Speed-reading pdfs using the “Spritz” technique

sprint reader logoA couple of days ago, I came across Spritz, a company that is popularizing a speed reading technology whose name I’ve forgotten called Rapid Serial Visual Presentation.  You get the idea once you look at the Spritz website. I immediately thought that this could be a way for me to read all those texts that I have to read as an academic a lot faster.

I am a very slow reader for three reasons: the first is that I look up every other reference and almost all footnotes. Sometimes I get so distracted that I even look up some of the references and start reading those instead (and so on).

Another reason is that when I read an interesting paragraph of sentence, that often triggers my own thoughts and associations and I start wandering off, thinking about how I might integrate that idea into my own text or whatever.

Finally, the third problem is that I simply read slowly, I guess because I really want to understand and thing through everything, rather than focusing on the essentials.

Now, I don’t want to philosphize too much about the pros and cons of speed reading here. Suffice it to say that if the aim is to get through a text reasonably fast at 350 words per minute with at least superficial understanding of the content (or even ridiculously fast at 800 wpm with probably minimal understanding but a rough idea), then this Spritz technique seems adequate to me, and I have indeed read two articles that way yesterday.

My point here is to show you how to read pdfs using this technique (not the original Spritz itself, cause their app is not released yet) because the reading apps that are out there at the moment seem to work only with plain text, while most academic articles come as pdf files.

It’s not a big deal, actually, but it took me a while to figure it out nevertheless. All you need is the Chrome bowser with the Sprint Reader extension installed. You also need to make sure that you have the native Chrome PDF Viewer enabled. This is the case by default, but if you are using Adobe Reader or Adobe Acrobat, you might have disabled it. In order to enable Chrome PDF Viewer, type “chrome://plugins/” into Chrome’s address bar and scroll down to find Chrome PDF Viewer and, well, enable it. The following will not work with pdf-files that are not displayed using Chrome’s PDF viewer.

You can now use Sprint Reader to speed read not only text on websites but also your locally stored pdf files (just drag them into the browser). Select the text you want to read, right click and select “Sprint read selected text”. There you go. (Needless to say that your pdf file needs to have actual text in it, not just a scanned image of text. If you have an image of text, you need to run some OCR on it.)

The Sprint Reader extension in action
The Sprint Reader extension in action

I recommend setting the “pause after period” higher than the default 450 milliseconds. I’m currently using 900. I also set the “pause after paragraph” to 2000, but that basically has no effect when reading a pdf, because there seem to be no paragraphs in there that the Sprint Reader would recognize as such, unfortunately.

There are also a couple of other drawbacks that we currently will have to live with, especially that the reader will – not surprisingly – read all the text in the pdf, which means it will also read the header on each page, the page numbers, and – most annoyingly, the text inserted on every page by various publishers, such as:

This content downloaded from xxx.xxx.16.16 on Thu, 17 Apr 2014 06:49:43 AM
All use subject to JSTOR Terms and Conditions

You will also encounter problems when there are tables and figures in the text, but hey, otherwise it works okay.

Why did I chose Sprint Reader and not any of the many other similar services and extensions such as Spreed – speed read the web, Spread Speed Reading Extension, Spreeder, or OpenSpritz, to name but a few? Well, it’s the best of them all. It’s not perfect (I did encounter a couple of crashes or whatever it was when it simply did not work until I restarted the browser), but it is very customizable and it works with pdfs, which not all of the others do.

Having said that, here are some suggestions for improvement in Sprint Reader, especially for people like me reading scientific texts:

  1. Add an option to exclude (i.e. hopp over) brackets containing a four digit number. Why? Because that would exclude all references provided using an Author-Year citation style and make reading academic texts a lot easier.
  2. Recognize abbreviations such as e.g. and treat the dots in these differently (i.e. don’t apply the “pause after period” rule). One quick way of minimizing this problem could be by checking whether the dot is followed by a capital letter. If not, it’s not a period.
  3. It would be nice to be able to exclude customized strings, such as “This content downloaded from xxx.xxx.16.16 on Thu, 17 Apr 2014 06:49:43 AM All use subject to JSTOR Terms and Conditions” (see above)
  4. Ignore hyphens at line breaks and join the words to one. The few cases where the hyphen should not be eliminated because it is a hyphened word, are negligible.
  5. Recognize paragraphs in pdfs (see above). I’m not sure whether this is possible, but I might as well add it to my wish list.
  6. Add a keyboard shortcut for quick rewind as, for example Spreed has it (I think it was spreed anyway). In addition, it would be nice to have the replay at a lower speed and then back to normal speed when passing the point where the rewind was initiated.
  7. The extension should not be dependent on the text selection in chrome, once it has started. This is feels almost like a bug to me, although it is not, but it is a bit of a pain in the *** that you cannot unselect the text you are currently reading without the reader losing track of it. It would be better if the reader would load the whole text into its own memory, once you press play, making it independent from what happens in the browser.

