Amiga "14-bit audio" refuted

[Thomas Richter]

Originally Posted by roondar

Do you have an idea why your figures are so much lower?

That depends on how you measured, and with which equipment. The results are relative to the measurement chain. 16bit output should give you approximately 96dB, probably 90dB with good equipment, but I can only measure 60dB here with my setup.

Still, there is quite some difference between 31dB and 60dB.

[Thomas Richter]

Originally Posted by nikosidis

Maybe it is Thomas A2000 that is that bad. I guess my A500 rev. 3 board is that bad too.
A1200 is nothing near that bad SNR. At least around 60db I would say is correct for A1200, EaglePlayer, 14bit amp.

Instead of making claims, please measure. There have been enough false claims already here. As in "The Amiga never creates dropouts, never!". Also simply wrong on my setup.

[Thomas Richter]

Originally Posted by Gorf

Therefore the whole test setting is biased:
To really make a useful comparison, you need an analog audio wave a staring point.
Each machine must be allowed to sample and playback that tone at a sample frequency that best matches it’s inner workings.

Please don't me call stupid. Please look at the test programs I provided. It does adjust to "the starting point" (actually, that is called "the phase&quot, and it does adjust to the frequency (actually, by including a phase drift). It also adjusts to the amplitude. Continuously following the phase is identical to adjusting the frequency.

The way how this works is that it fits a sine wave in double precision to the input, by continuously fitting the amplitude and the phase. Of course, if you can measure better, and provide improvements to the program, you are welcome.

Requesting a 16bit 44.1kHz output is actually nothing "unusual" if someone claims "CD quality audio" on the Amiga. CD quality is exactly that. (And not some other frequency). But, as always, the program allows to adjust the frequency, so please perform your own measurements and provide results.

Critique is good if it is based on facts, it is bad if it is based on fiction, aka "fan-boy-ism".

[roondar]

Originally Posted by Thomas Richter

Still, there is quite some difference between 31dB and 60dB.

Oh, don't get me wrong - this isn't about claiming that your measurements don't show anything relevant. In fact, even the original paper by Henryk already showed an about 10-11 bit effective resolution, as did buggs's results.

It was merely me wondering how come your figures were so different.

[nikosidis]

Originally Posted by Thomas Richter

Instead of making claims, please measure. There have been enough false claims already here. As in "The Amiga never creates dropouts, never!". Also simply wrong on my setup.

I'm sorry to say but how do I do that? Seams complicated.

I see the 2 progs. in your archive. I have Linux64 and Win.10 but no dedicated soundcard. I for sure have mic input.

Thomas: Did you add A-weighting filter?

It not to much trouble for you I would like to hear the last track I recorded in the thread from your A2000.
There are many quiet passages there. We could even compare the sine waves.

[nikosidis]

This was an interesting read!

https://www.audiosciencereview.com/f...t-is-quiet.14/

[Thorham]

Originally Posted by roondar

In fact, even the original paper by Henryk already showed an about 10-11 bit effective resolution, as did buggs's results.

I got the same result, by ear no less.

[malko]

Originally Posted by grond

There is nothing funny about tinnitus. I also have it and it got quite a bit worse with my covid infection.

I know. The opposite was not the meaning of my post as I have reduced a lot mine (almost vanished) with the grandmother's hint. Maybe I should have separated my answer like this :

Originally Posted by Bruce Abbott

[...] The older I get, the better Paula sounds!

Originally Posted by Bruce Abbott

[...] I have constant tinnitus which masks higher frequencies. [...]

A grandmother's hint : Rub a cotton bud with sweet almond oil and "brush" the inside of your ears with it. Practice daily for a few months and the tinnitus will fade away .

[buggs]

Originally Posted by Bruce Abbott

Another factor that 'scientific' tests often don't take into account is the response of the human ear, including psychoacoustic effects. For a fair comparison the signal should at least be 'A' weighted, so that components outside the normal hearing range are suppressed. Signals should also be evaluated at different levels corresponding to the dynamic range of music, rather than just at maximum volume where audio masking tends to suppress low level noise and distortion.

