Suggestion for upcomming hardware versions: HDMI audio

+1

5.1 so I can duct tape my surround system to my head.

There are two primary things why this is a bad idea:

1) It will cause the HMD to have an increased price, no matter how small it might be. FB/OVR are, at the moment, trying to get the best possible VR display device for as cheap as possible. Currently, they do have an added USB port, and there is a 1/4" hole (presumably, designed for a headset, but this is speculation.. it could be a vent as well)

2) Added weight.. it may not seem like such a big deal, however, we need the headset to be as light as possible. Ideally, it wouldn't weigh anything, damn that Newton fellow and his physics ><

In the end let the users decide what type of audio they want. Some entry level people may be fine with standard headsets, but others may want a super-high end headset that costs $300 (such as the Astro A50's). If they want HDMI audio, let those people get a sound card and headset that supports it.

The OVR DK2 uses a standard HDMI receiver chip, namely the Toshiba TC358779XBG, which already provides demultiplexed, decoded audio on a I2S interface. Datasheet: http://www.semicon.toshiba.co.jp/info/docget.jsp?type=datasheet〈=en&pid=TC358779XBG

TI has this nice little I2S interfaced audio DAC with integrated Class-D headphone amplifier TLV320DAC26, which you can connect directly to the HDMI receiver. Datasheet: http://www.ti.com/lit/gpn/tlv320dac26 Similar choice for integrated audio output solution, that offers slightly more features: TI LM49450

The control SPI interface (for setting the volume) can be trivially driven by the already present STM32L100 that's right now used for just reading out the inertial sensors. The sensors use SPI as well, so even if all the available SPI peripherals of the STM32L100 are already used, you can simply reuse that SPI; the STM32Fs don't have dedicated Chip Select pins anyway, you use an arbitrary GPIO for that. And since the audio parameters are set only once and then ocassionally this also doesn't interfere with the tracking reads.

"saviornt" wrote:
There are two primary things why this is a bad idea:
1) It will cause the HMD to have an increased price, no matter how small it might be.
(…)
2) Added weight.

There, I just added only a single IC, that costs 2.57$ @1ku to the BOM and I gave you an integrated headphone amplifier for the HDMI audio signal. The IC weighs less than 1g

Do you want me to draw the schematics for you?

"saviornt" wrote:

In the end let the users decide what type of audio they want. Some entry level people may be fine with standard headsets, but others may want a super-high end headset that costs $300 (such as the Astro A50's). If they want HDMI audio, let those people get a sound card and headset that supports it.

The HDMI signal already goes into the OVR. It would be trivial to present the carried SPDIF signal on some accessory plug. But I'd rather add <3$ single chip solution to the BOM.

"datenwolf" wrote:
There, I just added only a single IC, that costs 2.57$ @1ku to the BOM and I gave you an integrated headphone amplifier for the HDMI audio signal. The IC weighs less than 1g

Do you want me to draw the schematics for you?

It has been some time since I took a math class, however, I am pretty sure it goes something like this:

0g < 1g
$0.00 < $2.57

I could be mistaken though. As I have said, I am getting pretty "long in the tooth" as the saying goes.

I would like to see some schematics actually. I enjoy building guitar distortion pedals with spare transistors so this kind of stuff interests me.

"Oculouse" wrote:
I would like to see some schematics actually. I enjoy building guitar distortion pedals with spare transistors so this kind of stuff interests me.

Actually on the analog side there's not a lot of interesting stuff going on. Either IC I linked is a fully integrated audio solution. A digital signal with a I2S wire format goes in, an amplified analog audio signal with enough power capability to drive a headphone speaker comes out.

Unfortunately the HDMI receiver used in the DK2 has a BGA package so it's virtually impossible to perform a HW modification. At least not if you don't have a BGA rework station.

If the HDMI receiver had a package with easily accessible pins (a QFP variant for example) it would have been feasible to use some fine varnish insulated copper wire to connect the I2S signals to one of those audio solution ICs, you'd glue on the back to the PCB (dead bug construction method).

Interesting, thanks.

"saviornt" wrote:

"datenwolf" wrote:
There, I just added only a single IC, that costs 2.57$ @1ku to the BOM and I gave you an integrated headphone amplifier for the HDMI audio signal. The IC weighs less than 1g

Do you want me to draw the schematics for you?

It has been some time since I took a math class, however, I am pretty sure it goes something like this:

0g < 1g
$0.00 < $2.57

I could be mistaken though. As I have said, I am getting pretty "long in the tooth" as the saying goes.

So there is no value/cost tradeoffs in your mind? It has some cost, so it is flat out not worth it, even though the price is extremely minor?

Not really, no. Added benefits over the cost just wouldn't be there.

Just a guesstimate here, but why put it in when 75% or more of the people that use the headset won't use it? It would be just added fluff.