UPDATE: I just realized that the Sprint Reader sometimes seems to have problems rendering text in pdfs. I have a pdf in which it reads some of the text without problem but when I select and right click other parts, there is not even a possibility to select “Sprint read selected text”. In fact, this may be a problem of chrome (or the pdf?) and not the extension. Either way, be prepared to encounter this problem with your favourite pdf file…

IRIScan Book 3: pretty much useless

Image I just got an IRIScan Book 3, a portable scanner that looks promising when you check it out online. And I’m sending it back. The problem is that you can never rely on it actually scanning the whole page that you intend to scan. Unfortunately, I don’t have the time to make a video of this but I hope I will manage to do one before the 30 days trial period is over.

The device is a nice idea, but it simply doesn’t work. If you get one, you can expect to have problems with it all the time. No, my device is not defect. The product is simply badly designed. There are two main problems:

  1. As far as I can see, the scanner measures its movement through the rolling thingy on  the bottom, i.e. it does not use the actual optical sensor to register movement. As a result, it will stop scanning once the rolling thingy has rolled over the edge of the book so that it is suspended in the air. In other words: while the scanning sensor is still gliding over your text, the “wheels” tell the device that it is no longer moving and it will therefore stop recording the signal coming from the sensor. This may not be a problem if your book has a margin of at least 1.5 cm, but if your margin is narrower (or if you want to scan the notes scribbled on the margin) then the Iriscan Book 3 just wont work for you.
  2. The frame of the device is too thick so that it wont scan what is towards both ends of the sensor, i.e. the width of the scanned image is smaller than the sensor window on the bottom of the device suggests. Again, you could work around that making a mark where the sensor really starts and use the device accordingly, but the problem is that you may not always be able to do so because the fold of the book is preventing you from shifting the device far enough. Similar to point 1 mentioned above, this will not be a problem if your book has large margins, but I have one here that has just under 1.5 cm margin at the centre fold and the leads to the first letter or so of each row to be cut off.

I could also complain about the wifi functionality (on the Irisscan book 3 executive), but that has become a minor quibble compared to these problems. The problem with the wifi functionality is that the Iriscan Book 3 does not conntect to your existing wifi but it sets up its own wifi hotspot and if you want to connect your computer or ipad with it, you need to disconnect from your wifi (and hence the internet), connect to the Irisbook wifi, scan, connect back to your original wifi and only then can you upload the scanned documents.

Honestly: this could have been solved better. Why did they call it IRIScan Book, when it can scan single pieces of paper just fine, but not books? Anyway: I’m sending it back. Luckily they I.R.I.S. is giving a 30 day money back guarantee, so that should not be a problem. Despite this, I recommend anyone to think twice before ordering this. If you think you might want to use it for scanning pages out of books, don’t buy it.

Instead, I will try a fascinating software called booksorber. I had considered it before I bought the Iriscan Book, but I thought it would be nice to have a dedicated device for scanning, rather than setting up my camera everytime I want to scan a couple of pages, but now I’ll give booksober a chance. I’ll post my exoerience here as soon as I had time to test it.

UPDATE [31-10-2013]: If you are looking for an alternative portable scanner, you might want to have a look at the Xcanex document scanner by piQx which I reviewed in another post.

ASUS Transformer Book and Windows 8: a big disappointment

ASUS Transformer Book OK, this is a bit off topic, but I have to briefly air my frustration about the ASUS Transformer Book (TX300CA – C4005P), which I’ve been waiting for so long for. After a few days of good willed fiddling with it, I’m sending it back cause I know that this will just continue to frustrate me, even if I am starting to get used to some of the annoyances of Windows 8, that drove me crazy at the beginning, such as the layout of the on-screen keyboard or the fact that it does not always show up automatically when you are in a text field (it does work with genuine Touch Apps, but only sometimes when you’re in desktop mode), or that you have to aim for the little X in the corner in order to close it, rather than a bigger key or a swipe gesture.

I guess I needed to use a Win8 tablet in order to understand what a good job Apple did with the iPad. I was similarly annoyed with my iPad 2 when I first got it and I still think it has major faults (not to speak about the closed systems approach of Apple in general), but I have to admit that they at least got the basic essentials of the user interface working. Things like browsing a webpage and doing a quick pinch to zoom in to read small print or smooth scrolling or simply making it easy to tap on a link or button or whatever without having to be dead on centre, those are the things that I now appreciate on the iPad. Because they don’t work.

