There have been plenty of threads on this topic before. I don't remember the explanation, but yeah...if I recall correctly, they're essentially the same thing achieved by different methods (I think??). I'm one of the people that could care less as long as it sounds good. I do know that my new-ish DAC sounds better than my old Studer CDP, but I don't know if that's due to upsampling or not.
19 responses Add your response
I agree, Phild... we need to evaluate equipment on how it sounds and not (for the most part) on the technology used to get there. As far as your new DAC, digital filters have no doubt gotten better (and I believe will continue to get better), but in the end the performance of the product is purely dependent on the designers skill in implementing the whole package (digital and analog).
The method is the same for both up/over sampling. I think some manufacturers latched on to the high rez format buzz words of 24/96, 24/192 and just performed a small integer (or non-integer in the case of redbook CD to 96 or 192) up/over sampling in the first digital filter, followed by a standard 4x or 8x up/over sampling in the second digital filter. As the article states, if you add another digital filter it WILL change the sound for better or worse, depending on the computational power and sophistication of the filter. I also found it interesting that using two digital filters in series is a cheaper way to achieve a certain technical spec.
Is anyone else troubled by the statement in the article, "but the public really wants something like this. It's like trying to sell a seven-year-old on Santa Claus -- it doesn't take much selling."? It ticks me off!
I find it interesting that the statement "we need to evaluate equipment on how it sounds and not on the technology used to get there" is being put forward by someone who clearly understands the technology well.
If this statement is taken to mean that people can get drawn in by marketing hype around techie words, then I agree. If this statement means that performance should be evaluated based upon sonics and not on design, then I disagree. Case in point - The shoot-out between VHS and Beta format is a fascinating technology story. Warms my heart to know that R&D can really break through what look like fundamental performance barriers. I will always hold out hope for digital play back, and want to know the story along the way.
Just ordered a Belcanto DAC2 and now this! Are they really the same? I doubt. My imagination is that oversampling is a linear interpolation from one sample to the next, in 4 or 8 steps, while the interpolation from upsampling is of a sine-wave shape created by coresponding algorithms. Well, I can't pretend that I'm right on this.
Anyway, it's obvious that a conversion from 44.1 to 96 or 192 kHz isn't just oversampling in the mentioned way; there had to be 88.2 or 176.4 kHz. It's first of all a sampling conversion. I don't know if there's any advantage from that, but it surely is different from common oversampling.
And according to German magazines it sounds clearly better than with normal CD-Players' (oversampling) techniques. The reviews refer to DCS Purcell/Delius + Elgar, Chord DAC 64, Electrocompaniet ECD 1 and Musical Fidelity A 3-24. Especially the DCS combo is said to be very close to the hi-rez formats. If there is a possibility to switch between upsampling and non-upsampling, the difference is clearly in favor of the former.
Hellosimplymusic, I agree that the misconceptions regarding up/oversampling have reached critical mass and that the marketing hype will likely continue.
I wouldn't say I understand the technology well. I am not an EE specializing in digital design, although I am an engineer. That said, I do not think this is conceptually difficult to understand.
Judit, I wrote that to mean that I feel purchasing a piece of equipment based on the marketing hype is not likely to result in sonic satisfaction. The component may sound better, it may not as well. Our ears will ultimately determine our satisfaction, notwithstanding certain short-term psychological effects (eg. this "should" sound better affecting it "does" sound better -- not sure what this phenomenon is called?).
I am not familiar with the VHS/Betamax story, but I have heard it to be very interesting. Something along the lines of Betamax being technically superior, but marketing and timing resulted in VHS winning out. Is that essentially correct?
Ultimately, I think we all want better sound. Some may be quite content with knowing their ears tell them it is better while others want to know how/why as well. For me, I want to hear the differences for myself and know how/why also, probably the engineer in me coming out. But, no matter how interesting I find the technology, if it doesn't sound better I don't want it. And BTW, I don't consider "upsampling" new technology. Better filter design, yes, upsampling, no.
I have not heard any of the new formats (DSD or 24/96). They likely offer greater "potential" for sonics, but without a sizable library of software and clearly better sound, I probably will stay on the sidelines. I also have hope that significant improvements in Redbook CD will continue as long as 16/44.1 software is still produced. This, I believe, is good for us all!
Germanboxers, The VHS/Betamax story is much more interesting. As I recollect: The original betamax was a superior medium for encoding audio along with video. However, in the best spirit of technical competition, a multiplexing scheme was developed for VHS that kept them in the running quality wise. The format wars ensued with better quality being produced at every iteration. This was great reading back in the days when people were deciding what type of machine to buy. Ultimately Betamax lost out due to SONYs shortsightedness with regard to exclusive patent rights.
If you like how the Belcanto sounds, what's the problem? The sound is the most important part, at least for me.
Upsampling and oversampling are indeed the same thing. Both are sample rate converters. Both increase the number of samples by some factor (2, 2.18, 4, 4.35, 8, or whatever). Both will need interpolation to assign a voltage (amplitude) value to the newly created samples. Although a separate process, both may increase the word length from 16 bit to 18, 20, or 24 through the appropriate use of dither (basically white noise). Both may use a dac of various resolution (18/44.1, 24/96, 24/192 or whatever).
As Charles Hansen (research director for Ayre) said in the article I posted above, "Upsampling and oversampling are the very same thing and anybody who tries to tell you differently is misstating the case... And there's nothing unusual about putting two digital filters in a row -- virtually every digital filter is a cascade of 2x stages, because it costs less than accomplishing the entire filtering process in one go".
The reason why turning off upsampling on a capable player changes the sound is that you are taking one of the digital filters out, Particularly if it was designed to utilize the computing efficiency of two digital filters, it is likely to sound worse with only one. Again, Charles Hansen: "upsampling almost always makes a difference, and it can make an improvement".
