Some topics related to MP3 players

If only it were so. On March 5, German police raided a booth at the CeBit conference in Hannover. That booth, run by Meizu, contained an iPhone-clone product, but nobody cared about that. Instead, the contraband which absolutely had to be suppressed was a music player for which Sisvel (an Italian company which has done this kind of thing before) had not been paid royalties on its MP3 patents. The player, as it happens, did not even have MP3 playback capability, but that didn't seem to matter. The police duly cleared the booth of all mention of the offending device and saved another day for free enterprise.

This is a pure software patent action, and the U.S. has no part in it. Software patents are truly a global problem. (Police raids raise the stakes in interesting way, though; even in the U.S., things usually start with a polite letter from a lawyer first). Anybody who wonders why companies like Red Hat exercise great care around software patents (and MP3 patents in particular) need only look at episodes like this. The selling of enterprise Linux products is likely to be distinctly harder if your prospective customers see your conference booth being forcibly shut down by the authorities.

Meanwhile, it occurred to your editor, while thinking about music players, that little has been said about the Rockbox project on LWN in recent times. Rockbox, remember, is a GPL-licensed firmware which runs on a wide variety of music players. It offers a wider range of features, has more codecs, is more customizable, and has better accessibility support than the stock firmware on any of these devices. And it's free software.

Since LWN last looked at this project, the Rockbox developers have added a number of new features and new platforms. The abandoned 3.0 release has never happened; the Rockbox developers appear to have given up on the idea of formal releases for now. The daily snapshots generally work quite well, though, and there are lots of satisfied Rockbox users out there.

[PULL QUOTE: Despite the fact that Rockbox supports a lot of players, absolutely none of the supported platforms are currently in production. So anybody looking to buy a player which can run Rockbox must go digging around on auction sites. END QUOTE] The only problem is: it's not clear how many more such users may arrive in the future. Despite the fact that Rockbox supports a lot of players, absolutely none of the supported platforms are currently in production. So anybody looking to buy a player which can run Rockbox must go digging around on auction sites. Many Rockbox users do exactly that, but many more potential users would rather not get their devices that way.

Rockbox ports to current devices are underway, but the developers are fighting an uphill battle. Manufacturers tend to be uncooperative when it comes to releasing hardware information, so a certain amount of reverse engineering is required. And, by the time that work is done, the manufacturers have moved on to a new product. Music players are consumer electronics devices, and, like most such devices, their product lifetime tends to be quite short. So developers on a project like Rockbox will forever be trying to catch up.

Your editor, meanwhile, still lugs around his ancient iRiver H340. People look at it strangely, as if they expect there to be a hatch on the back so that the user can occasionally add another shovel full of coal. But it works beautifully with Rockbox, and a replacement looks hard to find. Your editor wishes that at least one manufacturer would realize that it could provide better functionality at a lower cost by designing its players to run Rockbox from the beginning. Perhaps the project needs better advocacy within the player industry.

There is another approach which could be considered here. The OpenMoko project is trying to rearrange the mobile telephone market by offering a completely open product. Surely a music player, being a much simpler device, would be amenable to the same treatment? As it turns out, there are a couple groups of people trying to jump start just this kind of effort. They have a prototype design, and a competing design as well. Both look like they could produce a respectable player at a reasonable cost - a player designed to run free software from the outset.

Designing a device which can run Rockbox and produce decent audio (and video) output is not that hard, given the components which are available. Turning it into a product which is small and sleek enough that people want to buy it seems likely to be harder. Getting a full device manufactured at a reasonable cost may be the hardest of all; that requires significant up-front money and a distribution channel which can sell enough units to make the whole thing cost-effective. There's also the little issue of those MP3 patents to take care of.

There is no real sign that the Rockbox player developers are thinking on this level at this time. One of the prototype designs carries a Creative Commons noncommercial license in an attempt to prevent others from thinking that way. So the resulting hardware may end up being little more than a kit for especially dedicated hobbyists. Unless somebody picks up the ball and tries to commercialize a product like this, Rockbox may be stuck in its role as the software of choice for last year's players. The good news in all this is that Linux-based tablet devices seem likely to become cheaper, more abundant, and more compact. Since these devices can make fine media players, we may eventually get our completely open gadget via that path. Modulo patent problems, of course.

I love my Sansa e250 running Rockbox, and no one looks at it funny.  Really, it's a very
pleasing device in every way.  I bought it specifically to run Rockbox on, since the port was
ready while it was still in stores. Unfortunately the one on the shelves now, the e250v2,
isn't supported yet.  However, that and several other Sansa models are being worked on,
because they're well designed and popular gadgets. It shouldn't be long before you can buy a
Rockbox target in stores again.  One is enough, if it's a good one.

Something they don't warn about on the web site is that the default UI theme is practically
unusable, particularly if you're over 35; you should plan to choose and install a "skin"
immediately.  In practice, that just means copying a few extra files to the player, no big
deal.

A final warning: "party mode" sounds like fun, but ... well, it's easy to switch off again.

The default UI theme changed last month, of course: are you referring to the new one
(CabbieV2) or the old? I don't think much of CabbieV2 either, but the new grey gradient-shaded
backdrop for the menu screens looks awesome (I've had iPod users on the train asking me how
they could make their iPods look so cool).

(I don't much like the default theme either, and spent five minutes hacking up a
mostly-textual detail-packed WPS, and five minutes with the GIMP turning out a blue-to-red
gradient-shaded backdrop...)


btw, the power-consumption problems on PP devices (which I think includes your Sansa as well
as the iPod) are now largely solved. :)

Oops, I should have checked again.  When last I checked, the default font size was still too
small to make out even with my reading glasses on.

I did notice that the battery didn't last as long as it should.  Time to upgrade!