While I appreciate the amount of pixels on the Transformer Book screen, this leads to buttons, menus, and text being incredibly small on your screen so that they are not only almost illegible but, most of all, un-tapable with your fingers. Only when I increased the screen display to 150 percent did the device become somewhat usable via touch screen. In the default mode, it you just keep tapping and tapping until you finally hit something and get a response (though it might be the wrong one, cause you tapped slightly too far right or left or so). Why can a convertible in this upper price range not be configured so that it is usable from the start?

And let me say, that even with 150 percent screen display, I often have to tap multiple times to, for example close a tab in Chrome. I guess you have to be dead on centre on whatever you are trying to tap and there is no tolerance or smart algorithm that makes Win8 realize that when you tap 5 pixels next to a link, maybe you were trying to tap that link. Oh, and speaking of browser. I installed Firefox first and it seem like they havent managed to produce a touch screen compatible version yet, cause what happens, at least on some pages (most importantly: google search results), when you try to scroll by swiping up: you are selecting the text on the page. WFT? – Anyway, this is not ASUS’s fault and probably not that of Win8 either, but it’s part of the experience.

Similarly, with the Evernote App: I wrote a note on the Transformer Book up to today but it did not sync until today. And it did not do so until I went back and tapped sync and waited in the App until it had synched 100%. Apparently it does not synch automatically in the background. WTF?

Another very basic problem was that when I first started the Transformer Book, the PC clock was apparently set to a date several weeks ago. OK, you cant expect the time and date to be right when you get the device, but did Win8 bother to check time and date with one of the many NTP servers out there? No. I had to manually trigger that my going to time and date settings and clicking on “Sync with NTP server” or something like that. And when did I do that? After I spent about half an hour trying to figure out why I was getting security certificate issues from Chrome all the time. It would not let me visit the most basic websites, claiming that there was a problem with that site’s security certificate. – The reason was that my PC time was weeks in the past and the website was claiming to be way in the future, and that is of course suspicious. So that was another wasted 45 minutes, just because Windows 8 was unable to automatically correct the date. What year was this operating system produced again?? Oh, and another consequence of the wrong date during installation was that when I corrected the date, the 30 day trial version of McAffee expired and Windows started bugging me that my computer was not safe. Not that I am in any way interested in McAffee bloatware, but what if I had installed other trial versions that I actually did want to try for 30 days?

It’s about time for me to mention one of the major failures: at some point the TouchPad stopped working! And no, it was not because I deactivated it with Fn + F9, it truly did not work. When I went to the Device Manager, the Touch Pad was listed as a “hidden item” because it was apparently not recognized by the system. (NB: the TouchPad is built into the Keyboard dock of the ASUS Transformer). I tried updating the driver, but it said that it was up to date. The way I eventually solved it was to boot Windows 8 in Safe Mode (Took me another 30 minutes or so to figure out how that works in Win8, since the F8 button that used to bring up the Boot menu in previous versions of Windows no longer works with Win8. It was so complicated (multiple clicks into some menu) that I don’t even remember how I did it. But when I eventually booted into safe mode, the Touch Pad started working again and continued to do so when I booted normally again. (Yes, I did try to reboot normally before, but to no avail.) I mean how can a Touch Pad just stop working on a new device??

But it fits with another annoying problem: the left swipe, which shuffles you through open apps, does not  works when the device is turned upside down and it also sometimes just stops working in other situations was well (i.e. nothing happens upon left swipe) and you have to use other ways to move around and then at some point it works again. And I think the same has happened with the right swipe, which is supposed to give you the charms menu, but even when the charms menu works, it seems to give you different options depending on whether you’re in desktop mode or metro mode (or whatever it’s called). But in both modes you will still give you some similar options, like the Control Panel, only in metro mode it’s called PC settings, and it only gives you very basic settings. Once you realize that, you still need to figure out which settings you can change via PC settings and for which ones you need to go to to the control panel. This is just an example of a broader problem: even when you decide to ignore those metro tiles and work mainly in desktop mode (which is not easy with a touch screen), it still feels like there are two worlds on your device: the touch metro world, with its apps and settings and what not. And the familiar windows 7 desktop. And the problem is that you don’t really understand how they interrelate.

For example, I installed Sugarsync from the App store, as well as the desktop version of sugarsync, and I still havent understood how they relate to each other. Admitedly, I have not spend much time trying to figure that one out, but that’s the point: I don’t want to spend weeks figuring basic things out all the time. I’m spending enough time doing this anyway cause I’m a gadget and tech guy. So I expect that basic stuff just works intuitively, and that’s simply not the case with Win 8. I can understand now, why PC sales dropped so dramatically: I will avoid win8 on my PC as long as I can too. Problem is: I won’t be able to avoid it on a Windows tablet…

Want some more failures and disappointments? – I really did not install a lot of Apps (afterall, there are really not that many in the App store!), but when I tried to do so. Microsoft app store was unreachable several times. (And no: my internet connection was fine, I tested that). But not only was it not reachable, but what annoyed me even more was that Windows 8 blamed my internet connection and suggested I should fix it when it was clearly a problem on the side of the App Store thatfailing to reply on time or whatever. I know, those kind of error messages have a long tradition in Microsoft Products, but I was under the impression that in Win7, things had improved somewhat and even functions like “Try to solve this problem for me” sometimes actually solve the problem.

