Kindle USB Audio Adapter

[NiLuJe]

@geekmaster: Yup, still got that PW1 and that K4, thanks . The PW1 was my main "at home" reader until the PW2 came out, and the K4 was invaluable as a dev device (and was the one that ended up accompanying me everywhere, until the H2O). The PW1 is currently between parenthesis in my device box since I've lent it to my brother, but it still there, alive and well . The PW1 is still terrible compared to the PW2/3, but that's certainly not your doing .

And, yeah, @knc1 is correct, there's a 'plugged in' check to avoid a various collection of weird issues. Depending on devices/FW combo, actual issues may only appear in one 'side' of the equation (i.e., etiher USBMS -> USBNet or vice-versa), so the check has to wrap both :/.

@knc1: I've always felt a bit iffy about donations in this specific context (I might have written a bit more about that at one point), so, no PayPal . But if it were to happen, I'm afraid that, yeah, I'd need a "prebaked" version when it comes to serial ^^.

[knc1]

Originally Posted by NiLuJe

- - - -

@knc1: I've always felt a bit iffy about donations in this specific context (I might have written a bit more about that at one point), so, no PayPal . But if it were to happen, I'm afraid that, yeah, I'd need a "prebaked" version when it comes to serial ^^.

I am OK with that position.

I have some practice shipping to England, I guess France shouldn't be that much different.

And I can 'pre-bake' it.

I'll start a thread on the subject but handle any transactions by PM.
If we get nothing from the members, well, that is the way it goes.

= = = =

It may not be possible to use the same USB port for USBnetworking and as an OTG port at the same time.

But even so - the USBnetworking package needs to either check for or to over-ride the kernel's module set.

And we really can't ask you to fix something you don't have the hardware for.
(Besides your going to like this screen. )

[geekmaster]

There was a collection for YifanLu back in the day, and he used it to buy a kindle (greatly contributing to it getting jailbreak support).

Yeah, I could use kindles newer than PW3 (for support and development), if we have any generous souls out there. But I think our audience is a bit smaller than a few years ago...

[geekmaster]

Okay, I reinstalled usbnet as recommended after a firmware upgrade. Now sshd starts and I get a shell, but the password generated by kindletool does not work. I can still login with telnet, but I see /etc/shadow contains no password hashes. I was also told that the root account is locked. So should I create a root password from telnet then, or is there something else I should do instead?

EDIT: Okay, I made backup copies of /etc/shadow and /etc/passwd, then I changed the password to what kindletool said. Passwd says "too weak" but accepts it anyway. Now I can login with SSH. Even better, I can leave it ALWAYS in usbnet mode, and use SCP to transfer files. No more fighting with the unreliable touchscreen. Though this makes it unreliable as a reading device, it is fine for development and testing purposes (at least for apps that do no depend on a touchscreen). And this makes me want to add mouse (and mouse cursor) support anyway.

[knc1]

/etc/passwd:

root::0:0:root:/tmp/root:/bin/sh

also change the home directory if different on the K1
Usually, the above means just delete the character between the first and second colon.

login: root<cr>
no password required

You can do the above with telnet/vi (or whatever editor is on a K1).

[geekmaster]

Originally Posted by knc1

/etc/passwd:

root::0:0:root:/tmp/root:/bin/sh

also change the home directory if different on the K1
Usually, the above means just delete the character between the first and second colon.

login: root<cr>
no password required

You can do the above with telnet/vi (or whatever editor is on a K1).

So apparently that exclamation mark between the first two colons disabled the account?

Except I need a password if I am going to be using SSH over wifi, right? Or can that be an empty password as well? In any case, it is now my "fiona" password...

[knc1]

do the above
setup pubkey auth.
(works for both cable and wifi)
undo the above passwd change if really concerned about a very dirty wifi environment.

[Marco77]

I uploaded the kernel source on Github for 5.6.5, 5.7.3.1 & 5.7.4.1

https://github.com/mchubby-3rdparty/...rivers/usb/otg

Apparently the new .ko module is created because the MXC_OTG config item is selected in https://github.com/mchubby-3rdparty/...ario_defconfig.
At some point scripts/gen_initramfs_list.sh includes both fsl_otg_arc.ko and ehci-hcd.ko in the intird
https://github.com/mchubby-3rdparty/...initramfs.list

[geekmaster]

Originally Posted by Marco77

I uploaded the kernel source on Github for 5.6.5, 5.7.3.1 & 5.7.4.1

https://github.com/mchubby-3rdparty/...rivers/usb/otg

Apparently the new .ko module is created because the MXC_OTG config item is selected in https://github.com/mchubby-3rdparty/...ario_defconfig.
At some point scripts/gen_initramfs_list.sh includes both fsl_otg_arc.ko and ehci-hcd.ko in the intird
https://github.com/mchubby-3rdparty/...initramfs.list

I hope we can port this to the older soundless kindles, starting with the K4. Even if we need to flash a new kernel, that would be fine as long as the user agrees. The "uninstall" could always flash the old kernel back to allow an official amazon FW update, if necessary...

