exxos's Atari, Amiga & retro forum

troed wrote: ↑Fri Mar 30, 2018 4:57 pm As to the cause ... well. I don't think it's age related (I remember it from way back when the machines were new) and it's also something that's common to different machines doing different things.

I don't get why my machines don't do this, at least the ones I tested. Wonder if its common on NTSC machines rather than PAL or something ?

I have a new PSU now and fully recapped mainboard and the situation is MUCH better now! Constant high pitched whistling is gone and the audio output is overall cleaner. The first WAV in attachment is right after power on the STe, with real keyboard beep sound (to compare sound level with noise level), second WAV is normalized noise (for better hearing the interferences). I think that the occasional interferences is the problem of all STe's, but there are so quiet that most people can't hear them. Therefore, the audio output seems to be clean to them. But if they would listen the audio through amplifier with higher volume (and with headphones), they would hear the whistling too (no matter what graphic mode and resolution is used, it's always the same). I remember similar high pitched whistling problem on Amiga machine's audio output about 25 years ago.
Besides the audio, I think that the recap also helped to the video. The picture is now cleaner, sharper, more stable and it does not bend, especially in color resolutions.

So my opinion is that it helps a lot, but not solved the problem fully... because it is simply the property of all machines.
Take a look at attached WAVs and let me know your opinions.

Yeah the infamous 2600Hz is still there - as well as the video frequency (~15.7kHz). Depending on which quiet section I analysed the first harmonic at 5200 was more or less pronounced.

Has anyone tried my DAC-fix kit to see if it helps with this problem or not yet ? (lost track sorry!)

exxos wrote: ↑Thu Apr 05, 2018 9:11 pm Has anyone tried my DAC-fix kit to see if it helps with this problem or not yet ? (lost track sorry!)

No, I don't tried your DAC-fix kit yet. Is it difficult to apply it? Too much soldering? Unfortunately, I don't have an advanced skills with it.
By the way... may you take a short recording of audio output from your STe for comparison?

Thanks a lot!

Paulo P wrote: ↑Fri Apr 06, 2018 6:54 am By the way... may you take a short recording of audio output from your STe for comparison?

Install guide is here..
http://www.exxosforum.co.uk/atari/last/S ... /index.htm

Paulo P wrote: ↑Fri Apr 06, 2018 6:54 am By the way... may you take a short recording of audio output from your STe for comparison?

I don't hear anything from headphones.. but volume is very loud when press a key. So I think its a good test overall. I will try and record it.. not sure when I will get chance though.

I haven't tried any fix yet because I don't have the required hardware.

But here is a useful tool for sound analysis.
Audacity has a nice spectral analysis view, but it is hidden. See picture.

I have loaded the file ste-left.wav I posted there.
In the top left of the window, there is actually a menu if you click on the file name. In that menu, select Spectrogram. Then instead of the waveform, the window will display the instantaneous spectral analysis for every sample. And of course, you can still play the sound to hear and see exactly what happens.

Extremely useful tool, extremely easy to use.
Now, hardware experts, go ahead!

Maybe I figured out an interesting thing. Hiss is not the same when in color or mono mode. In color mode, the peaks are about 2600Hz and 15000Hz, in mono mode 2600Hz and 6700Hz.
I recorded the same situation in both cases. The scratches in the beginning are memory test of TOS 2.06 during boot, the second longer scratch is CosmosEx harddisk driver loading. Both waves are amplified by 36dB. It's also interesting that predominantly affected is left channel.

Have you run memory test software and checked the audio noise ? Can't remember if its been talked about before, though I think some noise is RAM access noise. So more RAM access, the worse the noise should get..

exxos wrote: ↑Tue Apr 24, 2018 7:52 pm Have you run memory test software and checked the audio noise ? Can't remember if its been talked about before, though I think some noise is RAM access noise. So more RAM access, the worse the noise should get..

No, I didn't run memory test software yet... but I have this feeling that the more RAM access, the worse noise is in the audio out. No matter if it is a direct stereo output or monitor audio (pin 1).
So, the question is why and if there is a possibility to defuse this noise and interferences.
And why only in left channel?