Rockbox Player - Project to design and build a Free/Open hardware audio player

[roach]

I have some experience with designing and building MP3 players. My most recent design is an ARM7-based player with external MP3/AAC/WMA/WAV/midi support via a VS1033 from VLSI...

...and then I saw RockBox.

I want to design a player that will do all the marvelous things RockBox is capable of. Themes, plugins, games, Video, Software codecs, you name it.

So I need a push in the right direction with regards to hardware. What kind of CPU does the community recommend? How much RAM? Hard-drive or Flash-based?

Take as an example the new iPods. I mean, if RockBox ports to the 5G and 6Gs, then someone, somewhere must know something about the hardware, right?

Ideas, suggestions and, most of all, information are welcome!

[Mad Cow]

So you can actually make a functional MP3 player? How do you print the PCB's and write firmware for it?

[saratoga]

So you can actually make a functional MP3 player? How do you print the PCB's and write firmware for it?

You can have PCBs made for relatively cheap, and write your own firmware.  The housing would be the hardest part I think.

I want to design a player that will do all the marvelous things RockBox is capable of. Themes, plugins, games, Video, Software codecs, you name it.

All of this is documented in the wiki.  I recommend looking at the 68k targets first since they are most mature and have the most optimizations commited.

[Mad Cow]

You can get PCB's printed? I thought that was just for huge orders from big companies. The housing wouldn't be too hard, check out the projects on http://www.benheck.com. I still think writing firmware would be the hardest, because you can't get much more than rockbox devs anyway.

For the future I would recommend a dual-core CPU, because I've read that will be the fastest way to decode video and audio without an external chip. The problem is, nobody knows how to use both cores yet.

[Mad Cow]

You can get PCB's printed? I thought that was just for huge orders from big companies. The housing wouldn't be too hard, check out the projects on http://benheck.com. I still think writing firmware would be the hardest, because you can't get much more than rockbox devs anyway.

For the future I would recommend a dual-core CPU, because I've read that will be the fastest way to decode video and audio without an external chip. The problem is, nobody knows how to use both cores yet.

[roach]

All of this is documented in the wiki.  I recommend looking at the 68k targets first since they are most mature and have the most optimizations commited.

Do you have a link to the actual page? Or search terms I can use? The problem with wikis (IMHO, and not just the RockBox wiki...) is that they're such a pain in the a$$ to navigate. You can only get to a page by searching for it. God only know how many "hidden" pages are on the average wiki, just because there are no external links...

[roach]

For the future I would recommend a dual-core CPU, because I've read that will be the fastest way to decode video and audio without an external chip. The problem is, nobody knows how to use both cores yet.

For a music player, or even a full-fledged PDA, this is waaay overkill. I was thinking an ARM-9, prolly running at about 200MHz.

[saratoga]

All of this is documented in the wiki.  I recommend looking at the 68k targets first since they are most mature and have the most optimizations commited.

Do you have a link to the actual page? Or search terms I can use? The problem with wikis (IMHO, and not just the RockBox wiki...) is that they're such a pain in the a$$ to navigate. You can only get to a page by searching for it. God only know how many "hidden" pages are on the average wiki, just because there are no external links...

The problem with users is that they can't be bother to look for themselves.  The iriver port is linked on the front page of Rockbox.com as well as the wiki.

http://www.rockbox.org/twiki/bin/view/Main/IriverPort

Next time look before asking.

For a music player, or even a full-fledged PDA, this is waaay overkill. I was thinking an ARM-9, prolly running at about 200MHz.

FWIW the Ipod is a dual core ARM7 running at <=75MHz.  If thats overkill, your ARM9 core would be roughly 3x more overkill  Should make porting easier though.  It'll be so many times faster then any other Rockbox target you won't really have to worry much about optimization.

[markun]

I was thinking an ARM-9, prolly running at about 200MHz.

I think 200MHz ARM9 CPU is fine if it's not much more expensive than a slower CPU. You can change the clock to save battery time (like we do on most targets). The Toshiba Gigabeat we are working on has a 300MHz ARM9.

I want to design a player that will do all the marvelous things RockBox is capable of. Themes, plugins, games, Video, Software codecs, you name it.

You just need a DAC and a (color) display for that.

So I need a push in the right direction with regards to hardware. What kind of CPU does the community recommend? How much RAM? Hard-drive or Flash-based?