But I thought the whole POINT of kernel modules was to add or change kernel features WITHOUT needing to replace the kernel, so hopefully all we need for sound support is a set of kernel modules and a startup script to load them.

With the cheap (88-cent with free shipping) USB sound adapters work find on my PW3, using the OTG module that comes with the latest amazon firmware, porting this to earlier soundless kindles seems like a worthwhile pursuit (and easier to achieve now that it was when the K4 was new).

And regarding needing half a GIGABYTE of RAM to support TTS? Ridiculous! I think we can sound at least as good as Stephen Hawking with a mere handful of KILOBYTES of RAM.

Just do it in software, like they do on itty-bitty-RAM devices like Arduinos:
https://github.com/geekmaster/TTS
And that code linked above does PWM sound on a GPIO output pin. Rather like sending PWM sound to a kindle power LED, right?

Which makes me wonder, just how many of those "test point" pads, scattered all over the front (and many more on the back) of modern kindle internal circuit boards, are connected to unused GPIO pins on the SoC, just waiting to talk to hackish things like bit-banged SD cards, PWM sound (without USB adapters) and whatever else imagination can provide...

[knc1]

Originally Posted by Marco77

- - - -
At some point scripts/gen_initramfs_list.sh includes both fsl_otg_arc.ko and ehci-hcd.ko in the intird
- - - -

It has to, if it is going to use 3G.

The external USB connector is not the only USB port on the SoC.
The USB port that the 3G module (and hub) is on, is a host port.

[Cinisajoy]

Originally Posted by geekmaster

I still have not installed the TTS voice pack on my PW3. After all, I am going deaf, not blind. Apparently, deafness is not peaceful silence. I have lawnmowers and whistles and vacuum cleaners and sirens all blaring in my head all the time (making peaceful sleep difficult to acquire). Not that imaginary noise keeping me awake at night matters that much, when I have to get up and pee every half hour. Getting old sucks -- avoid it if you can. Yeah, that, only worse than Beethoven described. I still love my music, but some notes cause pain and others I cannot hear, and I cannot tell the difference between high-end audio gear and dollar store headphones. And the difference between hearing threshold and pain threshold is nonexistant for many frequencies, making listening to music something I like to save for concert-going these days.

USB audio for sound-deficient kindles? Great! But not all that useful for me, other than a technical curiousity.

You have my complete and totally sympathy. I always hear the fan noise. It doesn't matter the size of the fan. Even the fans with no visible fan blades. You know the Dyson ones. I can hear their internal fan.

[Doitsu]

Originally Posted by geekmaster

IT'S ALIVE! IT TALKS!

I did not install the voice update yet.

I just plugged in my cheap chinese $0.99USD (free shipping) USB sound dongle, with heaphones already attached, and my kindle immediately started talking to me about "no voice files detected" and such.

Maybe they've taken a page from Apple and include a low quality voice and offer high quality voices as optional downloads. Or the low quality voice was included as a fallback voice in case the user accidentally deletes the downloaded TTS files.

Theoretically this lower quality tts voice would also require less RAM and should also work on Kindles with less memory such as my PW2. (Of course, it's also possible that they only included a TTS generated audio file of that specific error message.)

Either way, I'm curious what the sound quality is like.

If it's not too much of a hassle, could you please record the error message that you hear when you attach your usb audio adapter?

[HarryT]

Originally Posted by Doitsu

Maybe they've taken a page from Apple and include a low quality voice and offer high quality voices as optional downloads. Or the low quality voice was included as a fallback voice in case the user accidentally deletes the downloaded TTS files.

If it's only used to report an error (as appears to be the case), it would not of course need to be a synthesised voice, but simply a recorded error message.

[geekmaster]

Originally Posted by Doitsu

If it's not too much of a hassle, could you please record the error message that you hear when you attach your usb audio adapter?

It sounds human. I strongly suspect they hired a professional voice actor to read that message, in which case posting it could cause MobileRead to receive a DMCA takedown notice (if the local forum moderators didn't beat them to it)...

[Barty]

Originally Posted by geekmaster

And regarding needing half a GIGABYTE of RAM to support TTS? Ridiculous! I think we can sound at least as good as Stephen Hawking with a mere handful of KILOBYTES of RAM.

That's for the new voice that sounds a lot better than the Hawkings voice. But who knows, maybe Ivona does not need that much RAM either and Amazon just does not want to support old devices.