Rockbox Development > New Ports
Rockbox Player - Project to design and build a Free/Open hardware audio player
squarefoot:
--- Quote from: wodz on September 25, 2017, 01:48:43 AM ---The main problem is RAM size. Rockbox needs more then 2MB of ram (I'd say 4 is bare minimum). Ah, and porting to new architecture (Xtensa) is *a lot of* work.
--- End quote ---
Ouch... Very true about the different architectures, also I didn't realize so much ram was required; this leaves out most cheap low power boards unfortunately; there are some ARM Linux oriented boards that would be more than capable of running Rockbox, though I think they would draw too much power to be considered for a portable player. I played a bit with Orange/Nano PI boards, they're cheap and fun and CPU speed throttling seems to work well under Linux, but I have no idea on how much work would be needed to port Rockbox as a OS-less app to them whilst maintaining enough power management to keep the device portable.
This one for example is ridiculously cheap: http://www.friendlyarm.com/index.php?route=product/product&path=69&product_id=197
Bilgus:
I too have toyed around with building a RB capable player.
Unfortunately, it is really hard to compete with economies of scale.
Even at the $8 price tag for that board you are forgetting about
all the ancillary devices that are needed.
Off the top of my head..
Screen, DAC, amplifier, battery, charging circuit, buttons, case, and whatever else I might have forgotten.
IMO in then end you are probably better off buying a ready made solution.
majianjia:
--- Quote from: Bilgus on September 25, 2017, 03:00:28 PM ---I too have toyed around with building a RB capable player.
Unfortunately, it is really hard to compete with economies of scale.
Even at the $8 price tag for that board you are forgetting about
all the ancillary devices that are needed.
Off the top of my head..
Screen, DAC, amplifier, battery, charging circuit, buttons, case, and whatever else I might have forgotten.
IMO in then end you are probably better off buying a ready made solution.
--- End quote ---
--- Quote from: saratoga on June 23, 2015, 12:45:29 PM ---
--- Quote from: MakingIt on June 23, 2015, 12:06:49 PM ---THis dates from a year ago. Has any work being done at all here? I could quickly do some quick PCBa design around it.
--- End quote ---
I don't think so. I don't even recall seeing anyone use Cortex M processors for a portable audio player yet.
--- End quote ---
It is true, building from zero is time-wasting. :(
But the reason I registered on RB Forum a few years ago and now i am checking this thread is I actually built one of this music player (please see the attaches).
This portable player was built about 4 years ago when I was a univ student who couldnt afford a HIFI player. The core is Cortex-M4F (STM32F4) which i dont think RB has the porting. I am glad if anyone can prove me wrong here.
The most difficult and time-wasting part for me is to program the GUI. I am a "MCU" guy; I can possiblly do everything from layout to mechanics but very poor on GUI programming.
For your interests:
- 168MHz Cortex M4F
- 128k RAM + 64k CCM RAM for MP3, FLAC and APE codec.
- 1M on-chip flash
- DAC: PCM1792
- Power supply: +- 6.5V for amplifiers.
- Independent clock source
- 8M flash for Chinese Font.
- QVGA screen.
- 3500mAh BAT for 10 hour replay.
- ~300g/~120*38*68mm
Currently supports: WAV to 192k 32bit, MP3, FLAC to 96k 24bit, APE 44k 16bit.
Adding codec isnt very difficult. I didnt do it just because i dont need them.
I had been using WAV alone for about 2 years before I addedd other codecs.
It is running with a very old version of RT-Thread, and it's deprecated RT-GUI.
Since it is old, I am looking to built a new one and maybe an fully open source one (including hardwares, and mechanical parts).
I might use a cpu that RB has already been ported on, like Rockchip's CPU, and a case that printed by 3D-printer instead of the aluminum case.
If there is porting already for Cortex M, i will be happy to use Cortex M.
The RAM size and flash size isnt a problem on cortex M4/7. Some of chips can run codes on external QPI-Flash, most of them support external RAM.
From my point of view, a Cortex M4/7 is more than enough to run RB.
Sorry for my English.
coacharnold:
OMG ... thats REALLY cool .... I'm not on the code developers side of things here .... but I'll vote for this .... especially since RB seems to have been slowing down the last couple years.... How much did that cost to build?
COUNT ME IN
T
uhmgawa:
If I were looking for a solution to this in the year 2018, I'd use a Raspberry PI Zero W:
1Ghz ARMv6
512MB DRAM
micro SD, WiFi, BT, uart, gpio, i2c, i2s, SPI
usb host, usb OTG, ...
US$5 at Microcenter (currently price, appears to be a perpetual promotion).
The issues are:
1. building a daughter board to supply display, input controls, i2s audio, voltage regulation and battery maintenance
2. deciding upon a strategy for intake of audio data.
For #2, a minimalist approach would require a usage-friendly method to share the microSD card
between system and music data. Or USB host could be the data intake method. Or a second
microSD on the daughter card (overkill). Or forget about flash and just pull in whatever is needed
via wireless.
Slight modifications could produce an automotive in-dash version.
Could be a fun project if you really want/need a standalone music appliance in 2018+.
Navigation
[0] Message Index
[#] Next page
[*] Previous page
Go to full version