Hard-drive or Flash-based really only depends on what you want. Flash is more robust but with a hdd you can put a lot more music on it.

For a flash-based player I don't think you need that much RAM, as reading from flash doesn't cost you much battery time compared to a hard-drive.

If you go with the hard-drive then you want a lot of memory (at least 32MB I would say) so many songs can be played from RAM before the disk needs to spin up again.

It might speed up the port if you use parts already supported by other players:

http://www.rockbox.org/twiki/bin/view/Main/DataSheets

It's not completely up-to-date. I'll add some more Gigabeat parts to the list later.

[roach]

The problem with users is that they can't be bother to look for themselves.  The iriver port is linked on the front page of Rockbox.com as well as the wiki.

http://www.rockbox.org/twiki/bin/view/Main/IriverPort

Next time look before asking.

Ouch. Well, first, rockbox.com is a GoDaddy parked domain. You must mean rockbox.org.
On the rockbox.org front page there are no links to any iRiver port (unless you mean the "Daily Builds" page?).
And I'm not looking for a port. I'm looking for hardware specs and suggestions (though Rockbox is maybe the wrong place to be looking, but I figured if you're writing iPod firmware, you must know the iPod hardware pretty well, right?). But thanks for the link, I'll take a look at the iRiver port.

FWIW the Ipod is a dual core ARM7 running at <=75MHz.  If thats overkill, your ARM9 core would be roughly 3x more overkill  Should make porting easier though.  It'll be so many times faster then any other Rockbox target you won't really have to worry much about optimization.

LOL  . Great! Guess I saw "dual core" and thought "Intel Pentium" :-). I don't suppose you might know where I could find a BOM? So if the iPod has a dual-core, and RockBox supports the iPod, then I guess dual-core is not as much of an obstacle as MadCow was talking about earlier? I've never worked with dual-core, so I'm curious...

It might speed up the port if you use parts already supported by other players:

http://www.rockbox.org/tw.../bin/view/Main/DataSheets

This is an excellent step in the right direction for me. Thanks Markun!

[Bagder]

IMO, dual-core is far from necessary and Rockbox doesn't even yet run on both cores of the portalplayer chips.

[roach]

IMO, dual-core is far from necessary and Rockbox doesn't even yet run on both cores of the portalplayer chips.

I was thinking ARM9 for the higher speed, if I'm going to be deconding on the controller, rather than using a third-party chip. If I can dedicate an entire core to just decoding, that would be sweet. The advantages to software decoding are pretty attractive, if only for the future-proofing...

BTW, Thanks to everyone for your input!

[edit]: Found a partial dissection of the 5G iPod, if anyone's interested, over at howstuffworks: http://www.howstuffworks.com/ipod.htm

[Bagder]

if I'm going to be deconding on the controller, rather than using a third-party chip

All modern DAPs decode music with the microcontroller. The dedicated chip solutions are all gone since they are too limited.

But yeah, an ARM9 will be plenty powerful enough and a second core is just making things complicated if you can get away with a single one - IMHO.

[plywood99]

Saratoga, you sound like a jerk. Your elitist attitude does nothing. Why not help the guy? I think it is great that roach wants to make a player for Rockbox.

This is my first post but a long time reader of this forum. I have a 2g Nano and patiently await a Rockbox version for it. But some people on this forum are quite snobish to say the least. This is not just my opinion either. Others Iv'e talked to have negative opinions about this forum too. Shame really, Rockbox is great and I appreciate the hard work that goes into making such a project. Some folks just need to chill a little...


Ply

[saratoga]

Saratoga, you sound like a jerk. Your elitist attitude does nothing. Why not help the guy? I think it is great that roach wants to make a player for Rockbox.

This is my first post but a long time reader of this forum. I have a 2g Nano and patiently await a Rockbox version for it. But some people on this forum are quite snobish to say the least. This is not just my opinion either. Others Iv'e talked to have negative opinions about this forum too. Shame really, Rockbox is great and I appreciate the hard work that goes into making such a project. Some folks just need to chill a little...

You registered just to call me a jerk and talk about the Nano port in the wrong thread?  Maybe you're not really helping the content level here yourself.

Anyway I pointed out that we have all the info he asked in the wiki, and then linked it. I'm not really sure how else I"m supposed to help.