 Begin December, 1176 Hambone Road, pulses are not real world, sine waves are and he is right. It was a reaction on my video The Truth About Nyquist and why 192 kHz does make sense. The implicit message he gave was wrong though. Digital Audio does not output pulses but sine waves, just as analog equipment. Yet I understand this remark. When Digital Audio was introduced, marketeers didn't have the foggiest idea what it was and asked techies for input. They came with a schematic representation of sampling that shows bars that touched a waveform at one of the top corners. This led to a wide belief that Digital Audio worked with bars for pulses and in an abstract way it does, but not in a real analog life. Let me use a metaphor to explain, the carpenter. For digitising is already ages old, as any carpenter can show you. If you ask a carpenter to copy you a shelf, the first thing he will do is get out his folding ruler and start measuring. He could have taken a large piece of cardboard to copy the shape of the shelf, that would have been very precise but also extremely inconvenient. Wind and rain might damage the template and it's hard to transport anyway. Instead the carpenter takes his rule and notes the measurements in his notebook or if it's a junior in his smartphone. It might be clear that the notebook is far easier to take to the workshop and the junior even has the measurements stored in his computer at the workshop even before he gets there. Now let's use some expensive words, as we say in my country. The template shape would have been analogous to the shape of the shelf while the measurements are a numerical representation of the shelf. The scientific name for numerical is digital. It comes from the Latin word for the first calculator of millennia ago, the fingers. The Latin word for finger is digitus. This information is utterly useless of course, and as a member of the society for the preservation of useless information I couldn't resist. Back to the carpenter. It might be clear that a good carpenter has learned how to measure his project correctly, that is, with sufficient precision while using as little measurements as needed. Back in the workshop he takes the rough material and draws the wanted shape based on his measurements. He could of course have drawn his measurements as bars given sufficient measurement the outcome would still be somewhat rounded since the jigsaw can't saw right angles in one go. In the digital audio things are about the same. Let's do a crash course digital audio. Like the carpenter, the analog to digital converter first makes sure correct measurements are done. The rule is to measure twice as often as the highest frequency that needs to be digitized. For sound 20 kHz is seen as the highest frequency so 40,000 measurements per second would suffice and a very steep filtering has to keep out anything above 20 kHz to avoid errors. In practice 44,100 times is used, technicians speak of a sampling rate of 44.1 kHz. Time to digitize. We start by taking a very small piece of an analog waveform that, for better understanding, is an unnatural straight line. Horizontally the time is projected while the amplitude is projected vertically. Now the signal is measured every 44.1,000 of a second and the values are written to a table. This table can be stored for future use or transported to another location. At the place of reproduction the table is read again and used to plot the voltages by means of a digital to analog converter. When passed through the reconstruction filter the straight line reappears unless something went wrong. The biggest problem that can occur is inaccurate timing of the reconstruction of the voltages. For then the straight line doesn't become a straight line again as can be seen here. The second potential problem of course is that the plotted voltages deviate from the original, being higher or lower than intended. But this hardly ever happens nowadays. At least not to the degree that it will become easily audible. OK, 1176 hand bone wanted a sine wave so let's digitize a part of a sine wave. It's the same again, voltages are measured and stored in a table. If we now take away the waveform you can see on the left a graphic representation of the values on the right. And you can still see the waveform in it. In playback that waveform will be reconstructed perfectly minus small errors that always occur during conversion or analog recording by the way. But there are those and I expect 1176 hand bone to be amongst them that think the reconstruction will be like this. Again I can't blame him for this is what the salespeople and advertising agencies have been promoting for years. It's wrong though. If we look at the first vertical jump that would need an instant rise in voltage, which is simply impossible since we have filtered the signal at 20 kHz the system is limited in speed. If you would try to achieve this instant rise, the result would be more like this, slowly rising in time. It might be clear this representation is incorrect. Now if you connect the dots knowing the slowness of a 20 kHz limited system you will end up with the original waveform again. It's like the jigsaw that is unable to saw right angles. Perfection is a goal, not a reality. But a well-designed built and set up digital system can come very close nowadays. The only real limitation today is the filtering that is needed at half the sampling rate. A filter at 20 kHz will cause time smearing below 20 kHz, something advanced long filters can almost solve like the ones by CORT, PSOR or even others. Or the appetizing filters and the related filtering in MQA designed by Stuart and Craven and found in more and more DAX these days, especially when combined with 192 kHz sampling where less aggressive filtering can be used. This is the area where big steps are made recently. Also the time accuracy, as I have shown in the example with the straight line, appears to be more critical than expected. Both the time smearing filters and the timing accuracy of the digital signal are time-related and it's only recently discovered how time-critical our auditory system is. And if you think these things are unique to digital order, you'll think again. While on flutter, rumble, tracking distortion, channel imbalance and many more limitations pester analog audio. It is therefore impressive how well both analog and digital audio can sound and how bad it can sound when done wrong. Still, the most impressive improvements happen in digital audio, especially in file-based audio and I will track new developments for you. So, if you want to stay informed, subscribe to this channel or follow me on Twitter, Facebook or Google+. See the show notes for the links. If you have a question posted below this video, please don't ask me for buying advice. See my About Questions video to find out why. If you like this video, please consider supporting the channel through Patreon as these super exclusive videos too. Just one dollar a month will do. The link is in the show notes. And don't forget to tell your friends on the web about this channel. I am Hans Beekhuyzen, thank you for watching and see you in the next show or on theHBproject.com. And whatever you do, enjoy the music.