Excuse me for having an acute case of technical nostalgia, but let me rant about what kind of performance you could get 20 years ago for very affordable money (400-500€/$). TLDR: it's so good, it still holds up today - or would, if hardware and software interfaces hadn't made it obsolete.
M-Audio Delta 1010 Digital Recording System -intended for and marketed towards the "semi professional" sector, that is home and small professional music studios
PCI card with S/PDIF in/out. Breakout box with external power supply (9V AC), BNC wordclock in/out, MIDI in/out, and 8ch in/out via balanced TRS, each switchable between -10/+4 levels. In total, 10ch in/out, hence the model name.
PCI card and breakout box (steel with aluminium front plate) connected by standard DB25 cable. Nice: widely available, nothing proprietary.
A look inside the breakout box with converters and connectors:
What we see here is eight input channels buffered by 2x NE5532 each (presumably for desymetrisation and level switching), and 4x 2ch AKM AK5383 ADC chips.
Removing the ADC board, we have this:
Basically the same story: eight output channels, again using 16x NE5532 buffers, and this time 4x 2ch AKM AK4393 DAC chips.
The whole system is 24/96, with a 36bit DSP for routing and mixing. Those AK4393 were called "wonder DAC" at the time, because their performance was exceptionally good, without being prohibitively expensive. It were exactly those days when 24/96 finally became mainstream, and affordable in really good quality.
PCI card:
These DAC chips had a manufacturer rated SNR of 120bB, the ADCs only marginally less. The whole M-Audio implementation was rated 109 (ADC) and 108 (DAC) dB by the manufacturer, at a distortion of <0.0015%. Personally I only measured noise, round-trip (output to input, thus both combined) revealed 105dB, and frequency response was super flat (+-0.1dB) from 20 to some 30 or 35 kHz. It's long ago and I forgot the exact numbers, but back then it was stellar. It easily doubled as a makeshift amateur measurement device for older gear, as long as you regarded its own limitations - which frankly were much better than most musical gear you could buy back then. Mixers, synthesizers, fx boxes, etc. - infact you'd have been hard pressed to find any professional electronic instrument or other gear with a lower noise floor than this.
This system was utterly and thoroughly audibly transparent, and would comfortably land in the "good" category of ASR measurements today, 25 years later. All that for 400 euromoneys from 2003 or so. Plus super stable system performance too: even a moldy Intel Core2Duo PC could do 128 samples buffer size with this interface, stable with no audio dropouts until full CPU load. That's 2.6ms latency at 48kHz. Later quad core CPUs allowed 64 samples or 1.3ms latency. Half that at 96kHz. Stable. For all practical musical purposes, may it be recording or live performance or any combination, that's true realtime. Even the MIDI standard is much slower than that. Add flexible routing and zero-latency 36-bit mixing and monitoring via onboard DSP, comfortably handled by software driver control panel, and you got a winner in terms of practicality too. Those 5.1 and 7.1 multichannel surround output modes supported by the driver seem like free gimmicks at this point.
In that light, I wonder where all the technical progress went since then. 8ch AD and DA i/o for that price at that quality, plus extra features, 15+ years of driver support (this 1999 model got drivers up to and including Windows 7, plus updates), great audio and mechanical quality that even surpasses the "semi professional" target market, for this little money? Even after inflation, where do you get that today?
M-Audio Delta 1010 Digital Recording System -intended for and marketed towards the "semi professional" sector, that is home and small professional music studios
PCI card with S/PDIF in/out. Breakout box with external power supply (9V AC), BNC wordclock in/out, MIDI in/out, and 8ch in/out via balanced TRS, each switchable between -10/+4 levels. In total, 10ch in/out, hence the model name.
PCI card and breakout box (steel with aluminium front plate) connected by standard DB25 cable. Nice: widely available, nothing proprietary.
A look inside the breakout box with converters and connectors:
What we see here is eight input channels buffered by 2x NE5532 each (presumably for desymetrisation and level switching), and 4x 2ch AKM AK5383 ADC chips.
Removing the ADC board, we have this:
Basically the same story: eight output channels, again using 16x NE5532 buffers, and this time 4x 2ch AKM AK4393 DAC chips.
The whole system is 24/96, with a 36bit DSP for routing and mixing. Those AK4393 were called "wonder DAC" at the time, because their performance was exceptionally good, without being prohibitively expensive. It were exactly those days when 24/96 finally became mainstream, and affordable in really good quality.
PCI card:
These DAC chips had a manufacturer rated SNR of 120bB, the ADCs only marginally less. The whole M-Audio implementation was rated 109 (ADC) and 108 (DAC) dB by the manufacturer, at a distortion of <0.0015%. Personally I only measured noise, round-trip (output to input, thus both combined) revealed 105dB, and frequency response was super flat (+-0.1dB) from 20 to some 30 or 35 kHz. It's long ago and I forgot the exact numbers, but back then it was stellar. It easily doubled as a makeshift amateur measurement device for older gear, as long as you regarded its own limitations - which frankly were much better than most musical gear you could buy back then. Mixers, synthesizers, fx boxes, etc. - infact you'd have been hard pressed to find any professional electronic instrument or other gear with a lower noise floor than this.
This system was utterly and thoroughly audibly transparent, and would comfortably land in the "good" category of ASR measurements today, 25 years later. All that for 400 euromoneys from 2003 or so. Plus super stable system performance too: even a moldy Intel Core2Duo PC could do 128 samples buffer size with this interface, stable with no audio dropouts until full CPU load. That's 2.6ms latency at 48kHz. Later quad core CPUs allowed 64 samples or 1.3ms latency. Half that at 96kHz. Stable. For all practical musical purposes, may it be recording or live performance or any combination, that's true realtime. Even the MIDI standard is much slower than that. Add flexible routing and zero-latency 36-bit mixing and monitoring via onboard DSP, comfortably handled by software driver control panel, and you got a winner in terms of practicality too. Those 5.1 and 7.1 multichannel surround output modes supported by the driver seem like free gimmicks at this point.
In that light, I wonder where all the technical progress went since then. 8ch AD and DA i/o for that price at that quality, plus extra features, 15+ years of driver support (this 1999 model got drivers up to and including Windows 7, plus updates), great audio and mechanical quality that even surpasses the "semi professional" target market, for this little money? Even after inflation, where do you get that today?
Last edited: