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

[JimZipCode]

Someone did something similar:

Raspberry Pi zero portable music player
https://hackaday.io/project/26157-pipod

Cost: roughly 47 pounds plus the price of a 64Gb SD card.

Not clear to me if this has a headphone jack, or if it just streams to an output device

[Bilgus]

https://cdn.hackaday.io/images/5173271505765650120.JPG

[JimZipCode]

https://cdn.hackaday.io/images/5173271505765650120.JPG

It's a little clearer to me now.   

[JimZipCode]

remember Rockbox has only very crude support for touch screen devices (it is made for physical buttons). 

I feel like physical buttons are more appropriate for this application.  Skip a song etc; a lot of the things one wants to do with a player, are things you shouldn't need to look at the player for.

Also, on a small player (say around Sansa Fuze+ size, or Agptek Rocker; let alone a Sansa Clip) getting my sausage fingers to touch the screen in EXACTLY the right place to do the action I want, would probably be a nightmare.  It just doesn't seem the right use-case for a touch screen.

[progmars]

While Raspberry Pi Zero and similar generic board solutions (C.H.I.P. etc.) might be good for experiments and DIY players, we shouldn't think that direction if the goal is to create something that could be mass-produced.

I'd say, the optimal solution might be to find some way to work together with some large Chinese company that already has experience and resources for mass-production (AGPTek? Somebody else?). Then the RB community experts might suggest choices of components that are the easiest to deal with in current Rockbox and the company could try to do the investigation from their side to see which component choices are the best options for them for various reasons (sourcing, prices, previous experience etc.).

But to make it possible, the company should be convinced that it is in their interests to develop such players. Some Chinese companies seem to be afraid to deal with open-source and actively contributing back, even if initially it might be in their interests. Why not leverage the fact that Rockbox is well known, popular, stable? There is no need to develop their in-house firmware.
But a large company of course will already have their own firmware and they might be interested in developing and maintaining it instead of Rockbox for various reasons. Marketing is one of the biggest "bad guys" - if the marketing department decides to have some specific and maybe unique feature on the device, then they won't consider Rockbox or open source solutions because they are afraid from competitors. They prefer to develop a buggy feature themselves and sell it instead of letting open-source community to develop it and improve. That's the dark side of the corporate world.

So, the question boils down to this: where to find a large company who already has experience and resources and willingness to use Rockbox as their firmware and contribute back?

For my own personal desires - I need a replacement of Clip+ - player, FM, voice recording, small size, physical buttons, a clip. All of it and for a reasonable price (under 50 $). Don't care about HiFi and aluminium. Just a fully functional player for an active person.

[uhmgawa]

While Raspberry Pi Zero and similar generic board solutions (C.H.I.P. etc.) might be good for experiments and DIY players, we shouldn't think that direction if the goal is to create something that could be mass-produced.

Mass production assumes there is a market to consume that volume.  In 2018 does one still exist?

For perspective this thread started 12 years ago and I jumped into the discussion when I was designing an imx433 linux based player.  Today we could find an MMU-less 100Mhz SoC which has enough bandwidth to achieve an acceptable integer based decode.  However I don't even know if integer decoders are available for all current formats of interest as my now horribly dated reference had been libmad and vorbis.

I fear at this point very few actually give a hoot as most of the music listening population have some sort of smartphone/tablet/whatever attached to their being during all waking hours.  Do you need a 64-bit 1Ghz ARM core to decode music?  Of course not but "I already have one" so do you make a case for a more optimally engineered solution when the proposition boils down to carrying around something extra and less able than what people already have?  The argument of "but longer runtime for the dedicated player" falls flat when life for most stops when the battery in the phone dies.

Unfortunately and regrettably, at this point I think the market has come an gone for a more practical solution.

[progmars]

Yeah, for the most part demand for mp3 players has gone down rapidly since smartphone revolution.
Still, when entering "mp3 player" in aliexpress, the number of results is overwhelming, so obviously some companies somewhere are still making money on them.

For me personally, a dedicated mp3 player is a good option when taking long walks somewhere where there is a chance to be robbed, especially if somebody sees you with headphones. And of course physical buttons and blind control is much more convenient than a smartphone, even with hands-free headphones. But I guess, I'm a minority and majority of modern civilization can't imagine leaving their phone home. So, mp3 players have become a niche product.

[coacharnold]

I'm going to have to agree .... I have a player attached to me belt clip as we speak ... and a phone sitting on the desk in front of me.  I prefer having a standalone player that I can hook into a portable headphone amp and listen to.   I don't have to deal with any of those connections when I pick up my phone to look up something or to read the news while listening.   

Forget looking up on alien express, I have a FIIO player that I use know bacause the DAC is way better than whats on board my rockbox players.   this player is LOW on the totem pole from very nice portable players that are out in mainstream, running from 100$ to 2000$ dollars.  I'd kill to have a standalone player that I could plug a usb dac/amp into.   This project is super relevant .

10 years ago this was about hacking cheap 2nd and 3rd generation players but now it's very much a platform that shouldn't be ignored.

The rockbox UI is still better than the one on my fiio or any of those players on alien express ..... or any of the expensive DAPs out there now.

tim

[uhmgawa]

This project is super relevant ... it's very much a platform that shouldn't be ignored.

If you believe that, you have a pure marketing problem on your hands.  The engineering is comparatively trivial.