Yes, the dB(A) curve or the later ITU-R 468 are quite common for psychoacoustic weighting of the SPL. In this case, it probably wouldn't do poor old Paula any good, I guess. There's little relevant energy around DC or near the end of the normal hearing range.
Your suggestion towards different levels (or amplitude) ranges as test source sure makes sense in cases of DACs with known nonlinearities. As it happens, I had such a test on my TODO already.
Another part to look at is the OpAmp. Typically, the TL084 was mounted in earlier Amiga models, replaced later by the LF347 with more reserves towards bandwidth at higher amplification. I wouldn't rule out the OpAmp as a source for overshoots, resulting in undesired harmonics.

[buggs]

Originally Posted by Thomas Richter

That depends on how you measured, and with which equipment. The results are relative to the measurement chain. 16bit output should give you approximately 96dB, probably 90dB with good equipment, but I can only measure 60dB here with my setup.
Still, there is quite some difference between 31dB and 60dB.

Can you please stop toting these numbers around? You should have realized by now that they are utterly wrong. No, they are not a little off, and they are not caused by your setup being inappropriate. Modern hardware from reputable brands does deliver north of 90dB dynamic range, reliably. They don't top out at 10 bit and neither is the Amiga limited to 5 bit and change.

Your whole calculation is wrong, plain and simple. If your algorithm doesn't come nowhere near plausible and/or expected results, than it is defective. An unexplained difference of more than 30dB between your results vs. the expected numbers amounts to a disqualification.

[Thomas Richter]

Originally Posted by buggs

Can you please stop toting these numbers around? You should have realized by now that they are utterly wrong.

You are invited to compute better, really. I provided the algorithm, and this is what I get with the equipment I have. Note that there is additional noise in the loop (quantized twice), plus some phase uncertainty. If you have better equipment, and/or better algorithms, please provide them.

[Thomas Richter]

Originally Posted by buggs

Modern hardware from reputable brands does deliver north of 90dB dynamic range, reliably.

With a 16 bit quantizer? How is that supposed to happen? With ~6dB per Bit (you should know that), you are limited to 96dB. With 14 bit, you should be able to get around 84dB, only 12dB less (obviously, for 2 bits missing). We are off by a lot from that.

If you need higher quality, you need oversampling, interpolation and even more bits. That's not what I recorded with. This is just 16 bit out, 16bit in, with more noise in the loop.

[Thomas Richter]

It's also interesting to look (or hear) the error signal, which is easily arranged with the algorithm provided. For the PC output, the error signal is so low I cannot hear anything, and audacity's spectral analyzer doesn't show anything. For the Amiga output, I have multiple peaks in the spectrum, most notably one at 56Hz probably coming from the video signal at around 70dB, and its higher harmonics, and another series of peaks at 880Hz and 1670Hz (double and four times the frequency) due to quantization at ~80dB and ~50dB. Thus, not really much of surprise here. Outside error signals due to improper isolation of the audio circuit, and higher harmonics of the signal itself due to quantization defects.

[roondar]

Originally Posted by Thomas Richter

With a 16 bit quantizer? How is that supposed to happen? With ~6dB per Bit (you should know that), you are limited to 96dB. With 14 bit, you should be able to get around 84dB, only 12dB less (obviously, for 2 bits missing). We are off by a lot from that.

If you need higher quality, you need oversampling, interpolation and even more bits. That's not what I recorded with. This is just 16 bit out, 16bit in, with more noise in the loop.

Most modern audio hardware does 24 bits, though. So it surely should be doable to reach over 96dB?

[nikosidis]

The software used here is questionable and using SoundBlaster Live is too. The analogue input probably generate a lot of noise.
A high-end soundcard or other pro. measuring equipnent is need to get more realistic data out of this.
On top of that bring inn a Amiga 1200 to get the best out of Paula.

The best proof is still always in the listening.

