The txtr beagle Takes a Great Idea and Turns it into Crappy Hardware

When txtr unveiled the Beagle ereader yesterday I fell n love. It’s a great looking minimalist device with a 5″ screen which is going to be sold cheap, run on a pair of AAA batteries, and not require any cables at all. It’s a smartphone companion which connects via Bluetooth.

The device was based on a great idea, but as yesterday wore on one of my readers noticed that the Beagle had a couple serious shortcomings. These were details which I missed because I was insufficiently pessimistic about how txtr would screw up the new ereader.

It turns out that the Beagle is limited to only storing 5 ebooks at a time and it is not an ereader so much as it is a page viewer. It’s running on a very weak (and cheap) CPU which can do little more than show page images. This explains part of the reason why txtr’s Android app is required to transfer ebooks; that app has to convert the text into page images before they are transferred to the Beagle.

On a related note, this means that this device, which is not actually an ereader, does not support the Epub and PDF claimed in the press release and product page.

While this is the first time I have seen this type of conversion on what was intended to be a mainstream device, it is not exactly new. In the past I have come across projects to turn portable DVD players and digital picture frames into faux ereaders. The first step in loading the ebooks was to convert them to page images which could then be viewed on the device via some type of image viewer. This is an old trick which I think dates back to the late 1990s.

So why is this an issue?

First, the so called Beagle ereader is missing a lot of the basic features we have come to expect like a TOC, resizeable text, and so on. But the greater issue here is that txtr could have released a real ereader for only $15 to $20 more – a negligible increase in price considering what would be gained.

While that might seem like a lot, given that the Beagle sells for under 10 euros, it’s really not. That 10 euro price is going to be subsidized by carriers, and the actual parts cost for the Beagle is somewhere above $50. While I don’t have actual figures, I do know something about the cost of the screen, and if we take the estimated costs for Amazon’s $79 Kindle as a guide we have enough info for a back of the envelope calculation.

That Kindle was estimated to cost $85 to make, but for now lets ignore most of the items on the list and focus on just 2 details: the screen and the logic board.

Amazon reportedly paid $30 for each E-ink screen. That figure is off, but I cannot say by how much (it’s confidential). But I can say that txtr will likely be paying a lot more for the 5″ screens. They’re not buying in as high of a volume as Amazon so they won’t get the same discounts as Amazon. And that means txtr could be paying at least $40 a screen, and the cost might be above $50.

That screen is the single largest cost for the Beagle – or for any other E-ink ereader for that matter. Another major cost is the logic board. This contains the CPU, Flash storage, and other chips which make the ereader go.

Amazon’s $79 Kindle was estimated to have $30 in chips inside, all packed on a single board the size of a laying card. Assuming that the Beagle went with cheaper components, it likely used a logic board which cost $10 to $15.

And that means, folks, that the Beagle could have been a real ereader like the cheap Kindle had txtr spent an extra $15.

The decision to not use the more expensive chips might make some sense given how txtr expects the Beagle to be used but it is still a false economy. They pared the costs down to the point where the lost abilities outweigh the savings.

If you still don’t agree with my conclusion, let me give you a counter example. The Pyrus Mini, a new ereader with a 4.3″ screen, is going to cost 50 euros. This is a very basic ereader, but unlike the Beagle it actually supports Epub and PDF. Admittedly, Trekstor has earned the epithet Dreck-Store in Germany, suggesting that they make awful products, but it is still cheap.

If Trekstor can afford to make and sell the Pyrus Mini for 50 euros then txtr could have made the Beagle into a real ereader. And that is why I am pissed with txtr. While they have great ideas, apparently one of the steps in bringing those ideas to market involves screwing up on a simple and basic point.

This is something they seem to do with all their hardware. For example, the original txtr eReader never made it to market because txtr couldn’t figure out how to get the costs down. That thing stayed at a price above 300 euros, and that just wasn’t practical in 2010 (or any time since).

P.S. If you’re thinking that transferring the images over Bluetooth is slow, you might be wrong. These will be grayscale images with a minimum of data. The file sizes should be quite small and thus quick to transfer.

