The art of DVD encoding - Part One
Part One - Background
Following a trend of minimalist video editing applications, lower end DVD authoring packages produce DVDs with the minimum of user input. Although fantastic for the beginner, this design feature can to take far too much away from the more demanding end user. Two cases in point are opposite ends of an extreme: the user wanting to create the best possible image quality; and the user aiming to cram in as much video as possible on to a single DVD.
Despite these limitations, DVD authoring applications combined with more powerful MPEG2 encoders give far more flexibilty to the budding videographer. Be warned that your DVD authoring application should be capable of importing MPEGs – and author the DVD without re-encoding your expertly encoded video. This issue is addressed later with a list of applications meeting this criteria.
Before explaining the process of DVD creation it's worth explaining the fundamentals of DVDs. The typical commercial DVD contains over two hours of high quality video on a dual layered DVD. You may have noticed a slight "jump" in the middle of a film - this is where the laser jumps from reading the first layer of data to a second. Alas, this method is not available to the home user and DVDs created on home PCs are limited to the physical data limit of a single layered DVD (approximately 4.37Gb).
Update: Dual Layer DVDs are now widely available, but require a compatible dual layer burner to burn a Dual Layer disc. You can therefore fit two hours of maximum quality on a dual layer DVD.
So how does this affect the amount of video on a DVD? Well, video encoding coverts the picture into data and then compresses this data by using a technique of comparing one frame of video (frame A) to the next (frame B), discarding any data duplicated in frame B. More specifically, MPEG files have a "Group of Pictures" (GOP) structure, consisting of complete (intra) frames "I", predicted frames "P", and bidirectional frames "B". Predicted frames are based on past frames, bidirectional based on past and future frames. Secondly, the amount of data consumed every second by is known as the video bitrate: the higher the bitrate, the more data per second and the higher the quality - you get less video on a DVD the higher the quality of the picture.
The maximum bitrate under the DVD standard is 9800kbits/sec (this includes both video and audio). It’s important to adhere to this standard to ensure the created DVD plays in standalone players. At this constant maximum bitrate, you’ll get 60mins of video on your DVD. As we know, the lower the bitrate, the lower the image quality but the more video we can fit on a DVD. So if we reduce the bitrate by half, we can get 120mins of video on a disc. A maximum Constant or Variable bitate should be used to encode videos less than 60mins long - this will maximise video quality by ensuring no space is wasted on the DVD.
Notice the use of the word “constant” above. This is important: an alternative to constant bitrate (CBR) is variable bitrate (VBR). When you’re encoding more than 60mins, 2-pass VBR should be the preferred alternative. The encoder runs through the video twice, the first “pass” analyses the data, the second actually encodes the video, using a higher bitrate in scenes with high motion. This will then “smooth” your video and should achieve better quality video compared to CBR for videos greater than 60mins.
If you do the maths, you’ll see that these bitrates don’t quite add up – and that’s because the audio is treated separately to the video. Another DVD standard is PCM audio, this is uncompressed audio with a datarate of 1536kbits/sec. This may seem high – and it is. Although not part of the DVD standard, you can compress the audio to squeeze more video onto your disc! If we use MPEG layer II audio at a bitrate of 224kbits/sec, we can either slightly increase the video bitrate or fit an extra 12 more minutes of video onto the disc. AC3 audio would work in a similar vain.
So there you have it: we can alter the constant bitrate of the video to control either quality and the amount of video we fit on a DVD. We can also use a variable bitrate to maximise overall quality when using a lower average bitrate and we can then compress the audio to further control the amount of video we can fit into a DVD.
Part two: using TMPGenc to encode DVD compliant video…