Which Amiga computer technically has the best audio quality?

[sokolovic]

Originally Posted by Turrican_3

Possibly not true?

The Factor5 logo of the Turrican3 Amiga intro apparently did, or at least this is what it claimed in the lower-right corner.

So is BC Kid from the same developpers.

[Turrican_3]

Chris Huelsbeck rules!

[AmigaHope]

Originally Posted by Turrican_3

Possibly not true?

The Factor5 logo of the Turrican3 Amiga intro apparently did, or at least this is what it claimed in the lower-right corner.

I mean the rear channel is just encoded as phase inversion, so any stereo sample can be Dolby surround.

[CCCP alert]

Just to clarify, I know Portia and Paula in all systems is a fixed spec, ignoring the by product of 56khz when productivity screen modes are used.

I am literally talking about the quality of motherboards and the relative effects in changing the reality of what is output on a top end hi-fi by the introduction of unwanted noise/sound altering effects diminishing the potential the chip offered. Many CD players use the same off the shelf DACs but there are noticeable differences in the quality and range of frequencies output from the device.

This has nothing to do with the tech specs of the DACs, all CD players had the same DAC specs but they didn't sound the same from model to model, they had slight differences in quality. Of course the health of the 3-4 decade old motherboard components will effect quality too in reality, like the dreaded crackly sound of duff CD32 mobo etc.

An R1 SID in a first production run silver label NTSC C64 has more bass and less treble output at the monitor port audio out than a final production run of the PAL silver label machines, identical tech specs so it's a motherboard thing not a sound chip thing. Audiophile magazine testing type difference.

[Turrican_3]

Originally Posted by AmigaHope

so any stereo sample can be Dolby surround.

Yes, I understand this is the way it works. But I didn't claim otherwise!

[h0ffman]

A1200
Great audio output, allows more top end and good volume.

A600
Due to an error in the audio circuit design it sounds really muffled. This can be fixed by removing C321 and C331 from the underside of the board then it matches the A1200. The only drawback is it reduces the overal output level and thus brings the noise floor up a little.

A500
Muffled output when compared to the A1200. I don't know of any fix for this.

I've run side by side comparisons of the 1200 / 600, before and after the fix listed above and it truly is night and day. Sadly I don't own a CD32, A2000, A3000, A4000 or A1000, but I would hazard a guess that the CD32 matches the A1200.

[no9]

Originally Posted by TuRRIcaNEd

Paula does not generate her own tones - she literally takes data from Chip RAM, is able to process that data through many functions that are useful in terms of audio and musical reproduction, and feeds the processed data to a DAC setup which converts the signals to an analogue audio signal pumped out of the stereo phono sockets. The only part of that system that might affect the final output is the character of the DACs - and the whole point of a DAC is that they're

Data is not sound. It is only a set of characters that might represent sound. So this in essence means that Paula generates its own tones internally and does so on its own, feeding itself from the memory field to which it is pointed. The way Paula interprets the data given to it determines the sound characteristics that are produced. The difference here is that Paula has access to the outside world to pull data from, while SID is limited to its internal 24-bit phase accumulator. But the sound is prepared and rendered inside, in both cases.

Originally Posted by TCD

It would be interesting in a 'blind test' to see if people can even hear the difference between real hardware and emulation.

It is easy to spot the real hardware since it pollutes the sound with noticeable noise and other interference. It might be harder if we take noise out of equation or add the same amount to the emulated ones.

If you manage to not spoil the test by reading the description you may check it by yourself [ Show youtube player ]

[ShK]

Originally Posted by CCCP alert

Paula in all systems is a fixed spec, ignoring the by product of 56khz when productivity screen modes are used.

And filters.

[TCD]

Originally Posted by no9

If you manage to not spoil the test by reading the description you may check it by yourself [ Show youtube player ]

Ah, nice one

[Photon]

Originally Posted by ShK

And filters.

Brilliant! Also, you can mod all Amigas' filters, e.g. to be the same as A1200. This thread has component values.

Probably an audio expert could improve a little on the circuits and provide a mod to lower the noise floor and provide a nicer filter.

In the synth and amp world, what you want is not necessarily and open filter. I think it's the same for samples and softsynths. You want a filter that's not in the way of what you want to do, and it's not often you want to do synthesis/instrument/modulation stuff very near the limit of Paula (close to 28kHz) because of the coarse steps in period - notes will be off key, vibrato and slides will be steppy. I guess you also want the filter to have a nice curve and not be a cutoff filter completely, for instruments where high pitch modulation is sampled in.

