a GOP
This option allows you to limit the maximum number of frames in a GOP. If the encoder reaches this number of frames a new GOP is started.
Output bitstream for edit (closed GOP)
If you wish to edit the encoded video stream select this option. It removed inter-GOP dependency. Some DVD players are reported to introduce artifacts when fast forwarding video without closed GOPs.
Detect scene change
This option allows the encoder to look for sudden changes in the video, adding a new I-frame when they are detected. This option should usually be selected, however if you have high motion video sequences it can introduce more I-frames than necessary, lowering the overall picture quality.
Force picture setting
This option allows manual setting of many of the encoder options on a per- frame basis. It shows each video frame in a scrolling window. Right click on a frame to change it settings for that frame. As you can imagine, this is time consuming work. If you have short (a few seconds) video sequence and have problems encoding it you could try this. Defiantly not recommended for beginners.
I picture only
Forces the encoder to produce a sequence of I frames only. Since the sequence contains only I frames each frame can be easily edited after encoding since there is no dependency on other frames. Also motion artifacts will be completely eliminated. Since the video sequence will not make use of compression provided by P and B frames a massive bitrate (perhaps 20Mbps) will be required to make the image look acceptable.
Don't use this option unless you know what you are doing.
Forces the encoder to produce a sequence containing only I and P frames (no B frames). Since the video sequence will not make use of compression provided by B frames a large bitrate (perhaps 5Mbps) will be required to make the image look acceptable.
Don't use this option unless you know what you are doing.
The default setting, using I, P and B frames. This obtains maximum compression of the video.
Output YUV Data as basic YCbCr not CCIR601
This affects the range of values accepted for luminance and chrominance signals. YCrCb is raw samples, taking the range 0-255. In TV systems CCIR601 is normally used, which restricts the allowed range to 16-235 for luminance and 16-240 for chrominance signals. If you are mastering for TV (e.g. VCD, SVCD, DVD) then do not tick this option.
Use floating point DCT
The discrete cosine transform (DCT) is part of the encoding mechanism for compressing MPEG video. There is a floating point implementation and an integer (fixed point) implementation. The floating point DCT implementation introduces less error than the fixed point implementation, but takes longer. For Pentium-IV (SSE) systems this option is always enabled.
No motion search for still picture part by half pixel
This option prevents the encoder from looking for motion changes of half a pixel between consecutive frames. As a result, still picture video sequences may look better. This option should not be used for normal video sequences.
Soften block noise
MPEG uses blocks of 8x8 pixels to encode video sequences. Block noise is evident when the bitrate is too low to accurately encode the video. If you experiment with low bitrates you'll quickly see the effect - the picture becomes blocky. This is because high spatial frequency information is reduced. Enable this option if you are encoding at low bitrates or you see blocking artifacts in your encoded video. If you are using high bitrates and do not see block artifacts, disable this option.
MPEG-1 specifies 3 "layers" for audio information. They all operate in a similar way, with different levels of complexity in the encoder and decoder. In each the audio is split into 32 sub bands and a psychoacoustic model is applied to decide what audio information can be discarded with the least impact on the listeners perception.
Conventional MPEG decoders can only decode layers 1 and 2, so these are the only ones supported by TMPGEnc. Layer 3 (also known as MP3) encoding and decoding is significantly more complicated. Some decoders support layer 3 audio, but this is not (currently) supported by TMPGEnc.
Select the sampling frequency that you wish to encode to from one of the following options:
The average human ear can hear sounds in the frequency range 20-20000Hz. To reproduce a sound at a given frequency the audio sampling rate must be more than twice that frequency. This means that with a sampling frequency of 32000Hz (32kHz) some high frequency information will be lost. The 2 higher sampling frequencies (44100Hz and 48000Hz) have cut-offs above normal human hearing range.
For higher sampling frequencies the audio bitrate must be increased, else the benefit of using the higher sampling frequency will be lost due to an increase in audio artifacts.
Select the audio bitrate (in bits per second) from the drop down selection. As a rough guide for layer II stereo audio, 64kbps will give AM radio quality audio, 192kbps will give FM radio quality, 384kpbs will give CD quality audio.
VCD1.0 and VCD1.1 require an audio bitrate of 224kbps. VCD2.0 allows bitrates of 128, 192, 224 or 384 kbps for stereo streams or 64, 96 or 192 kbps for mono streams.
Includes CRC checks into the audio data, allowing the decoder to perform error checking. If you are recording to CD, DVD or hard drive, error checking elsewhere in the system makes this option redundant. Use it if you are transmitting using an unreliable media.
This is a single bit flag carried in the MPEG audio stream for information only. It does not affect the encoder or decoder. It simply indicates that the material is original.
This bit indicates that the audio material is protected by copyright. Decoders should look at this bit and not allow the stream to be copied if this bit is set.
The private flag can be switched on or off and is not used by the encoder or decoder. In clever systems it could be used to trigger certain events, or even a low data rate stream.
A pre-emphasis filter may have been applied to analogue recorded audio to change the frequency response characteristics. This option indicates that a similar de-emphasis filter should be used in the decoder to reproduce the original audio. With digitally sampled streams this is not necessary, and the emphasis setting should be set to None.
Select one of the following types:
This option allows comments to be inserted into the video stream that you are creating. The comments will be ignored by the decoder. For instance you could include a copyright message using this option.