Tuesday, June 12, 2007

OQO Model 02 Review Part 2: Pen Navigation, Handwriting and Voice Recognition

This is part two of my OQO Model 02 review. In Part One, I discussed the OQO's general form factor. As with all my reviews, Part Two will focus on usability, user interface and user experience as it relates to pen inking and voice recognition.

The unit I'm reviewing is the OQO model 02 with Vista Business installed. It's amazing how much you can get done on the OQO with Vista installed without needing to use the keyboard or mouse. In fact, this entire review is being done using voice recognition only. Even my outlines and notes were dictated using voice recognition.



Pen Navigation

While the OQO offers a superb Thumbstick for navigation, I prefer to use the pen and the built-in tablet navigation features of Vista and Internet Explorer seven. First let me talk about why I prefer pen navigation to the built-in keyboard and Thumbstick. First, I can hold the unit naturally. To use the Thumbstick I have to have the screen in its up right or open position and I have to hold the unit a certain way: my right thumb must control the Thumbstick and my left thumb must control the left and right mouse buttons. As I mentioned in part one of this review holding the OQO in landscape mode is just not as easy as holding it in portrait mode. So by using a pen I can close the screen and hold the unit in any way that feels natural to me including not holding the unit at all like laying it down on a flat surface or on my lap.

The OQO sports an active digitizer vs. what is commonly known as a touchscreen or passive digitizer. A touch screen allows you to use anything including your finger to touch things on the screen, drag-and-drop, push buttons. The disadvantage of a touchscreen is that you have to actually drag your stylus or finger across the screen to move the cursor. Because a touch screen can't differentiate between a pen and a finger it also can't differentiate between your hand and your stylus. So, when writing on a touch screen you have to be very careful not to let your hand touch the screen. Otherwise the touch screen driver will get confused and think your hand is your pen and vice-versa and you'll end up with a muddled jumble of scribbles on the screen.

The OQO's active digitizer solves these problems. The OQO's digital pen emits a magnetic signal that the active digitizer can track and read. This means you can hover the pen about a half inch above the screen and the OQO will read it. Move the cursor on the screen by simply pointing with the pen tip without actually touching the screen. This is more natural because when you want to group select you can simply press the pen down on the screen draw a box around your selection and release the pen. Very natural and simple. It also means you can hover the pen above a button or control to bring up its tool tip. these behaviors are impossible with touch screens.

Vista has two builtin features: Pen Flicks, and Grab-and-Drag that when combined with the OQO's active digitizer make pen navigation natural and my preferred navigational method when untethered from the keyboard and mouse.

Pen Flicks allow you to do basic navigation like scroll up, scroll down, page back and page forward in the browser and edit commands like copy paste and undo by simply flicking the pen in a certain motion. For example, to page up in any document simply flick the pen in an upward motion. To page down, simply flick the pen in a downward motion. To page back in the browser, flick the pen to the left. The gestural motions are very intuitive, simple and execution is fast. It works much better than using the Thumbstick to move the cursor up to the back button or over to the scroll bars.

Both the Internet Explorer Seven and Outlook 2007 incorporate a great tablet feature called Grab-and-Drag. Here's how it works: You put Internet Explorer Seven or Outlook 2007 into Grab-and-Drag mode (by clicking a button). Now the cursor looks like a hand. When you press the pen down the hand grabs the page. You can now pan the page by simply moving the pen around.

Grab-and-Drag and Pen Flicks work well together. I use Pen Flicks to quickly get to a part of a scrollable view. Then I use Grab-and-Drag to fine tune my position. I may use Pen Flicks to quickly get to a paragraph at the end of a document and then use Grab-and-Drag to scroll the paragraph as I'm reading it. The two work seamlessly together and are very natural to invoke and use.

Occasionally I will use the OQO's builtin Thumbstick and mouse buttons the pen navigation with Pen Flicks and Grab-and-Drag feel more natural and comfortable and frankly is faster.



