True confession: I’m a student again. But this time, it’s not at Penn.
I’m back in the saddle writing papers and cranking out projects, and I’ve been thinking how nice it would be if my computer could read my MS Word drafts to me while I chop away at carrots, making dinner. If NPR can talk to me, why can’t my own paper?
So I started looking for text-to-speech apps that work with MS Word. First I ran across an add-in called WordTalk, and I’m here to tell you: it is awful. Just don’t.
Then I discovered that MS Word has its own text-to-speech tool buried in its toolkit as well. To try it out, add Speak to your Quick Access Toolbar, as shown:
This is a step up from WordTalk, but it isn’t as smooth as my favorite: SoundGecko. But the trouble is, SG is a web browser plugin, so it’s not right there inside MS Word.
Then it dawned on me. When last I test-drove the app, the company was close to adding PDF support. A Word document can easily be converted to a PDF. I checked in with SoundGecko again, and hooray! PDF support exists – but for a small price. Only $2.95.
SoundGecko provides a peek at this functionality, so I gave it a shot. It’s great. Here’s the text I fed it and the recording it delivered:
Text sample: “The Constitutionally mandated collection of seemingly simple census data – just counting people – depends on a highly accurate sense of geography through the lens of the location of where people live. This rich, accurate trove of geospatial data is valuable in itself, regardless of the actual counts or attributes of the households surveyed. In fact it is so complex, it exists on its own, without any census data mixed into the database. The MAF/TIGER database contains these geographic data – and only that data – in vector data format. Vector data are the building blocks of shapes in space: nodes, edges and faces. Nodes are points, defined by a latitude and longitude pair of numbers. Edges are lines, used to delimit linear features like railroads and highways. Edges have start and end nodes, which give a line direction (from start node to end node). A single street according to human experience may be composed in the MAF/Tiger database of many edges chained together block by block, and given a single label. Even more specifically, street segments are encoded with a right to left address range, and a left to right address range – quite a useful feature for a census worker in the field! The last sort of “geometric primitive” stored in the database is a face. Faces are polygon surfaces defined by bounding edges – for example, a census tract bounded by streets”
Compare that to the awful recording of the same text sample WordTalk gave me:
The difference is night and day. The $2.95 version also comes with
email-reading (in the works), choice of preferred voice, support for 30,000 words per document, and RSS-to-voice support. Moral of the story?
Let us know if there’s anything else you like out there for speeding up your draft-proofing process!