Why is Google Drive so slow for downloading and uploading?
Google Drive is a cloud folder synchronization program. It is easy and efficient when you only have one or two computers to sync.For companies with many computers/employees, it is inefficient and insecure, for example: If 100 computers sync to one shared folder, and if a user updated a 50MB file, it will create 5GB of download usage, which will slow down your network dramatically.Cloud folder synchronization also wastes local storage space. In the above example, it will use 100 times of local storage space. If your company has just 1TB of data, the amount of storage space wasted is 99TB! Because business data is sync-ed to many computers/devices, it cannot be centrally managed. Data breach can happen on any device and there is no way to prevent it.A cloud file server/drive mapping based solution is much more efficient and secure. (DriveHQ is a leading Enterprise Cloud File Server/WebDAV Drive Mapping service provider.) It works just like a local file server/local drive, there is no learning curve. It does not sync data locally, which saves local storage space, bandwidth and allows the system to be centrally managed. System administrators can set fine granular access control on different folders.
If a human ear can hear no frequency greater than 20 kHz, and we need no more than 40 kHz sampling rates, why does equipment exist that plays and records at 96 or 128 kHz sample rates?
You're right, the human ear cannot hear past 20kHz. In fact, as we age, that number drops off, easily to 15kHz. But for the sake of our discussion, let's assume that we would like to reproduce a 20kHz tone.You are also correct in following through the Nyquist sampling frequency of just over 40kHz is the minimum we need to properly represent a 20kHz signal—otherwise we will experience aliasing, in which our samples could accurately represent a plurality of possible signals, as shown here:Undersampling(from the aliasing article on Wikipedia, http://en.wikipedia.org/wiki/Ali...)Why is 40kHz not enough? I made two graphs especially for you! Here is a 20kHz signal, with a slightly-more-than-40kHz sampling frequency:"Sufficient" samplingAlthough we know the frequency of the signal from our sampling, the reconstruction will not get us the amplitude without many cycles, as a result of a few factors[1]. A typical song isn't just a single tone, either... we may not have so many cycles to go by. A brief, loud, high-frequency sound could be be almost completely missed here.One benefit of oversampling—measuring at a frequency much higher than the Nyquist-Shannon-theorem-suggested minimum—is that we can determine the nature of the signal more quickly. Here is 128kHz sampling on our 20kHz signal:OversamplingSo why are so many sampling rates in the 44.1kHz range? First, as I already mentioned, the adult ear doesn't hear 20kHz frequencies anyway, so this is already oversampling. Second, there is always going to be a tradeoff between data cost and sampling resolution. How frequently you sample a signal depends on error size you are willing to tolerate, filtering in reproducing the signal later, the costs of data storage, etc. There are variable sampling rate schemes which can describe a signal with the same accuracy at a lower cost, but at the end of the day, you can't sample with an infinite frequency. Don't say that near audiophiles though, because they have feelings too.[1] The filtering of greater-than-20kHz components in the reconstruction (Ian has since written a good answer about this), discretization by the analog-to-digital conversion of the samples, and the fact that the song audio will contain higher frequencies than we care to record.