40 thoughts on “The txtr beagle Takes a Great Idea and Turns it into Crappy Hardware

    1. It’s a mini that does have potential in the wanting department. However, most of those shortcomings are so important in my eyes that it’s an absolute no-go. The 5-book limit isn’t the biggest problem, nor is the lack of an USB connection. The lack of native rendering of epub and PDF could be, though. How do those page images look after the app has rendered them? I prefer to use PDF on my 5″ BBMini (hand-made PDF, made up for a 5″ screen, so no reflowing required). It has a font (size + family) that is easy for me to read on a 5″ screen and a page-layout as I want it to have.

      I think I’ll stick with my own 5″ for now (until that 4.3″ arrives that is, I’m really looking forward to that one!)

      1. Also with the Beagle, you couldn’t highlight a word and do a dictionary lookup. I am doubtful the mini will have dictionary lookup, or even the ability to swap out a different one, but hopefully I am wrong.

        I was pleased to hear from Nate’s buddy over at Good E Reader that the Kobo mini is using the same firmware as Kobo Touch, because the firmware was upgraded to included multiple foreign language dictionaries you can select as your default.

        I am probably going to remain a dedicated 5 inch PRS-350 user for the foreseeable future, although it would be nice to synchronize it with my tablets like you can with the Kindle. If Amazon can made a 5 inch or smaller device, I would buy it for that reason.

  1. I suspect the reader’s inability to natively process ebooks is intended to be a “feature” to help sell it to the carriers: “They can only use it to read books they buy from *you*.”

  2. I don’t think the price on the Pyrus Mini is that great either though. 50 Euros is around $64 American. That’s not that much less than the ad-supported Kindle non-Touch. With the screen and the chips costing that much, they don’t seem to have much leeway.

    I wonder if either B&N or Amazon will later release a low-cost 5″ reader. Kobo has.

    1. Converting european consumer electronics prices to the US isn’t just a matter of currency conversion.
      That 50 euro price would most likely translate to US$50 or even US$40.

  3. passing data from one device to another is not a serious problem, you do not read the whole book in one, but for parts) Now You could preset the font size before transferring the image quality is the true problem. With 4GB of storage would have to try it with comics

  4. The one cost you’re not accounting for is the cost of developing the firmware. This is not cheap. Nor easy, as some manufacturers are discovering.

    It’s a lot cheaper to develop a page-turner than an eReader. And it’s a lot cheaper to develop the eReader conversion software for a computer/tablet than for an embedded device.

  5. Storing text as image and sending images rather than text sounds too cpu intensive and bandwitdth intensize. Are you sure you are right about image files?

    1. I don’t see how this could be more CPU intensive than a lot of the games you can play on Android smartphones. And the images will be 600×800 grayscale. They won’t be that large – probably somewhere under 30 KB each.

  6. As an avid reader, trying to bring other people to read more, I was exited to hear about this txtr beagle. I was imagining giving these away as gifts. And then I read this. Heartbroken I tell you…

  7. I don’t agree.

    txtr’s argument is that people generally want an eReader to just read the books. You don’t need to be able to change fonts on the fly once you’ve chosen the font that works for you. You don’t need hundreds of books – five books is about right for me for a fortnight’s holiday, if not on holiday I get through no more than three a month. I don’t want any distractions (if I want to shop for books, or browse the web, I have an android for that).

    I personally don’t ever refer to a dictionary while reading, even in other languages. If I was studying a text it’d be a different matter, but most of the time I just want to read.

    So for me, this device sounds great – the battery life, capacity and features fit in with what I actually need for an eReader. The real problem with me is if it locks into the txtr store – I want to be able to load ePubs from different sources, and it’s not yet 100% clear if that’ll be possible.

    At the expected price, too, it’s low-risk. I don’t have to justify to myself trying it for just 8 pounds. At 20 or more, I’d be thinking much harder whether I’d rather go for one of the fully-featured eReaders, but at such a low price point I don’t even mind too much if it ends up being rubbish – I’ve not spent so much I’d feel I had to stick with it.

    A brave venture by txtr, I applaud them for daring to be different, and I hope my mobile provider offers this.

    1. I’ve uploaded epubs from other sources to my txtr account. I don’t expect it to change when the beagle comes out. The beagle has 4GB memory and can story many books but can only remember the last page you were on for five books.

  8. The numbers don’t make sense. 4Gb of storage and only 5 book capacity implies 800MB/book, unless most of that space is used for the OS and display program. If each book is 100s of MB due to the image format, it will take ages to load. Max Bluetooth transfer speed is about 2.1mb/s, so maybe about 200KB/s, or 5 seconds to transfer 1MB. The 24mbit/s high speed Bluetooth standard actually uses a 802.11 wireless co-channel for data transfer. Bluetooth itself is much slower. If a book is 100MB, 5 sec/MB gives’s 500 seconds of transfer time, or 8 1/3 minutes. If a book truly is 800MB, that’s over an hour. If most of the memory is taken up by OS and application program, how big is an image based book likely to be? An 800×600 image at two bits/pixel (4 shades of gray) is 960kbits, or 120kbyte for an uncompressed image. An uncompressed 300 page book would be 36mb, transferring in 180 seconds, or 3 minutes. Maybe not too bad, if the Bluetooth link can sustain the 120KB/s rate. However, differences in phone firmware, distance, and so forth could significantly drop that rate, resulting in greater than 10 minute transfers, perhaps much greater. You could do compression, but file sizes are probably still several MB, so any slowing of the Bluetooth transfer could still stretch things out. And forget phones with earlier Bluetooth, those slow transfer speeds really would take an age. So the point is, either very little of the 4GB storage is actually available for books, or file size is big and transfers will take a long time, or some combination of the two. Something is not adding up.

    1. Maybe he files transfer in compressed mode but decompress later.
      And they need to leave a lot of extra space in case somebody wants to read WAR AND PEACE, The AMBER CHRONICLES omnibus, and a few other monster books. ;)

      1. According to Wikipedia, “War and Peace” is 1440 pages. By my previous estimate of 120KB/page uncompressed, an image-based file of the book would be 1440×0.12, or 172.8MB. At 5sec/MB, that’s a best case 864 seconds, or 14.4 minutes. JPEG or some other compression scheme might improve things by a factor of 2-10, down to a minute and a half for a compressed 17.2MB file. Note that an ePub of the same book is 1.5 to 3.5MB, depending on exact format. So again, either most memory is not available, or transfer time is long, or some combination. What I think is going on is that by doing an image-based approach requiring proprietary phone-based conversion software, these guys have invented another kind of data silo. You can only read the book on their device. Customer lock-in. It would be interesting to see if they convert standard ePub to their format, or if you have to buy from them. I find it hard to believe that the processor doesn’t have the oomph to do ePub on the device. If they can’t do that, uncompressing page by page images is likely to be slow too.

        1. I bet the beagle displays png, not jpg. Decompressing the jpg file might take to much processing ability. These images are probably also 4-bit grayscale, not color, so they could be even smaller than you might expect.

          1. You’re probably right. JPG is not such a good choice for line drawings and letters. PNG is patent-free GIF, better suited for text images. I’ll have to look, but probably some kind of LZW lossless compression followed by Huffman-style encoding. Quick to decode. But if they can do all that, they could probably do ePub, which uses some kind of LZW related compression. To me, the device and marketing plan look like a data silo for cell phone companies masquerading as a device with cheap (subsidized!) up-front cost. Face it, you can buy a low end Kindle outright for $69. Their unsubsidized price is probably not much less than this.

            My file size estimates were for 2bits/pixel, or 4 levels of gray. 4-bit grayscale would be 16 levels of gray. Their display is probably not that good. So my estimate of file size is probably pretty close to minimum size. Also, maybe they went with Bluetooth to avoid the cost of a mechanical connector, as in USB? The cost of packaging and mechanicals can exceed the cost of the electronics.

    2. I don’t understand the 5 title limitation either. Just based on my own guesses about the image size I thought that hundreds of ebooks could fit into the 4Gb of storage. Maybe the microprocessor isn’t capable of addressing that many different storage locations. I don’t know.