Handwriting Recognition

Handwriting recognition on the OQO model 02 can be smooth and very accurate but you have to know how to set things up correctly. The most important thing to remember is that speech recognition consumes lots of processing power. The Via C7 - M is incredibly good at power management. This means it steps down processing power whenever it can. Unfortunately handwriting recognition can suffer if the Via steps down the CPU and there aren't enough CPU cycles left for the handwriting recognizer. If the OQO's processor is anywhere between sixty percent and one hundred percent handwriting recognition accuracy is good. Anything under 60% and you're likely to see handwriting recognition errors.

There are two problems that occur when the processor is running at under 60%: First, the text input panel (T IP) is slow to start recognition. If you're not expecting this, you'll begin to write but the T IP won't show your digital ink. Thinking you've done something wrong, you'll probably go back and rewrite what you thought the TIP missed. Unfortunately, the TIP did not miss what you wrote it simply was lagging. Now, as you're rewriting the T IP tries to catch up with you and you end up with a jumbled mess that is generally misrecognized as a complete non sequitur.

A similar problem can creep up in the middle of writing in the TIP. Everything may be fine until a background service or process kicks in and the TIP is left with no CPU cycles. Suddenly, you run out of digital ink and quite naturally go back and rewrite what didn't show up. Once again, the TIP catches up with you when CPU cycles become available to it and you end up with a jumbled mess.

After some experimentation I found the 60% CPU threshold is a good minimum when running on battery power. Create your own power scheme and set the maximum CPU to 60% when running on batteries. Now you can disconnect and take notes in meetings to your heart's content.

The other thing you'll want to do when you get your OQO is disable the pen buttons on your Wacom pen. Why? Because you will inadvertently hit them while writing and trigger a right click. Dreadful. I've enabled the eraser mode so I can flip the pen upside down and scrub out mistakes like I would with a pencils eraser. Call me old school but 12 years and a public school and five years at an engineering school are hard to break.



Voice Recognition

I was fortunate enough to work at Apple's Human Interface Group, a part of the Advanced Technology Group awhile back. My team focused on speech recognition and synthesis so voice recognition is something that's near and dear to my heart. Back then I had a $14,000 DEC speech synthesizer that sounded like Colossus from the Forbin project and a dedicated speaker dependent voice recognizer box that was 10 times the size of the OQO and couldn't do dictation. Man, have things changed.

The speech recognizer in Vista is superb. It requires no training (although training will dramatically improve recognition) and is speaker independent.

It's amazing sitting here with no keyboard or mouse in front of me. As I mentioned before this entire review from start to finish has been done using only voice recognition. I find voice recognition is much faster and oftentimes more accurate than handwriting recognition; which is faster and more accurate than typing on the built-in keyboard on the OQO (at least for me).

Running voice recognition on the OQO will take between sixty percent and seventy percent of the CPU. You can use voice recognition in power saver mode, when the maximum CPU is set to 50%, but you will notice a slowdown in recognition. Accuracy will not be affected.

Do not attempt to use the OQO's built in microphone for speech recognition. It simply does not have the fidelity for good recognition. Go out and buy yourself a good USB microphone and you won't be sorry. I use the logic tech USB microphone -it's good quality and doesn't look too dorky.

To show you how big of a dork I really am, I also purchased NaturalReader 2007 a speech synthesis package for Vista. The $39.00 download adds toolbars to Outlook 2007, Word 2007 and Internet Explorer Seven that allows any text in those programs to be read aloud. Not terribly useful but comes in handy I want an Email or an RSS feed read to me while I'm working on something else.

What I'm really interested in are how alternate forms of input and output can enhance the ultra mobile experience. Keyboards and mice are great when you're sitting at a desk but the whole purpose of the ultramobile platform is to be mobile -in the field, on a factory floor, in a car. In these environments voice commands and dictation and speech synthesis could provide solutions for interesting new use cases.

