W3C home > Mailing lists > Public > public-sdwig@w3.org > July 2019

Re: [sdw] Video Metadata Question: Use cases & requirements for interpolation schemes (#1136)

From: Rob Smith via GitHub <sysbot+gh@w3.org>
Date: Mon, 22 Jul 2019 15:25:31 +0000
To: public-sdwig@w3.org
Message-ID: <issue_comment.created-513839986-1563809130-sysbot+gh@w3.org>
@chris-little Thanks for [your comments in #1130](https://github.com/w3c/sdw/issues/1130#issuecomment-511751960), which raise a couple of points.

WebVMT is designed to store the original values in the file, so no reverse calculation would be required. Interpolation schemes are defined in the [data model](https://w3c.github.io/sdw/proposals/geotagging/webvmt/#datamodel) to ensure consistent interpretation, though intermediate values are not explicitly recorded in the file. For example:

NOTE Interpolate cam1 position from (0.3, 51.1) to (0.2, 51.5) over 1 minute

00:00:00.000 -->
{"moveto": {"lat": 51.1, "lng:" 0.3, "path": "cam1"}}
{"lineto": {"lat": 51.5, "lng": 0.2, "path": "cam1", "end": "00:01:00.000"}}

Data should be sampled at a sufficiently high rate and resolution to accurately represent the characteristics of interest. For example, taking one photo per minute of a football match will give a fair record of who played, but not of who scored. The purpose of interpolation is not to 'fix' poor quality data, which is why I'm interested in identifying valid use cases.

Is there a drone/wind use case where an interpolation scheme would add legitimate value to the data, rather than just masking a quality issue? This could make a case for including suitable features, but my concern is that the underlying problem is out of scope.

Finally, there is nothing to prevent users applying their own interpolation scheme, which is not included in the specification, if they believe it will add value to their own use case.

GitHub Notification of comment by rjksmith
Please view or discuss this issue at https://github.com/w3c/sdw/issues/1136#issuecomment-513839986 using your GitHub account
Received on Monday, 22 July 2019 15:25:32 UTC

This archive was generated by hypermail 2.4.0 : Thursday, 24 March 2022 20:31:01 UTC