Just because a group of potential users are adamant a product should exist, does not make it a commercially viable proposition.  Even if a group of capable users vowed to support this as a labour of love until their last breath, prerequisites exist beyond your control: production life of the commodity SoC, display, battery, etc.. being leveraged.  Unless you are prepared and financially able to maintain the infrastructure to sustain all of them forever.

If you can create a market able to justify such an effort at both an agreeable entry and sustaining price point, you certainly have a compelling case.  But in this day and age it sadly is extremely unlikely to happen.  Witness alone the 12 year life of this thread, where the beginning arguably coincided with the greatest potential interest which has been declining ever since,

No I don't particularly like it either but I've accepted the practical reality.  If you disagree, I'd suggest some type of community kick-start vehicle to assess the potential.


 

[coacharnold]

I really don't care about being commercially viable... Why on earth does this forum exist... It's for rock box..  There has been virtually no new devices in a very long time. This is a FUN DIY project... I'd like to build my own superior DAP...  Much like the Rock Box DAPs I was using 10 yrs ago.   If you think that's silly.  Then shut this site down what's the point

[uhmgawa]

I really don't care about being commercially viable...

Let me rephrase more clearly the concern as being economically viable.  If this was an inexpensive proposition, those impassioned with the concept and so inclined would have already done so.

Why on earth does this forum exist... It's for rock box..  There has been virtually no new devices in a very long time. This is a FUN DIY project... I'd like to build my own superior DAP...  Much like the Rock Box DAPs I was using 10 yrs ago.   If you think that's silly.  Then shut this site down what's the point

I'm uncertain how you read an assessment of the effort being "silly" from my comments.  But to be clear my point was only the financial and logistic burden of implementing the device effectively and sustainably.  Case in point, my suggestion of a kick-start approach to estimate the potential enthusiast base.

[RandomInsano]

Exhuming this thread. I won't make any promises, but if I made five Raspberry Pi based daughterboards would there be devs willing to port Rockbox to the BCM2835 that's in the Raspberry Pi Zero? I built a prototype tonight and am streaming music on it in Linux now. The Pis will be hard to get...

https://imgur.com/a/svWt4Dq

I bought some cheap junk off the Internet (PCM5102 I2S DAC, SSD1306 OLED module) that are widely used by Raspberry Pi and Arduino users and I think is what's in the Hackaday article. With some light electrical engineering it's possible to make something smaller than the volume of an iPod classic. I don't expect it can be done for less than $50 USD because people's time costs more than the bill of materials but I paid ~$40 CAD just buying off Amazon (already had the Pi)

The *real* question I think is does anyone want to try and port Rockbox natively to the Pi? There are bare-metal toolchains now (I use a different Pi to emulate a Commodore 1641 disk drive https://cbm-pi1541.firebaseapp.com/) and it's well documented. If there's interest I may just build a few with these parts and send them out. I would only send to active port maintainers though.

I'll float the hardware side of this idea past some friends to see if they'd want to tackle that side of it, but as it stands I built mine in an hour and didn't need a daughterboard.

* https://www.digikey.ca/en/products/detail/texas-instruments/PCM5102APW/3902495 (~$7 in orders of 100)
* https://www.digikey.ca/en/products/detail/midas-displays/MCOT128032CY-WS/15854837 (~$11 in orders of 100)

If there are better options for screen and DAC I'm open to suggestions but I want them to already be in use with the Pi and have their source documented.

No promises at all, but I think this could be fun.


Back the next day with some adventures. This device from Pimaroni is about 66% of the way there (need a battery) to give an idea of how small it could actually be:
https://shop.pimoroni.com/products/pirate-audio-headphone-amp?variant=31189750480979

[speachy]

TBH running this as a native Linux player ought to work quite well; we could use the SDL port directly on top of the framebuffer (ie no X), so all that really needs doing is working on the mechanicals

What's your proposed solution for a battery?   That's one area where the RPi hardware falls pretty flat; IIRC it needs 5V (and is relatively thirsty, with poor platform power management) with no provision with the SoC for proper battery management, so you'd need an external BMIC. And it would need to be a relatively beefy battery too.

That pirate audio amp board is pretty sweet though!

[john.cooper]

I've tossed around using RBPi0-W as an audio platform in the past.  It isn't practical for a battery powered device but if one were going to use it, say in a vehicle, that's a different story.  These days auto manufacturers appear to have woken up and are covering that base to overkill.  But it is fun to experiment and if you're a daily commuter there can be some satisfaction in that approach.  As you're already paying the architectural and power consumption tax for a real SoC able to support a virtual memory POSIX runtime, why not use the resources to a unique differential?

The RBPi0-W is network connected and with a Raspian install you have ssh access into your audio  development / playback system, with a full toolchain, GIT, and anything under the sun which suits your technical fancy.  From an engineering geek perspective, little else out there can compete with this matchbox sized platform.  About the only bolt-on hardware needed is I2S and UI support.  And even those can be kludged for a prototype.

[coacharnold]

I'm so ready for this,  this is the part of this project i have been eagerly awaiting to come back to life.    I don't write code,  but I'm happy to invest in hardware to help test.  PLEASE LETS GET THIS ROLLING.

There has to be a battery solution?  I'm thinking 10 hrs of life is perfect.   no one makes something for RbPie0 that can handle what we need???

That pirate board is really cool.  Any idea if it works with squeezelite?

Tim
NJ, USA