Imagine a mechanical engineer in the field examining an oil pipeline. As the engineer examines the pipeline she enters her notes into a special application using dictation. A special language module helps the application recognize engineering terms specific to this task. The application accepts voice commands and responds with speech synthesis. Using WWAN, the engineer can pull up specs from the home office and have the application read those specs back to her, all while she's busy physically examining the pipeline. This use case demonstrates how an ultra mobile PC like the OQO combined with speech recognition and synthesis allows the user to perform computing tasks while engaged in a completely unrelated activity. In this case the OQO conforms and adapts to the user's environment to provide desktop application and functionality where none previously existed.



Conclusion

The OQO Model 02 provides a wealth of alternate input mechanisms and performs all of them admirably. Its active digitizer combined with pen flicks and Grab-and-drag is a natural and intuitive way of navigating. Vista's handwriting recognition is superb and the OQO's processor has no problems keeping up provided the CPU is throttled no less than 60%. One of the most exciting alternate input mechanisms is the promise that voice recognition holds with a device like the OQO. Ultra mobility means moving about and often times neither a pen nor the OQO's built-in keyboard are ideal input devices. As speech recognition and synthesis technologies advance it will be interesting to see the new use cases that arise out of ultramobile platforms like the OQO in the field.

10 comments:

Anonymous said...

Wow, fantastic review! That's great to see that speech recognition is so useful on the OQO 02. It makes it appear a lot more user friendly in it's small size. THANK YOU!

Al Iverson said...

Vista has a built in ability to take dictation in word or whatever? How does that work?

Anonymous said...

- Nobody does smaller better than OQO!

Anonymous said...

OQO is hiring!

Unknown said...

what is your opinion on using xp tablet instead of vista, will the performance be affected?

neil balthaser said...

Hi Paddy: Yes, speech recognition is useful on the OQO, especially when you're away from your desk. It's pretty cool to be able to do your normal routine using your voice only. e.g. "open outlook. compose email. to: paddy. subject: voice recognition on the oqo. message: hi paddy!"

Hi Al: Yes, Vista has excellent voice recognition built into the operating system. Almost any application can be controlled by voice: Internet Explorer, Word, Excel, Powerpoint, Outlook. You can dictate anywhere. Just say the name of a field like "Goto search field" and then dictate what you want: "OQO reviews" and then command the search with "Search."

Vista's voice recognizer examines the controls on the active window (buttons, fields, links, etc.) Generally, you can just say the name of the control like "Click Search button" or "Goto Search Field". You can even say the name of a link like and Vista will find it.

Hi Anonymous: Yep, no one so far does it better!

Hi Dick: I highly recommend Vista over XP Tablet. I have a blog entry that details why I feel this way. In a nutshell, while XP's handwriting recognition is good, Vista's is better. They fixed a lot of problems with the UI and added the ability for Vista to learn from its mistakes. Voice recognition on Vista is completley reworked and much better than XP.

Anonymous said...

Hi Neil,
Great review!! Just received my OQO and I must say that it is an amazing device. I'm running on XP tablet and after reading your review I am strongly thinking about switching to vista. What I'm really interested is whether it is possible to dictate using blutooth headset/mic.

Petr T.

neil balthaser said...

Hi Petr T: I tried using my Bluetooth mic. and it didn't work. The problem could be with my Bluetooth headset but I've found that in general Bluetooth headsets are not of a high enough fidelity for voice recognition. Generally, they are tuned for mobile phone, where a 5kbps to 8kbps sample range will suffice. However, to get the best results from voice recognition, you'll need a to use either a plug-in microphone or a usb mic that can delivery fidelity in the 32kbs plus range. I have had excellent results with my Logitech USB headset and mic.

Hope this helps and thank you for writing in!

Cheers,
Neil

Anonymous said...

Neil, are you using the Premium Notebook headset?...I'm trying to find the one you are talking about on the Logitech site.
thanks.

Anonymous said...

Hello Neil,

From what you described, the Vista Business can do all the good stuff that you mentioned already, is there anything extra that a Vista Ultimate can do to justify the extra $100 for it?

Also, since the Vista Business version of OQO 02 does not include a digital pen, so we need to buy a separate digital pen to fully utilize/realize the potential of the Vista, doesn't that seem quite unreasonable on the part of OQO company?

Davy