A lot of people I speak to would like to use the inking features of Windows 8.1 and their surface – especially the Ink-to-Text feature of OneNote however they struggle to make the process efficient for them and get frustrated when things don’t work out as expected. One of the biggest frustrations is not having your handwriting properly recognized by OneNote.
I’ve spoken about the feature in Episode 4 of the Surface Smiths Podcast and in Episode 5, I speak about the Surface App that you can use to customize the pen sensitivity to better suit your writing style and in Episode 6 6, I speak about how to train your surface to better recognize your handwriting.
This blog post provides step by step instructions on how to personalize the handwriting recognition in Windows 8.1 and the MS Surface Pro. These instructions should also work on the Surface 3 however I haven’t tested them.
- You already know how to use OneNote and how to use the Ink-to-Text Feature. If you don’t, please refer to Surface Smiths Podcast – Episode 4 or the MS Surface FAQ
- You have a working Surface Pen that is synced to your device
- You have about an hour to go through all of the training. Don’t worry if you run out of time, you can save and complete it later if you need to.
- Open Control Panel
- Select Language
- Select Options beside the language that you will be writing in (easier if you only have one language installed)
- Select Personalize handwriting recognition (If you have more than one language installed, you may get prompted to Choose the language for handwriting recognition personalization)
- Select Teach the recognizer your handwriting style
You will be prompted to train on both Sentences and Numbers, Symbols and Letters through a series of wizards. Get to work!
The sentences training panels will look similar to this based on your language selection:
Notice that there are fifty screens to complete
The Numbers, Symbols and Letters training panels will look similar to this:
Notice that there are nine screen to complete
IF things don’t work out perfectly, you can fine tune the recognition for specific errors or restart the training from scratch.