4 ms·

I bailed out halfway into my CS/EE university program (to start working in software instead) a very long time ago in part so I wouldn't have to finish the signa

I bailed out halfway into my CS/EE university program (to start working in software instead) a very long time ago in part so I wouldn't have to finish the signal theory/processing course, so please bear with me. "I made the assumption that by harsh you meant dissonant sounds of a relatively high audio frequency" Yes, that's what I'm hearing. Here is a comparison video where you can listen to the difference between the original NES (via an ADC, presumably) and the NES Classic Mini: https://www.youtube.com/watch?v=Z9EyI3mVhi4 https://www.youtube.com/watch?v=Z9EyI3mVhi4 Anyway, to me it sounds that by just filtering away some high frequencies you would get a better approximation of the original. Edit: Oh wait, I see what I did; I mixed up high-pass filter with low-pass filter (again). Doh. I do maintain that whoever named them made a mistake.

morganvachon 10y ago
A sibilance filter would be fairly easy to make, I built a headphone amp board with a sibilance filter circuit in it many years back.
- johansch 10y ago
  Is that something that can be cheaply done in software (with low latency - like less than 50 ms)?
  - nitrogen 10y ago
    An IIR low-pass filter can be implemented with a delay measured in tens of microseconds. With regard to naming of high-pass/low-pass filters, it might help to think of the opposite names (low-cut/high-cut). Pass is the opposite of cut/reject.
  - morganvachon 10y ago
    I have no idea as I'm not a programmer, I'm pretty much a hardware/EE guy. When I built mine it was less than 50 cents in discrete components and can definitely be done with SMT components to save space. I don't see why not though; I can get some pretty precise filtering using LADSPA plugins in software mixing apps.
metaphor 10y ago
Ahh, makes sense now. The emulator sounds like it does an outstanding job of functionally reproducing the APU, but clearly doesn't capture the analog LPF contributions of the original architecture[1] (2nd and 4th pages apply). An improvement to the NES Classic will necessarily have to come in software as a DSP LPF either within the emulator application or at the OS level (someone mentioned seeing alsa during boot??). Nevertheless, a passive analog hack is definitely not going to be applicable. [1] http://www.freeinfosociety.com/electronics/schemview.php?id=2405 http://www.freeinfosociety.com/electronics/schemview.php?id=...
- johansch 10y ago
  Oh come on, do tell us what NES emulator you made? :)
  - metaphor 10y ago
    What makes you think that I've developed my own emulator? Not that the assumption was incorrect...
    - johansch 10y ago
      The incredibly intricate understanding of the NES sound?
- johansch 10y ago
  "alsa-lib-1.0.29.tar.bz2" is part of the OSS compliance download package they released. (See https://news.ycombinator.com/item?id=12938817 https://news.ycombinator.com/item?id=12938817)
kaoD 10y ago
Nice link! Not only the high end is too crisp, the low end sounds "fatter" too.
- johansch 10y ago
  That guy's Youtube channel appears to be full of wonderful things, I just noticed: https://www.youtube.com/user/KM65536/videos https://www.youtube.com/user/KM65536/videos
jungletek 10y ago
>I do maintain that whoever named them made a mistake. Nah, they're named fine, you're just doing the incorrect mental-shortcut that I used to do: because there's 'low' and 'filter', you're ignoring 'pass'. The lows or the highs (respectively) pass-through the filter.

morganvachon 10y ago

A sibilance filter would be fairly easy to make, I built a headphone amp board with a sibilance filter circuit in it many years back.

johansch 10y ago
Is that something that can be cheaply done in software (with low latency - like less than 50 ms)?
- nitrogen 10y ago
  An IIR low-pass filter can be implemented with a delay measured in tens of microseconds. With regard to naming of high-pass/low-pass filters, it might help to think of the opposite names (low-cut/high-cut). Pass is the opposite of cut/reject.
- morganvachon 10y ago
  I have no idea as I'm not a programmer, I'm pretty much a hardware/EE guy. When I built mine it was less than 50 cents in discrete components and can definitely be done with SMT components to save space. I don't see why not though; I can get some pretty precise filtering using LADSPA plugins in software mixing apps.

metaphor 10y ago

Ahh, makes sense now. The emulator sounds like it does an outstanding job of functionally reproducing the APU, but clearly doesn't capture the analog LPF contributions of the original architecture[1] (2nd and 4th pages apply). An improvement to the NES Classic will necessarily have to come in software as a DSP LPF either within the emulator application or at the OS level (someone mentioned seeing alsa during boot??). Nevertheless, a passive analog hack is definitely not going to be applicable. [1] http://www.freeinfosociety.com/electronics/schemview.php?id=2405 http://www.freeinfosociety.com/electronics/schemview.php?id=...

johansch 10y ago
Oh come on, do tell us what NES emulator you made? :)
- metaphor 10y ago
  What makes you think that I've developed my own emulator? Not that the assumption was incorrect...
  - johansch 10y ago
    The incredibly intricate understanding of the NES sound?
johansch 10y ago
"alsa-lib-1.0.29.tar.bz2" is part of the OSS compliance download package they released. (See https://news.ycombinator.com/item?id=12938817 https://news.ycombinator.com/item?id=12938817)

kaoD 10y ago

Nice link! Not only the high end is too crisp, the low end sounds "fatter" too.

johansch 10y ago
That guy's Youtube channel appears to be full of wonderful things, I just noticed: https://www.youtube.com/user/KM65536/videos https://www.youtube.com/user/KM65536/videos

jungletek 10y ago

>I do maintain that whoever named them made a mistake. Nah, they're named fine, you're just doing the incorrect mental-shortcut that I used to do: because there's 'low' and 'filter', you're ignoring 'pass'. The lows or the highs (respectively) pass-through the filter.