And there is more: I calibrated the touch screen because my impression was that it was not reacting properly, i.e. not targeting the item I was trying to tap. The result was (and it again took me a while to find that out!) that the right swipe for the charms menu no longer worked. Its an official bug. Here: http://support.microsoft.com/kb/2795584. So although there is a possibility to calibrate the touch screen, this will result in Windows 8 becoming more or less dysfuctional, so that you are advised to reset the calibration.
Last, but not least, one of the bigger reasons for me to return the ASUS Transformer Book is that it is quite loud for my taste. Now, admittedly, I am not the typical user when it comes to fan noise. I cannot bear fan or hard drive noise and I don’t understand how people are willing and able to work on computers that not only have a clearly audible fan but one that is really loud. I probably could live with the of the fan of the ASUS Transformer book when it is running in its slowest mode (as it is during surfing and basic activity) but once the CPU gets just a bit busier (for example while installing an app), it accellerates so much that it becomes clearly audible and that is just a deal breaker for me, at least when so many other things are just not right.
It looks like I’m now going to be a bit more patient and wait for convertibles with the new Haswell processor to come out and I hope to find one with a Wacom Digitizer, so that I can highlight and scribble into pdfs while reading them. The Toshiba WT310 looks promising, for example, but if you have any other suggestions that come close to the following criteria, please leave a comment:
  • at least a i3 core cpu (i5 preferred, or even i7, but that might lead to problems with the next point)
  • low noise
  • at least 11 inch screen, preferably 13 inch
  • digitizer pen (i.e. non capacitive)
  • weight well below 1.9 kg for keyboard and screen combined.
  • touch screen
  • touch pad in front of the keys

Turn a double-page pdf into single pages using briss

Wow! I just discovered a little tool that does exactly what I’ve been looking for since ages. I’m sure you’ve also wondered about this after having scanned a book or book chapter: how can I cut those scanned double pages down the middle and make two pages out of it (because it is two pages in the book)? And, of course, you want to do it automatically. If you’re using Adobe Acrobat, you might expect that this rather expensive software has that function built in, but you will find: it hasn’t.

But thanks to Gerhard Aigner, you wont have to worry about this any more because he created Briss, a  cross-platform (Linux, Windows, Mac) java-application for cropping .pdf-files. Oh, and did I mention that it is freeware? You just open your pdf file, position the rectangles according to the position of the pages in the book and select “crop pdf”.

My only criticism (or suggestion for improvement) would be the following: while the program allows you to copy cropping frames across different sections of the document, it does apparently not allow you to copy the frame for the odd pages and use it for the even pages. So if you want to have odd and even pages in the same size and format, you need to draw your even-pages-frame on top of your odd-pages-frame to make sure they are the same size (only visual judgement!) and then drag it over to the even page. It’s feasible but a bit cumbersome. Or well, actually, it can be almost infeasible at times, namely when your frame as exactly the height of the page, because the the size of a frame changes if you move it across the edges of the page and since you will hardly be able to move your mouse 100 percent horizontally, the frame will get smaller as you drag it.

Oh, and drag’n’drop would be nice: i.e. the possibility to drag your pdf into briss directly from Windows Explorer. And an option “First page last”. The latter would facilitate scanning of brochures and booklets because in those the first scanned page will contain the front cover and the back cover.

Another think to be aware of is that the briss (just like Adobe Acrobat) apparently does not delete the parts of the pdf that are “cut off” but just makes them invisible. At least that’s what I gather from a comment by user “Viking2ev” on sourceforge. But that’s only a problem if you want to get rid of confidential information.

Oh, and irrespective of briss, during scanning, you need to watch that your book in the same position on the scanner for every double page. Otherwise you might have to do several runs with briss and then merge the files back into one with Acrobat or so.

Having said all that, I still think that briss is an excellent tool for scholars to use. Just go and try it out.

[UPDATE: I just found a very similar program that seems to be more flexible to use: It’s called Scan Tailor, it’s also Freeware and it seems to be maintained better as the last update is from May 2012 while Briss hasn’t been updated since July 2011). The only (big!) problem is that it is apparently unable to read .pdf-files. So your scans have to be in supported files types are .tif, .tiff, .png, .jpg, .jpeg … So if that is okay for you, try both and tell us what you think about them in the comments below!]