Yeah, the piss poor visual representations used to try to explain digital have done it no favors, but the other side of the issue is that digital devices are just a lot less expensive to produce than analog ones.....only thing better than cheap is cheaper. There are a lot of garbage digital devices out there that make digital a bad word.
To me it's all about the device itself. I mean you could take a completely crappy, cheap analog device and say "well look, analog sucks" but if you take a high end analog device it's going to sound fantastic. It translates to digital as well.
Junk is junk whether its analog or digital.
The other thing is that people automatically equate bigger numbers with higher quality. There is nothing in the audible spectrum that cannot be fully represented by 44.1kHz. Hell, 32Hz will generally take care of anything you really want to listen to (I mean how many people have heard 16kHz.....it's not pleasant). The problem as I understand it is that when you are taking an analog signal and digitize it that you absolutely have to make sure that the highest frequency feeding the ADC is less than half the sample rate. Every electronic circuit is susceptible to interference; RF, intermodulation, harmonic, etc., those are the things that can cause issues with your digitized product. Those have to be filtered. Filtering is tricky and from what I have heard it's one of the more expensive parts of the process. A good sharp filter that has little roll off is going to cost more to produce than something that has a gradual roll off. That roll off ends up attenuating the high end and makes it sound like mud. Oversampling is a trick that is used to make cheaper filters sharper and I do not know enough to say that higher sample rates benefit from the same thing, but I do have my suspicions that this is the case. I know that having a higher sample rate will allow the filter to cutoff later which isn't about quality as much as cost savings.
To me the bit depth is where the quality of the signal is made these days. With 16 bit you can only get to -96dB before it's just noise which seems like a lot of headroom, but I can hear a the benefit in 24-bit, especially when you are doing things like re-amping where the guitar peaks at -40dB. You get higher resolution and a much higher SNR and it's a lot more forgiving when you record something too low and add gain. I've recorded 16 bit at very low levels and when you add gain it does get noisy as hell. 24-bit and higher takes care of that. At least in my experience. And yes dithering can help, but to me that alters the original signal even if it's only slightly. 24 bit and I don't have to worry about it.
There's just no way to counter the number of mass produced, inferior products out there that are designed for profit margin over quality, but you can't deny that there is a market for it. I just wish that people would realize that if you buy something cheap it's going to be cheap and it's not the best representation of analog or digital.