  9. And of course, Kindle is itself a massive data silo scheme. You need the Amazon app or the Amazon device to read books available only from Amazon. Except with Kindle, you deal direct with Amazon and get a fairly high level reading experience, as opposed to dealing with a 3rd party provider selling Android handsets and wireless service, with a sideline in various barely capable companion devices linked to the Android handset by Bluetooth.

    Apple and iBook I think are similar. I don’t believe that iBooks are industry standard ePub, so another data silo.

  10. Take a look at http://www.youtube.com/watch?v=314v1H_aN2o&sns=em

    8- bit processor that cost $2, minimal hardware in the device itself, external transcoder (epubgrinder) to convert ePub text and any images into a series of bitmaps that can be rapidly dumped to the screen by the tiny processor. Sound familiar? Done right on an industrial scale, these things could be cheaper than dirt, especially if you use LCD I stead of e-ink. Maybe txtr is doing something like this? No mysteries or miracles, just clever engineering.

      1. Arguably, the beagle is a real ereader, but one that requires a companion device to massage the data first. Hey, my Sony PRS-350 requires a USB connection to a PC to get a book. The beagle’s Android platform dependence for companion device is probably a business model decision. No reason their transcoder couldn’t work on a PC. My guess is, they want something to offer the wireless companies in terms of revenue stream. Bluetooth fits into this, both from cost and power consumption points of view, and as a natural complement to cell phones. I still would like to know how long it takes to load a book using Bluetooth. As for BOM, has iSuppli looked at it? My guess is, down around the subsidized price of $13, with the e-ink display being the major cost. Maybe as high as $25, depending on display. With a non-touch monochrome LCD it would be dirt cheap.

        1. I know a little about LCD. E-ink, not so much. If your display numbers are right, that would mean that the low end $69 Kindle is being sold for close to cost. Amazon makes it up by selling ebooks that can be read only on their readers. So what the beagle guys have figured out is a way for an interested 3rd party (the wireless carrier) to subsidize their cost without getting into the book business. My initial cost estimate was based on cell phone business models. ISuppli says an iPhone cost $210 to make. The carrier values it at $600-700, but sells it to you for $199, covering the BOM cost (but not Apple’s profit selling it to the carrier). You pay the actual BOM cost, and the carrier makes back the Apple mark-up in a month or two of fees. The move to higher out of pocket handset costs generally might mean that all up-front carrier costs are paid for when you go on contract, not just raw BOM cost. So you’re right. If there’s any kind of substantial carrier subsidy to txtr, they could have put another $5 into the design and delivered a way more capable device. I suspect that the subsidy situation is different in Europe, possibly accounting for the extreme bare-bones approach to the beagle design. They could have had Kindle-like capability, using standard replaceable batteries and no cables as key differentiators.

      1. I am confused. So apparently txtr had a more or less standard ereader ready for market in 2010, but then essentially went bust and that ereader was never released to the market. In 2011, 3M dumped a bunch of money into txtr and touted the formerly dead device as a companion to their new 3M cloud service for libraries. And now, in 2013, 3M is using someone else’s private label ereader for libraries to load up with books from the 3M cloud and lend out, while txtr has gone off and done a minimalist ereader that requires smart cellphones to do the heavy lifting, with no mention of any 3M involvement. Actually, the new txtr would arguably be better for libraries to lend out to patrons, as it’s a lot less costly than a standard ereader. What’s going on?

        1. The circa 2009 txtr Reader first had production issues at the parts suppliers and then ended up costing too much. The BOM was based on a $400 retail and that simply wasn’t practical by the end of 2010.

          And yes, 3M bought into txtr so txtr could help build the 3M Cloud Library, but otherwise left txtr free and independent. 3M did get an ereader elsewhere, yes.

          That is complicated and inexplicable but it’s the truth.

Leave a Reply

Your email address will not be published. Required fields are marked *

You may use these HTML tags and attributes: <a href="" title=""> <abbr title=""> <acronym title=""> <b> <blockquote cite=""> <cite> <code> <del datetime=""> <em> <i> <q cite=""> <strike> <strong>