[Bruce Abbott]

I created a sequence of 1kHz tones at different levels ranging from 0dB to -90dB using Audacity on my PC, then played them on my A1200 in Eagle player with its 'calibrated 14 bit' engine. Going down to -65db, -67.2dB, -70dB and -75dB the tone got progressively weaker, but still sounded pure. At -76.5dB it suddenly dropped out. All levels below that produced nothing.

The 16 bit values in the -75dB file ranged from -6 to +6 = 13 levels, which is slightly less than 4 bits. Dividing 75dB by 6.06dB per bit we get an effective 12.4 bits.

That's not 14 bits, but is it a big deal? -75dBm is a pretty low level. With my amp set to normal listening volume the -75dB tone was inaudible. I had to crank the volume knob all the way up to '100' to hear it, which of course brought up some mains hum (probably caused by a ground loop between my amp and other components) and faint rumbling sounds whenever the Amiga's mouse moved or hard drive was accessed.

To get a better idea of what noise might be present I hooked my oscilloscope up to one of the A1200's audio outputs. There was ~5mV of high frequency pulses which appeared be coming from Paula and not completely filtered out, and the trace bounced around at a low frequency due to that ground loop, but at -60dB the sine wave looked pure. At -70dB it became distorted, indicating insufficient bits to form a good waveform. At this level it was very hard to get a stable display due to all the high frequency noise. So the scope showed a lot of inaudible stuff that made the signal look much worse than it sounded.

[Gorf]

Originally Posted by Thomas Richter

Please don't me call stupid. Please look at the test programs I provided. It does adjust to "the starting point" (actually, that is called "the phase&quot, and it does adjust to the frequency (actually, by including a phase drift). It also adjusts to the amplitude. Continuously following the phase is identical to adjusting the frequency.

The way how this works is that it fits a sine wave in double precision to the input, by continuously fitting the amplitude and the phase. Of course, if you can measure better, and provide improvements to the program, you are welcome.

At no point I called you stupid.
So what is this rambling about?

Your adjustments do not provide a fair comparison, due to all the differences these systems have....

Requesting a 16bit 44.1kHz output is actually nothing "unusual" if someone claims "CD quality audio" on the Amiga. CD quality is exactly that. (And not some other frequency). But, as always, the program allows to adjust the frequency, so please perform your own measurements and provide results.

This exact frequency is of course NOT useful for a comparison - as is ANY fixed frequency!
As said above: the starting point needs to be a analog audio wave - after all that is what both systems try to recreate: a good resemblance of a actual analog acoustic.

The closest thing in the digital realm is probably a Pulse Density Modulation like on a Super Audio CD at several MHz.

Starting from there might give us a fair method.

Such a sample now could be transformed into a 16bit 44.1kHz PCM sample on a PC und a 55.4 kHz 8bit PCM + 6bit PWM combination on the Amiga ...

[Thomas Richter]

Originally Posted by Gorf

This exact frequency is of course NOT useful for a comparison - as is ANY fixed frequency!

Unfortunately, that is exactly what we have: Digitally sampled analog audio waves. I tried to provide tools to generate such waves with any frequency and any sampling frequency you like.

Originally Posted by Gorf

The closest thing in the digital realm is probably a Pulse Density Modulation like on a Super Audio CD at several MHz.

Unfortunately, neither the Amiga nor the PC can provide that. I'm not clear why you are suggesting something like this - it is out of reach.

[Thomas Richter]

Originally Posted by roondar

Most modern audio hardware does 24 bits, though. So it surely should be doable to reach over 96dB?

Absolutely, you should. Though it also requires a high quality analog part. The SB Live! here doesn't do that. It's not what is available as raw data on the CD as well, though every decent CD player would try to improve the quality by oversampling.

[Gorf]

Originally Posted by Thomas Richter

Unfortunately, that is exactly what we have: Digitally sampled analog audio waves. I tried to provide tools to generate such waves with any frequency and any sampling frequency you like.

Unfortunately, neither the Amiga nor the PC can provide that. I'm not clear why you are suggesting something like this - it is out of reach.

Why?
You can download PDM encoded music - so called DSD formal. E.g. here:
https://bluecoastmusic.com/free-downloads