Why VOB Files Have Negative Audio Delay in Authoring Logs

When demuxing or authoring DVD-Video assets, processing logs frequently report negative audio delay values (such as -80 ms) for extracted VOB streams. This article explains why these negative values appear, how Presentation Time Stamps dictate synchronization within MPEG-2 program streams, and why encoders intentionally introduce audio offsets to guarantee smooth playback across standalone hardware players.

Understanding Audio Delay in VOB Containers

A VOB (Video Object) file is a specialized implementation of the MPEG-2 Program Stream (PS) container. It bundles multiple elementary streams—primarily MPEG-2 video and AC-3, DTS, or LPCM audio—into a single time-multiplexed file.

Each elementary stream contains its own timing markers known as Presentation Time Stamps (PTS). The PTS instructs the decoder precisely when a specific frame of video or frame of audio must be presented to the screen or speakers.

DVD authoring and demuxing tools calculate audio delay using a simple relative formula:

\[\text{Audio Delay} = \text{Audio PTS} - \text{Video PTS}\]

When an authoring log displays a negative audio delay, it means the first audio packet has a lower PTS value than the first video packet. In practical terms, the audio track begins chronologically before the video track.

Primary Causes of Negative Audio Delay

1. Decoder Buffer Pre-Filling (System Target Decoder)

The DVD-Video specification relies on the MPEG System Target Decoder (STD) model, an idealized hardware architecture with dedicated buffers for video, audio, and sub-pictures.

Video frames require significantly more buffer space and decoding time than audio packets—especially when dealing with bidirectional predictive frames (B-frames), which require the decoder to reference past and future frames before rendering. To prevent the audio buffer from starving while the video decoder initializes and decodes its first I-frame, encoders offset the streams. Audio packets are introduced into the stream early, yielding a lower initial PTS.

2. Frame Duration Discrepancies

Audio and video do not share uniform frame durations:

Because audio and video frame boundaries rarely align perfectly at time zero, authoring tools must offset one of the streams to achieve synchronization. Encoders frequently start an AC-3 frame slightly ahead of the first video frame, resulting in a fractional negative delay equal to a partial audio frame.

3. Stream Trimming and Editing (Open GOPs)

If a video source is cut or split at a point that does not coincide with a sequence header or an intra-coded frame (I-frame), the initial frames might be discarded during authoring because they cannot be decoded independently. If the corresponding audio packets are not trimmed with frame-accurate precision, the remaining audio stream will start before the first decodable video frame.

How Hardware Players Handle Negative Delay

Negative audio delay rarely causes synchronization issues on standalone DVD players. The DVD player's demultiplexer reads the PTS of every packet. When it encounters audio packets with a timestamp earlier than the initial video PTS, it simply decodes the audio and discards the preliminary samples (pre-roll), waiting until the audio PTS matches the video PTS before sending the signal to the output speakers.

Handling Negative Delays During Re-Authoring or Transcoding

While DVD players resolve negative timestamps transparently, modern containers (like MP4 or MKV) handle delays differently: