HDMI Audio Oddity

https://github.com/MiSTer-devel/Main_MiSTer/wiki

Moderators: Mug UK, Zorro 2, spiny, Greenious, Sorgelig, Moderator Team

Locked
Hewhoisred
Atari freak
Atari freak
Posts: 51
Joined: Sun Dec 09, 2018 6:53 pm

HDMI Audio Oddity

Post by Hewhoisred »

I just built my MiSTer yesterday. It is amazing. HUGE thank you to all of the devs.

FYI: my device is currently without it's SD RAM board, which I will assemble soon. So, I noticed a strange behavior in the cores (the ones that work without the daughter board). When plugged into my PC monitor, via HDMI, audio plays on the TG16, Genesis and ao486 cores, but does not play on any of the arcade or the Atari 2600 cores. What is strange is that, when I plug MiSTer into my living room HDTV via HDMI, all cores play audio just fine.

I'm not sure if this is just a rare quirk of my hardware, but it strikes me as odd that the monitor would play any audio at all. I would assume the problem would affect all cores, but that's obviously not the case.

Any insight on this?
Sorgelig
Ultimate Atarian
Ultimate Atarian
Posts: 6348
Joined: Mon Dec 14, 2015 10:51 am
Location: Russia/Taiwan

Re: HDMI Audio Oddity

Post by Sorgelig »

This has been already discussed in different topics.
Why it happens on some displays currently is unknown.
User avatar
ono
Atariator
Atariator
Posts: 20
Joined: Sun Aug 18, 2019 11:22 am
Location: Switzerland
Contact:

Re: HDMI Audio Oddity

Post by ono »

Sorry to dig out the old subject, that apparently was already discussed, but without clear conclusion what causes the problem.

Can HDMI Audio oddities of 8-bit computer can be related to a bias (aka DC offset) in the produced digital audio signal? I think this is a case for some 8-bit computers e.g. Atari. By bias I mean that audio signal values (e.g. signed integers) are not non-zero mean. This is apparently done to maximize the volume or some other reasons. While modern TVs supplied with analog biased audio signal may behave reasonable, I suspect they also may freak out when supplied with biased digital signal via HDMI - I believe this may be the case on MiSTer where the sound never leaves digital path.

Feel free to correct me if I am talking nonsense though. :lol:

Here's some more read related to Atari 800 emulator where I researched a bit this issue when my Samsung TV was producing noticeable humming when using Atari 800 emulator:
https://github.com/atari800/atari800/is ... -341995810

Image

And some more here: https://manual.audacityteam.org/man/dc_offset.html
Grown up playing/learning to program Atari 130XE. FM synth and tech nerd. | github.com/nanoant
Sorgelig
Ultimate Atarian
Ultimate Atarian
Posts: 6348
Joined: Mon Dec 14, 2015 10:51 am
Location: Russia/Taiwan

Re: HDMI Audio Oddity

Post by Sorgelig »

You need to understand retro sound system. Most old systems produce unsigned samples. It means from 0 to for example 1V. It didn't produce negative wave.
And when it comes to modern systems where signed samples required it may have issues on some of them. You are welcome to improve the cores. This task is not easy.
User avatar
ono
Atariator
Atariator
Posts: 20
Joined: Sun Aug 18, 2019 11:22 am
Location: Switzerland
Contact:

Re: HDMI Audio Oddity

Post by ono »

Sorgelig wrote:You need to understand retro sound system. Most old systems produce unsigned samples. It means from 0 to for example 1V. It didn't produce negative wave.
Yes this is something I've discovered about retro machines when submitting issue report to Atari 800 emu project.
Sorgelig wrote:And when it comes to modern systems where signed samples required it may have issues on some of them. You are welcome to improve the cores. This task is not easy.
I know it is not easy, but it is also challenging and cool to learn about. :mrgreen: By no means I am suggesting what should be done there. I am complete noob in this subject, just wanting to learn more, and maybe have some nice discussion.

Nevertheless, I was just wondering if this may be a root cause of HDMI audio oddities? WDYT? Many HDMI TVs do some audio processing on sound input by default and I can imagine these processing filters can easily go wild and filter out everything when input has large bias from zero mean, while they assume there is no bias at all.

I will study a bit about removing DC bias from the signal, e.g. using some DC trap recommended at some audio forums http://www.cs.washington.edu/education/ ... tion_1.pdf
Last edited by ono on Tue Aug 27, 2019 3:35 pm, edited 2 times in total.
Grown up playing/learning to program Atari 130XE. FM synth and tech nerd. | github.com/nanoant
Locked

Return to “MiSTer”