Amiga 1000 - Too Expensive at Introduction or not Priced High Enough?

[Keatah]

21 minutes ago, zzip said:

My understanding of digital audio is you want the sampling rate to be twice the highest frequency you intend to sample.   Recently watched a video of someone who went through the math on this and it was pretty convincing.   They say humans can't hear above 20,000, so that's why the 44.1khz sampling rate of CDs persisted so long (good for frequencies of up to 22,050hz).

True enough humans don’t hear much above 20,000Hz. So a single tone would pass a test set up like that.

 

As soon as you get multiple instruments going, triangles, cymbals, drums, guitar, brass, and all, you’ve got high frequency harmonics that humans don’t hear. And even parts of the instruments themselves can’t be heard either.

 

But guess what none of that math talks about? Guess what happens when those instrument play together?

 

The harmonics that aren’t heard individually now downmix into tones you can hear. Think 2 motors running at different speeds. There’s a lower frequency beat that forms. And anyone can hear that.

 

Now imagine all that happens with an orchestra or band. People hear that and describe it as definition and clarity.

 

Done nicely it’s an award-winning piece. Overdone and its cluttered and fatiguing.

 

Got a DVD audio recording of Alan Parsons’ “Try Anything Once”, with sampling rates much higher than standard CD. The sharpness and color is markedly noticeable. Unmuddied highs, dramatic separation. Dynamic range and SNR goes to the next level. Makes CD audio quite tinny and flat in comparison. 

 

21 minutes ago, zzip said:

 Given that,  even 96khz seems excessive.    It would be interesting to see if there's blind tests where people can tell the difference between 44.1, 96 and 192

On a single tone. No. On an orchestra recording? Very likely yes.

 

21 minutes ago, zzip said:

Back when I had my Atari 800XL,  I was trying to shape wave forms to make it sound more like real instruments-  like piano, trumpet, guitar etc..   But it turned out I could only take that so far.   I guess for that reason FM synthesis was never good enough for me.   I wanted the computer to sound like it was playing real instruments, and FM was usually a poor representation of that.   That's why I quickly jumped to a wavetable card..  OPL3 always sounded kind of weird to me.

Early on in A2 days, with the Mountain Music System or ALF Music Card, I was hard up to get real instrument sounds.

 

Somewhere along the way I decided a computer should sound like a computer and that was the end of it.

 

So early synthesis and FM was good enough. Many arcade cabs were like that. Many sounded like modified electric guitars - and I liked that.

[Keatah]

11 hours ago, zzip said:

3D Now! was simply AMD's marketing name for their version of MMX.   I think it was compatible with MMX with some extra added instructions.

I believe the K6-2 had MMX, 57 new instructions IIRC. And 3DNow! was an extension to that, 21 new instructions on top of the 57. The 21 were similar to SSE in how they used and sized the registers.

 

11 hours ago, zzip said:

There was an interesting Linux distribution called Gentoo that would recompile everything on your system with compiler optimization flags for your specific CPU rather than a lowest common denominator CPU.

Now that is interesting.

[Daedalus2097]

14 hours ago, Geoff Oltmans said:

To be fair, Wing Commander on an Amiga isn't a particularly great experience either. heh

True, though it got better with a little extra CPU grunt. Too much horsepower though and it ran too fast and became unplayable - much like many early PC games  To be fair, the CD32 (AGA) version is an improvement too, getting closer to the VGA version.

 

12 hours ago, zzip said:

My understanding of digital audio is you want the sampling rate to be twice the highest frequency you intend to sample.   Recently watched a video of someone who went through the math on this and it was pretty convincing.   They say humans can't hear above 20,000, so that's why the 44.1khz sampling rate of CDs persisted so long (good for frequencies of up to 22,050hz).   Given that,  even 96khz seems excessive.    It would be interesting to see if there's blind tests where people can tell the difference between 44.1, 96 and 192

On top of what has already been said, you get quantization errors that increase as you approach they Nyquist frequency (which is 22,050Hz in your example). While yes, in theory you can reproduce that frequency with a 44.1kHz sample rate, what it will look like is a 22,050Hz square wave, regardless of the shape of the wave you put in. And what happens if your frequency is, say, 18kHz? You get a horrible noise in your waveform because now you have a square wave with a variable duty cycle (sometimes 1 sample high, sometimes 2 samples high) and variable amplitude, because the peaks don't coincide with the sampling points and so you can't get an accurate representation of that waveform. 44.1kHz is certainly good enough for most purposes, but it's when you get into certain edge cases that you start to see (well, hear) the limitations.