Where also the melody is sampled in (straight sampled loops or playing .wavs), especially distorted high pitch stuff like metal solos, cymbals/hihats, parts of speech etc, you might want the filter to be almost completely open for those parts to not be filtered out.

But more than that, you want a nice filter. You'd be surprised how much sound is filtered out by old synth and amp circuits, that the users say is "how it should sound".

I think the nice filter part IS the job of the amp, and the line level signals should be as pure as possible without raising the noise floor (e.g. by picking up any flaws in the circuit).

As it is though, it's good that the Amiga provides the service of filtering a bit, so you can connect studio monitors, or good amps+speakers that have very little filtering in them, without getting TOO raw/noisy output. It saves having to put something external and dial it in until noisy artifacts are gone per song/game/demo.

[CCCP alert]

FYI would never use the filter because it's shit

[ShK]

Yes, except lowpass filters since all Amiga models have a fixed lowpass filter in their audio output circuit that cannot be disabled. These are typically RC (resistor-capacitor) filters implemented in hardware.

[Photon]

Originally Posted by CCCP alert

FYI would never use the filter because it's shit

Yes, you mean the LED filter I think. Usually we turn that off, yes. It was only to allow/hide low sample rate artifacts in 1985 when Amigas could have max 512k chip.

[CCCP alert]

Originally Posted by Photon

Yes, you mean the LED filter I think. Usually we turn that off, yes. It was only to allow/hide low sample rate artifacts in 1985 when Amigas could have max 512k chip.

Yes the LED filter is what I meant, coming from the SID filters it was horrible.

I suppose I could get a 32 band spectrum analyser and test all my kit but who knows what condition the SMT components on the boards of my 600/1200/4000 is like.

[pandy71]

Originally Posted by no9

If you manage to not spoil the test by reading the description you may check it by yourself [ Show youtube player ]

Surprisingly seem original HW is better than emulation - yes it is noisy but still way more open... this is even visible in waveform - A500 has distinctively different look. Strange...
Btw this kind of test content can be misleading (due overall low sample rate and lot of antialiasing audible in samples - it should be probably some sampled sound - probably period 128..124).

[hammer]

Originally Posted by kriz

They all use Paula the same sound chip.. If you want higher quality, get a sound card

Analog circuits' quality also influences audio quality, not just the DAC.

[no9]

Originally Posted by pandy71

Btw this kind of test content can be misleading (due overall low sample rate and lot of antialiasing audible in samples - it should be probably some sampled sound - probably period 128..124).

This example represents the real-life experience of an Amiga user in a common situation, not an abstract edge case. The way Amiga, as a whole system, operates on sound defines the quality that users face. There is nothing misleading in this, and it is far more representative of the Amiga in common applications such as games or demos.

[ross]

Originally Posted by no9

This example represents the real-life experience of an Amiga user in a common situation, not an abstract edge case.

So true.
This is what I always challenge those who say that the Amiga has 'simple' 8-bit PCM audio.

The 'colour' of its sound is perfectly recognisable, and is due to a number of factors.
And the more you raise the frequency and use advanced modes and machines with different filters (A1200), the more its distinct sound gets 'lost'
This doesn't mean that the sound of the A1200 is bad (indeed, it is the machine with the 'best' audio output), but it is a bit different from the 'classic' Amiga sound.
The first time I heard it I was quite disappointed, it sounded too 'shrill' and full of perfectly audible aliasing, which degrades the quality at low frequencies.
Then obviously the greater memory and quality of the samples and frequencies used made the difference and now it is undoubtedly the best.

Yes, the Amiga sound, compared to any machine, is much 'fuller' and different from any existing PCM (because in fact it is not, at least in terms and for what is available and used now).
I didn't say in absolute better, but different, and I like it

[Gorf]

Originally Posted by ShK

And filters.

I would like to add, that the strange filter design of the A3000 (RC lowpass with built-in load) has just -4 dB/oct after the cutoff frequency (6kHz-12kHz) and -5 dB/oct on higher frequency (12kHz - 24kHz)

So all in all it sounds quite nice compared to an A500 or A2000

[ShK]

Thanks for the input! Here's the updated table: