What Will Be The Best Audio Out?

[TrashyMG]

As I said, it's likely that the integrated DAC is just used for line-in.

Then wouldn't it be an ADC?

 

[mali]

^It's not my métier and I don't have much time at hand right now, check this for further info:
http://focus.ti.com/docs/prod/folders/print/tps65950.html

 

[shunun]

Can we expect a port for Audacity ?? Now what about external microphones, how can we connect one of those? Oh and what about recording quality? What kind of sample rates can be expected?

 

[mali]

^You got line-in on the ext port and microphone on the headset port. In those posts I mentioned were infos about sample rate too, iirc. The datasheet of the companion chip should contain more infos.

 

[TrashyMG]

^It's not my métier and I don't have much time at hand right now, check this for further info:
http://focus.ti.com/docs/prod/folders/print/tps65950.html

That chip seems to have a lot of different functions and seems to do the dirty work of all the small things the pandora does.. Here is a snippet from that PDF.

The TPS65950 device is a highly integrated power-management and audio coder/decoder (codec)
integrated circuit (IC) that supports the power and peripheral requirements of the OMAP™ application
processors. The device contains power management, an audio codec, a universal serial bus (USB)
high-speed (HS) transceiver, an ac/USB charger, light-emitting diode (LED) drivers, an analog-to-digital[
converter (ADC), a real-time clock (RTC), and embedded power control.
The power portion of the device contains three buck converters, two controllable by a dedicated
SmartReflex™ class-3 interface, multiple low-dropout (LDO) regulators, an embedded power controller
(EPC) to manage the power-sequencing requirements of OMAP, and an RTC and backup module. The
RTC can be powered by a backup battery when the main supply is not present, and the device contains a
coin-cell charger to recharge the backup battery as needed.
The USB module provides a HS 2.0 on-the-go (OTG) transceiver suitable for direct connection to the
OMAP universal transceiver macrocell interface (UTMI) + low pin interface (ULPI) with an integrated
charge pump (CP) and full support for the carkit Consumer Electronics Association (CEA)-936A
specification.
The Li-ion battery charger supports charging from ac chargers, USB host devices, USB chargers, or
carkits. The type of charger is detected automatically by the device, which provides hardware-controlled
linear charging with ac chargers, USB chargers, and carkits, in addition to software-controlled charging for
all charger types.
The audio codec in the device includes five digital-to-analog converters (DACs) and two ADCs to provide
multiple voice channels and stereo downlink channels that can support all standard audio sample rates
through several inter-IC sound (I2S™)/time division multiplexing (TDM) format interfaces. The audio output
stages on the device include stereo headset amplifiers, two integrated class-D amplifiers providing stereo
differential outputs, predrivers for line outputs, and an earpiece amplifier. The input audio stages include
three differential microphone inputs, stereo line inputs, and interface for digital micrphones. Automatic and
programmable gain control is available with all necessary digital filtering, side-tone functions, and
pop-noise reduction.
The device also provides a auxiliary modules, including LED drivers, and ADC, keypad interface, and
general-purpose inputs/outputs (GPIOs). The LED driver can power two LED circuits to illuminate a panel
or provide user indicators. The drivers also provide pulse width modulation (PWM) circuits to control the
illumination levels of the LEDs. The ADC monitors signals entering the device, such as supply and
charging voltages, and has multiple additional external ADC inputs for system use. The keypad interface
implements a built-in scanning algorithm to decode hardware-based key presses and to reduce software
use, with multiple additional GPIOs that can be used as interrupts when they are configured as inputs.

 

[mali]

Page 3 from this document:
http://focus.ti.com/lit/ml/swcs034a/swcs034a.pdf

Audio subchip:
The audio subchip has two interfaces: PCM for voice signaling and I2S for audio signaling. The Bluetooth
interface is also available. This uses the same clock and sync signals as for the PCM interface. Both ports can
function as master or slave.
Sampling modes available are:
Voice channel with 8-kHz or 16-kHz sampling modes
Audio channel with 8, 11.025, 12, 16, 22.05, 24, 32, 44.1, or 48-kHz sampling modes. 96 kHz is supported
on the RX path voice codec with MCLK = 26MHz.
The audio phase-locked loop (PLL) supports 26-MHz, 19.2-MHz, or 38.4-MHz ability to work with two ports with
the clock frequencies 19.2 MHz or 26 MHz, or with a single TDM port with the frequencies 19.2 MHz, 26 MHz, or
38.4 MHz. The voice PCM interface is available when the system clock is 26MHz.
The audio supports common features like pop-noise reduction, sidetone functionality, bass boost function, uplink
and downlink programmable gain amplifiers, and the DTMF tone generator.
The external vibrator control is provided in this module and can be controlled through an audio signal or direct
I2C writes to the registers.

I'm not completely sure of the input specs through this chip. All the datasheets and the TRM has finally become public so you can have a look for yourself: <a href="http://focus.ti.com/docs/prod/folders/print/tps65950.html" target="_blank">http://focus.ti.com/docs/prod/folders/print/tps65950.html</a>

That is the catalog replacement for the TWL4030 (like the 3530 is the replacement for the 3430).

I had asked Squidge to investigate this setup many months ago, but I don't think anything ever came of that. I see 8 or 16KHz sample rates selectable for voice, but that is for the dedicated voice channel to Bluetooth which is not used on Pandora. Audio to the Bluetooth chip comes from the OMAP chip directly over I2S which has much better sound quality options. I don't feel these low sample rate options apply to Pandora (see page 684 of swcu050b.pdf).

I see that the microphones and line inputs can be muxed over to the internal ADCs that connect to the main I2S interface back to the OMAP chip. This supports 13-16 bit data (I can't find a specific reference to quantization anywhere except 16-bit listed on the website so that's what I will go with) and up to 48KHz sample rates (96KHz is the max when transmitting audio only so doesn't apply here). I am assuming that this path can be used to get the best audio back to the processor. In fact, I felt confident that this is how it works when I laid out the design. I just need someone to actually test it.

edit:
The quote I meant from MWeston:

To MWeston - can you say what audio pin's on the TWL are used for pandora's audio IOs?

Internal microphone - TPS65950, MAIN MIC
External microphone (on headphone connector) - TPS65950, SUB MIC
Line Inputs - TPS65950, AUX INPUTS (shared with mics inside TPS65950 so you can set muxes to use one line and one mic channel if you want)
Line Outputs - PCM1773, line level output shared between headphone amp and EXT connector, no volume control (2.15Vpp output, 10K load)

Audio inputs use McBSP4 and audio outputs use McBSP2 so they can run simultaneously with different audio on them. BT is on McBSP3 for reference.

The audio output of the TPS65950 is not connected on the PCB but I read in the TRM it might be possible to route it out of the USB OTG port for four channel audio (six channel if you stream over BT too). That might take some clever coding to hack into the kernel, but neat if it works some day. I could be wrong about it working like this, but I found information in that carkit section for USB.

 

[Bosbeetle]

Line Outputs - PCM1773, line level output shared between headphone amp and EXT connector, no volume control (2.15Vpp output, 10K load)

So ext it is. Thats the most pure signal you can get than I let the amplifying to my stereo.

I am convinced I'll add a tv out cable to my order

thanks mali

 

[mali]

Thank urjaman and peca, who already gave the right answer
It's always good to back it up with sources, though.

 

[peca]

Thank urjaman and peca, who already gave the right answer
It's always good to back it up with sources, though.

Sources You say?
---
autor: MWeston; datum: 2008-11-05 06:22 AM CET
autor: MWeston; datum: 2008-11-06 01:53 PM CET
autor: MWeston; datum: 2009-12-01 02:38 AM CET

 

[mali]

Peca, I can't wait to see the results of your dissertation. It will be very informative, I suppose
It would be cool if you could make an English version of that diagram, so that we can put it in the photobucket

 

[BaDToaD]

It would be handy it OP made a short decent quality EXT to 3.5 mm jack cable. I'd happily pay £10-15 for a decent cable so I could bypass the headphone out straight into a portable amp.

 

[rabidpoobear]

OP's supposedly gonna sell EXT connectors so you could quite easily chop the end off a high quality 3.5mm and solder it onto the EXT with some heat-shrinkage and a stress relief. I want a lot of EXT connectors, I'll probably buy at least 10.

[ed] if anyone's found the EXT in small quantities already, let me know so I can order some. Last I heard you had to get runs of 10k or something like that.

 

[peca]

Peca, I can't wait to see the results of your dissertation. It will be very informative, I suppose

Nah, Your expectation seems to be too high. It may be informational for someone who doesn't follow forum and don't eat every bit of info as we do

It would be cool if you could make an English version of that diagram, so that we can put it in the photobucket

Yep, that quite easy, why I don't this earlier !?

edti2: I'm author of this picture and anybody can use it for anything.

EDIT:
@rabidpoobear: IIRC MWeston say that they buy some spare connector for selling to as. I want a few too.

 

[mali]

Great, thanks

 

[rabidpoobear]

EDIT:
@rabidpoobear: IIRC MWeston say that they buy some spare connector for selling to as. I want a few too.

TY for the info Peca, I heard that too. I was hoping that maybe someone knew if Digikey had small quant's up for order or something.

 

[WizardStan]

It would be handy it OP made a short decent quality EXT to 3.5 mm jack cable. I'd happily pay £10-15 for a decent cable so I could bypass the headphone out straight into a portable amp.

go to the dollar store (or the UK equivalent) and buy one of those 3.5mm-to-RCA adapter thing. I just saved you £9-14 (except you still have to buy the TV out cable, so I guess I really cost you an extra 1)

 

[cobaltage]

As Nikkopt mentioned, in principle, you want only one amplifier in an audio signal chain to get the "best sound." The headphone jack amplifies the signal (for the headphones, which are "passive monitors"), while a line-level out is unamplified and more suitable to routing the signal to, let's say, a stereo receiver, mixer, or standalone headphone amplifier. If you route the signal from the headphone jack to a stereo receiver, it goes through two amplifiers, each of which modifies (distorts) the signal.

In practice, most portable audio devices don't even have line-level out, and people make do with routing the signal from the headphone jack to whatever (car stereo, receiver, etc.). I've read that one can optimize the volume of the headphone out in order to reduce the distorting effect of the headphone jack's amplification, that is, to make that signal act more like a line-level signal. I don't know if, or to what degree, that's really true, though.

If you look at audiophile iPod mods, they bypass the headphone jack by using an iPod docking port to get a line-level out, and route the signal to a high-end headphone amplifier, pre-amp, or receiver (using cables that are as thick as your pinky). In that kind of mod, the goal is to bypass the internal headphone amplifier entirely. But, audiophile iPod mods don't do anything fancy like routing a digital signal from the iPod to a free-standing DAC -- the line-level out from the iPod's internal DAC is considered to be a perfectly good audio source for later amplification by ultra-expensive hardware. (Otherwise, audiophiles wouldn't use the iPod in any kind of set-up.)

I suspect that the Burr-Brown DAC module also included superior amplification -- that is, that the Burr-Brown DAC just provides a better analog signal path for headphones than could be achieved with the PCM plus a headphone jack. Certainly the audio specifications suggest that that was the case, since the specs on the headphone signal are extremely promising. I would be curious to know if there would have been much of a difference between the PCM line-level out and a line-level out from a Burr-Brown DAC, if it had been set up to provide one.

Edit: The audiophile iPod mod I'm thinking of is called the iMod by Red Wine Audio, by the way.

 

[BaDToaD]

Edit: The audiophile iPod mod I'm thinking of is called the iMod by Red Wine Audio, by the way.

I've heard good things (no pun intended) about the imod but isn't it limited to older ipods?

Back on topic the headphone out should still be of good enough quality to use with an external amp provided you can except a little bit of colouration from the pandoras built in amp. I may not even bother with an EXT cable if the headphone socket produces a decent sound. It may be good enough for on the move use. When I'm stationary but away from a decent music source I hope to run my Go Vibe DAC from the USB on the Pandora which negates the use of the EXT socket.

 

[nikkopt]

Well, i'm glad i was wrong about the line out, using the DAC on the TWL.

^Yes, there is an external amp reducing the quality. I don't know how we could compare the companion DAC to the amplified Burr Brown, but the integrated DAC was bad enough to justify adding that Burr Brown in the first place. As I said, it's likely that the integrated DAC is just used for line-in. I'll search for info this evening, if nobody else does it in the meantime (the info might even be on the other boards). Nikkopt's quote from notaz is _very_ old.

On the TWL: 16-bit Linear Audio Stereo DAC (96 kHz, 48 kHz, 44.1 kHz, 32 kHz, and Derivatives) not that bad, some onborad audio cards are still 16-bit

Anyway, saying that the amp reduces the quality can be subjective, without it you would hear nothing (headphones) . Yes, it does degradate the signal if you amp it again, like connecting it to your stereo, as Bosbeetle will do.

cobaltage it's funny as you can still see lot's of people hanging around with an iPod and an ext amp with 2 times the size of the player, connected to the headphone out, and they are proud of themselves as they hang around with a pair of $1K audiophile headphones ;P

Edit: BaDToaD, the internal amp is cool (150mW on a 16ohm load) unless you have some 600ohm headphones, removing the "ultra-portable" from the "pandora"..

 

[cobaltage]

Edit: The audiophile iPod mod I'm thinking of is called the iMod by Red Wine Audio, by the way.

I've heard good things (no pun intended) about the imod but isn't it limited to older ipods?

Back on topic the headphone out should still be of good enough quality to use with an external amp provided you can except a little bit of colouration from the pandoras built in amp. I may not even bother with an EXT cable if the headphone socket produces a decent sound. It may be good enough for on the move use. When I'm stationary but away from a decent music source I hope to run my Go Vibe DAC from the USB on the Pandora which negates the use of the EXT socket.

I read that the iMod is for older iPods. Apple must have moved something around inside the newest iPods that makes the iMod cable not work safely or effectively or something.

I agree that really, what sort of distortion or coloration you can live with is fine. In my experience, plugging an mp3 player with even a very good headphone out signal into a receiver has pretty variable results: sometimes it's fine, and sometimes the added effects of the two devices sound terrible. E.g., if the headphone jack has a bass cut and a mid level boost, and the next amplifier in the chain has exactly the same qualities, the combined result will almost certainly be awful. To further complicate things, the relative effect of each amplifier on the signal will be dependent on the volume setting of each. If you're talking in-car use, the audio environment of a running vehicle is hardly the kind of place where you're likely to notice relative differences in audio quality. The absence or presence of a flexible EQ of good quality is also a huge factor in what you can do with your set up.

This is one of the things I'll be excited to experiment with when I get the Pandora. There's really no way to know in advance what will work well with any particular set up.

@nikkopt: Some of those mods look completely ridiculous. It's amazing that people actually carry these two to three part Frankenstein's monsters around with them in real life.