 

Of course all of that pales into insignificance once you start involving lossy compression, which has a far greater impact on the sound.

Edited December 7, 2022 by Daedalus2097

[zzip]

15 hours ago, Keatah said:

As soon as you get multiple instruments going, triangles, cymbals, drums, guitar, brass, and all, you’ve got high frequency harmonics that humans don’t hear. And even parts of the instruments themselves can’t be heard either.

 

But guess what none of that math talks about? Guess what happens when those instrument play together?

If you are recording the instruments individually and mixing them,  then yeah maybe having a 96khz source will make a difference in the final product.    But if you have an audio track that is already mixed and mastered--  the the waveforms are already fixed.  Anything outside the range of human hearing has been removed via high-pass filter and I really doubt you are going to hear a difference between 1 44.1khz FLAC and a 96khz FLAC

 

15 hours ago, Keatah said:

Done nicely it’s an award-winning piece. Overdone and its cluttered and fatiguing.

 

Got a DVD audio recording of Alan Parsons’ “Try Anything Once”, with sampling rates much higher than standard CD. The sharpness and color is markedly noticeable. Unmuddied highs, dramatic separation. Dynamic range and SNR goes to the next level. Makes CD audio quite tinny and flat in comparison. 

There's been an issue in music, especially rock music--  the so called "loudness wars" where record labels would want their releases mixed a little bit louder than other releases so that it stands out on radio.  Over time the increased loudness ruined the fidelity of the recordings.   Back in the 70s, it was common to have instrument meticulously recorded.  you could hear every instrument clearly even on standard CD.    More modern recordings sound muddier and distorted, and not always intentionally 

 

Alan Parsons is an audio engineer.   I'm sure he could make it sound great even on CD.

 

 

[Keatah]

5 hours ago, zzip said:

If you are recording the instruments individually and mixing them,  then yeah maybe having a 96khz source will make a difference in the final product.    But if you have an audio track that is already mixed and mastered--  the the waveforms are already fixed.  Anything outside the range of human hearing has been removed via high-pass filter and I really doubt you are going to hear a difference between 1 44.1khz FLAC and a 96khz FLAC

Having that removed tends to cause a certain flatness and less clarity. You're not pulled into the depths where the musical instruments interact and play off each other.

 

5 hours ago, zzip said:

There's been an issue in music, especially rock music--  the so called "loudness wars" where record labels would want their releases mixed a little bit louder than other releases so that it stands out on radio.  Over time the increased loudness ruined the fidelity of the recordings.   Back in the 70s, it was common to have instrument meticulously recorded.  you could hear every instrument clearly even on standard CD.    More modern recordings sound muddier and distorted, and not always intentionally 

It's one of the reasons I prefer earlier music. The separation and individuality of instruments were there if you wanted to pay attention. And the performers didn't generally call audience members up on stage for golden shows either. And much less twerking.

 

5 hours ago, zzip said:

Alan Parsons is an audio engineer.   I'm sure he could make it sound great even on CD.

And it does. At first got the gold version of the album and was impressed. I had already expected a lot from AP. But I was still green and anything round and shiny would automatically sound better by default.

 

I do have an early CD of Survivor and it sounds like crap. The highs are totally blown out and sounds like it's playing a on a crystal radio earphone. It's one of about 15 out of 1600 or so discs I consider unenjoyable for technical reasons. Cymbals sound like TV static shifted up several octaves.

Edited December 7, 2022 by Keatah