Does ACR Cloud timestamp when it first detects a new song or after it has recognized it? Our music programmers would like to know. Thanks!
1 Like
Neither. It’s devious and possibly a bad design on our part (not ACRCloud’s).
We use the start time of the recording calculated back from the recognized segment.
The usual scenario goes like this:
- You start playing Song X
- Between 0 and 10 seconds from that moment ACRCloud begins a new 10-second sampling window.
- When that window closes, ACRCloud uses its fingerprint magic, recognizes and sends results to Spinitron, all of which takes a little time that’s variable.
The results we get include the timestamp of the start of the recognized segment and the offset in seconds of the start of the recognized segment in the recording ACRCloud thinks it comes from. We do a calculation to get the timestamp of the theoretical start time of your playout of the recording.
This relies on the debatable assumption that you started playing Song X from the start. But it goes some way to compensate for the variable delays in getting results, which may also be affected by segues and talking over the start of the recording, etc.