Playback & Transcript Sync
Why the transcript can drift from the audio you're playing: dynamic ads (DAI), variable bit rate (VBR) and non-standard bitrates.
The episode page links here when the audio you're playing may not line up with the transcript. There are three causes.
What is Dynamic Ad Insertion (DAI)?
DAI is a technology that allows podcast creators to swap ads in podcast audio, playing different ads for different users.
Ads are a major source of income for podcast creators, so Podwise plays the original audio for DAI to work properly.
However, DAI can make the audio being played differ from the audio Podwise transcribed. So Podwise has to find the differences between the two and align the playing audio with the transcript.
How does Podwise handle DAI?
Whenever an episode is transcribed, Podwise first pre-samples the audio, then deletes the audio used for transcription. When you start playing an episode that has been processed, Podwise immediately begins sampling the audio you're playing and compares it with the pre-sampled results to detect where the dynamic ads are. Once detection is complete, Podwise adjusts the outlines, highlights and transcript to match the audio you are playing.
While the audio is being aligned, highlighting is turned off, and the parts that differ are marked Dynamic Content in the transcript.
What is Variable Bit Rate (VBR)?
Variable Bit Rate (VBR) is a technology that compresses audio and video files by encoding different positions at different bit rates. A problem with it is that a player can't seek accurately to a specific position; it can only approximate it.
Unfortunately, there is no good solution to this yet. So when you start playing a VBR episode from the middle, the transcript can be off by a few seconds to tens of seconds. This doesn't happen if you start playing from the beginning.
If possible, you could ask the podcast host to consider using CBR (constant bit rate) encoding for their episodes to avoid this issue.
Why does the transcript drift away from the audio?
Some audio files don't follow the standard bitrate when exported, and don't add the padding their frame data needs. This makes the actual duration of the audio differ from the duration the player sees, so the player can't seek accurately to a specific time.
Podwise adjusts the transcript timestamps to match the player's time as closely as it can, but the adjustment isn't always accurate. When you start playing from a specific sentence, Podwise aligns the transcript with the player correctly. As playback continues, though, the transcript gradually drifts away from the player's time.
When the drift becomes too large, click Sync or the sentence that is currently playing to realign the transcript with the audio.