 The HB project and the HB channel are supported by Hi-Fi Klubben. That sound kills good music. There are many questions on upsampling and oversampling in the audio community. What do they do? What is the difference? And should I use it? Well, I do have some answers, but if you will like them, let me start with the difference between upsampling and oversampling. The name and the place it is used. Sort of. When in the late 70s of the previous century the CD and CD player were developed, Sony wanted 16-bit resolution but Philips found 14-bit sufficient since no one was able to produce 16-bit DA converters. Sony won and issued CD players that in practice only had 14-bit resolution. Philips applied oversampling and this way achieved almost 16-bit real-world resolution, although the converters varied between samples. What did they do? Well, a digital audio signal is no more than a large number of measurements of the amplitude of the signal. These measurements are called samples and the amplitude is measured in volts. For CD there are 44100 samples per second. In playback a very steep filter at 20 kHz must be applied to reconstruct the analog audio without conversion errors. Also see my video, the truth about Nyquist and why 192 kHz makes sense. The link is in the top right corner. What Philips did was take two of those samples, calculate three samples in between and do the same for the following samples. The advantage of digital audio is that you can easily apply all kinds of math to it, provided the calculation power is present. Philips called this four times oversampling and claimed rightly so that it gave higher resolution and lower distortion. This and the swing arm mechanism made the Philips CD players immensely popular. The only kind of DA converters at that time were the so-called ladder converters. A 16 bit ladder converter has 16 switches with a resistor in front of it. If all switches are on, the output voltage is two volts. If all switches are off, the output voltage obviously is zero volts. If only the top switch is on, the output voltage is one volt, the second switch provides half a volt, the third one a quarter of a volt and so on. People that are familiar with binary counting see what the plan behind this is since binary counting works identical. In other words, the most significant bit switches the top switch, the second significant bit, the second switch and so on. The problem with ladder converters is that when you come to the least significant bits, the voltages are so low and the resistor values so critical that it's quite difficult to make them sufficiently accurate. The 16 bit has to feed about 30 microvolts. In the DIY Audio Society, these ladder converters are still very popular and here they use a number of these converters in parallel to even out the errors. There are also manufacturers of these non-oversampling Nos-for-Short DA converters. The trick Philips applied was more or less a softer version of the stacking of the number of chips on top of each other. But instead of placing them on top of each other, they were placed after each other. But if an oversampled signal contains no more information than the original, why then oversample? Well, as said, to a certain degree it can increase the resolution of the DA converter by replacing the amplitude accuracy by time accuracy, but there are limits to that too. Then remains another advantage, although some see it as a disadvantage. By oversampling, the filtering is moved from the analog domain to the digital domain. The signal is oversampled in the digital domain and according to many digital filters do their work with less audible artifacts than analog filters. The resulting digital signal has a higher sampling rate and thus needs less steep analog filters that do sound better. OK then, but where does that leave oversampling in this story? Well, that's just semantics. Opsampling and oversampling are essentially the same. But the word oversampling is used by manufacturers and upsampling is used when you at home increase the sampling frequency. That can be inside your DAG that offers that as an option, but then it will be probably called oversampling. But it might also be done inside a computer using special software and then it's called upsampling. There is software that upsamples audio files, thus write new audio files. This has the advantage that the upsampling process can take all the time needed for you will play the file no sooner than it is ready. Another solution is real-time upsampling. The computational power of modern computers is more than sufficient to do this real-time provided the software used is clever enough to decouple varying calculation latency from the output signal. So, should I upsample? Well, that depends on a number of factors. Some people don't like upsampling and go for Nostax. But if you don't belong to this group, upsampling or oversampling might be good options. Again, both do the same but upsampling is outside the DAG while oversampling is the same but then performed by the DAG according to the conventions on the web. The question now is what is better, having the DAG do it or the software in your computer? There is only one answer possible on this question. Whatever sounds the best? If you own a DCS, a PS Audio Direct Stream or a Quart Dave or something in the same class you might find that you better have the DAG do its work. But if you own a two-year-old, 150 euro DAG I am almost certain that you better have your computer do the upsampling. The problem is that there can be many settings in upsampling software and that a set of goodies and some technical awareness come in handy in setting the best way of upsampling. Also do realize that one DAG will perform better with one setting while the other DAG might require another setting. Filter settings are highly bespoke. Upsampling might result in a better sound or it might not. It is a complex matter and if you don't want to study a lot and search the forums try the default settings of the software you use. If that sounds better, use it. If don't, don't use it or visit forums, study manuals and other sources and make it your hobby. Whatever you do, there is one thing you shouldn't do. Ask me for help. I just lack the time to study filter settings in detail let alone help you find the best setting for you. I will keep looking for ready-to-use solutions and report those to you. So subscribe to this channel, follow my Facebook or Google Plus page or my Twitter account. You can also post questions there but please don't ask me for buying advice. View my questions video to find out why the link is in the top right corner. You'll find more information below this video on YouTube. If you like this video, please give it a thumbs up and tell your friends on the web about it. 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.