About patents... typically the hardware for these gadgets includes a hardware MP3 decoder
on-chip, and the vendor gets a license along with the chip, so it's actually easier for player
vendors than it would be for Red Hat.

I can't think of a current MP3 player which does MP3 decoding in hardware. One of the reasons
for the huge delay before the not-release of Rockbox 3 was the implementation of a software
codec layer (with the corresponding explosion of supported codec format), because the Archos
was the only platform Rockbox worked on which had a chip to do it. Everything else uses CPUs:
it's cheaper (modulo patent rubbish), more flexible, and with modern CPUs faster too.

The European Patent Office takes very little notice of article 52, or more accurately,
interprets it very differently to you or I. The sisvel patents are most certainly algorithmic
patents, although some of them are described in terms of hardware, and can equally be
implemented in hardware or software, they are largely stuff that most people here would
consider shouldn't be patentable. MP3 is an interesting example because the original work did
actually require some real research into psychoacoustics, and was dealing with audio
representations, rather than software as such, and thus a better case can be made for granting
patents on it than most algorithms. Nevertheless, having those rights for 20 years so we
_still_ can't freely implement mp3 is clearly excessive, even if you don't think it was plain
wrong in the first place.

Sisvel's aggressive enforcement tactics are legendary. They are just as vile as the more
famous  US-based patent trolls, and indeed have gone as far as sueing other MP3 patent holders
(Thomspon, Fraunhofer).

I have a wonderful Cowon iAudio X5 with 30GBs running Rockbox. Sound is great, radio works,
etc. But I'm now selling it on a certain auction site (your chance, three days left, hint,
hint). Mainly, because I'm not totally satisfied with the feature set of Rockbox. E.g. it is
not able to handle more than one genre per track. Implementing such features is (for me) much
easier with Linux, so my next MP3-Player will be an N800 or an EEE PC. I'm just waiting for
32GB SD cards to become widely available under 300€. So, if I would design a new audio player
(why MP3?, isn't that the bad version of Ogg Vorbis?), I would let it run Linux, not Rockbox,
just because this is where all the developers are.

> why MP3?, isn't that the bad version of Ogg Vorbis?


Because everybody and their mom has MP3s. Everybody uses it, everybody supports it (say a mp3
cdrom in a car stereo), you can buy DRM-free songs in MP3 format, etc etc.

I like Ogg Vorbis more, but it's pointless to re-encode mp3 format into o.v. if you can help
it.

> N800 or an EEE PC

You'll want a N800. EEE gets rather (suprisingly) hot and the battery life is way too limited
(2.5-3hr, goes faster then you think it does and it takes something like 5 hours for a full
recharge). I am typing this out on a EEE, btw.

Personally I think the Neo1973 (or whatever they are calling it nowadays) has a good chance at
being my first 'mp3 player'. (or more accurately Ogg/Flac/mp3 player)

For the time waiting for the Neo, I'll buy an N800 :~)

Um. More than one genre per track in which codec exactly? I suppose you 
could have multiple genre comments in Ogg, but I don't think Rockbox 
currently has infrastructure to handle multiple instances of any tag (sign 
of its MP3 heritage; periodically I think of fixing this and then put it 
off because it seems so unimportant).

You can have multiple genres in both Ogg Vorbis and MP3. In Ogg Vorbis it's done by having
multiple tags with the same name and different values, in MP3 it's a convention (but supported
by some programs) to have a comma separated list of values on the genre tag, IIRC.

Of course, few people care about this. The good thing about using Linux or any other
all-purpose OS for an audio player is, that you can choose between a couple of audio playing
applications matching your desires. If no program fits your needs you will find one in your
preferred programming language to hack. This is harder with a specialised system with a
smaller user and developer base, such as Rockbox (which I actually find very impressive
compared to any proprietary MP3 player system software).

I believe the intent is ironic, a play on the tendency of all too many politicians to equate
free enterprise with whatever kind of intellectual monopoly is being pushed at the moment.

Read it as a statement dripping in sarcasm, since that was the tone with which it was written.

My mind boggles at the hundreds of competing player designs released, most with no unique
features to recommend them.  The only sense I can make of it is that they must be personal
projects of younger sons of Japanese and Taiwanese businessmen, and that breakeven is not
required.  If so, perhaps that identifies a route to getting Rockbox players produced.
Complete Japanese language support may be more important than has been realized.  Getting such
players marketed outside Japan or Taiwan would be another hurdle.

The next time people expect you to have to load coal into your music player, point out that
it's a tool of the 20th century, not the 19th. The hatch on the back is so that you can load
new uranium fuel rods every few months, not coal.

... Just don't do this as you are talking to the TSA guy at the airport, 
about to board the plane.  He's likely to take you literally, if not 
because he believes it himself, then because he expects others in the line 
might.

... And don't do it /on/ the plane either.  You're likely to get the thing 
diverted, and end up being the subject of a rather forceful manual 
inspection of the interior of your hind side!

=8^(

> Despite the fact that Rockbox supports a lot of players, absolutely none of the supported
platforms are currently in production.

While this may be true ...

> So anybody looking to buy a player which can run Rockbox must go digging around on auction
sites.

... I take exception to that. I had no problems buying my Sansa C200. From a short research at
price-comparison sites it seems that enough Sansas are in stock in the home of our esteemed
editor, too. But your friendly neighbourhood electronics store will probably not know these,
so it's mail-order time.

IIRC Sandisk gave at least a little bit of support to the developers of this port. People may
want to reinforce this behaviour with their money.

For massive storage (i.e. HD-based) your best bet is the iPod Video which also seems to be in
stock still.

Be sure to watch for versions and generations. Producers change their HW more often than their
model names, and not all versions work with RockBox.