In the end, it's the sound that matters. If it sounds better and the mfg happens to market using the "upsampling" hype, then fine. I do appreciate mfg's who talk straight about the topic though and if two products sounded identical (or maybe I should say the final subjective value I assign to the two products are the same), I will award the straight talking mfg with my business. That's just my philosophy.
Germanboxers, «...it's the sound that matters.» Of course you're right. And I'm about to believe your right with up-/oversampling, too until now I didn't find any clue that the interpolation with upsampling is based on a sine-wave algorithm instead of a linear one (as with oversampling). But I won't give up completely.
Instead I've found this article (on that site):
Upsampling CDs to DVD-A high-bit standard is just the same as oversampling. Or is it?
by ANDREW HARRISON
While there is certainly overlap (but never uplap!) in the use of the terms oversampling and upsampling, some guidelines can be given to differentiate the processes.
Oversampling is typically used to describe a technique used when transferring between the analogue and digital domain, where a signal is sampled many times over and above that actually required by the sampling frequency.
Oversampling in the context of the D-A process involves multiplying the sampling frequency by a whole number, typically between 4 and 32, or even higher. For example, in 8x oversampling, CDs base rate of 4.4.1kHz is raised to 352.8kHz by introducing seven new empty samples between the original data samples. These new samples, though, are often not just empty strings of noughts, but based on mathematical models to assist the DAC to work more linearly with the extracted data.
Oversampling, as well as easing the workload of the anti-aliasing filter, which can now operate more gently at a higher frequency, can also reduce distortion created when those analogue signals are first turned from continuous, analogue waveforms into stepped, digital, stair-like curves. This quantization noise is now spread over a larger band after oversampling, and can even be somewhat shifted out of the audible envelope by the technique of noise-shaping. Sony/Philips Direct Stream Digital, as used
in SACD, takes this idea to its limit, in order to dump high levels of digital noise up to higher frequencies than are not directly audible.
Upsampling is a solely digital domain process where the data stream is also stretched out by interpolation guessing the points in between, again mathematically and is typically used to refer to small, non-integer changes, such as from 44.1 kHz to 48kHz. When the change is larger than this, such as 44.1 kHz to 192kHz, upsampling is a more popular term.
'There is apparently no extra information in the upsampled signal that was not present in the initial signal, says Mike Story of dCS. With a 44.1 kS/s input, both the input data stream and the upsampled data stream will only contain a spectrum that must be between 0 and 22.05 kHz and is probably only between 0 and 20kHz.'
'This conventional analysis starts from the viewpoint that the behavior of the ear can be described in mathematical terms using Fourier analysis. This assumption is probably pretty good it means we are interested in frequency responses, for example, and these do provide good guides to the performance of equipment and to descriptions of what we hear. The analysis was right at the heart of the definition of the audio coding used on CDs.'
For those working with audio, it is also apparent that theories based on these descriptions are not completely adequate, and that there can be significant differences in the performances of pieces of equipment with similar "conventional" specifications. It seems that two things are going on here: the ear may have more than one mechanism at work; and sine waves may not be the best function to use as the basis for analysis. On the mechanism front, it seems highly likely that the ear has a sound localization mechanism ("where is it?") that is fast, and independent of the mechanism that says "its a violin", and that is related to transient response. There may also be a third mechanism at work. On the analysis front, it may be that some form of wavelet is the best basis for mathematical modelling. The problem here is that sine-wave theory is relatively simple, and has been fully worked out by generations of mathematicians, following on from Fourier. Wavelet maths is just plain hard work, and does not yet have anything like such a solid core of mathematical results to call upon. Our ears, however, are not waiting.
It's me again...
Having read this site a second time, I'm convinced that the sampling rate conversion makes the difference between up- and oversampling and possibly in sound quality. One thing that's rather clear is that such a conversion involves a sine-function interpolation (!). And that's exactly what probably makes the sonic advantage. (Without engagement...)
Martian, while I agree that the term "upsampling" is generally used in conjunction with 96kHz and 192kHz, both non-integer multiples of 44.1, I disagree that there is any fundamental difference from oversampling. I do believe that 96kHz and 192kHz are used specifically to wrap the manufacturer (and convince the consumer) that this is somehow as good, the same, or close to the same as the higher resolution format DVD-A (true 24 bit, 96kHz recordings).
"Sample rate conversion" simply means changing the sample rate from one rate to another. If you recorded a concert at 88.2 kHz DAT, the sample rate conversion was a simple 1/2 of the original rate to burn it to redbook CD. Most DAT's, however, are 48kHz or 96 kHz, so the sample rate conversion to Redbook CD is a non-integer conversion.
The "bits and pieces" (pun intended) of actually changing the sample rate are the same whether it is a non-integer or integer change. Extra samples are created and voltage values are assigned for these extra samples based on interpolation of samples before and after. Several techniques are available for the digital filter designer to choose from, one of which would be a linear interpolater. But an interpolation algorithm based 2nd order, 3rd order, or sine function can be used as well. This is the same for both integer and non-integer sample rate conversion (over/upsampling) and is just one of many factors a digital designer must consider.
I will leave you with this Wadia quote from the article you linked:
When used to convert a CD signal to a higher sample rate, the process of sample rate conversion is mathematically synonymous with over-sampling. Whether this process is performed in a digital filter housed in the same chassis as the D-to-A converter or in a ieparate chassis has little bearing on performance. Any advantage that can be claimed for a rate-conversion system can equally be achieved in a sophisticated over-sampled system such as the Wadia DigiMaster.
Wadia oversampling advocating:
«...Any advantage that can be claimed for a rate-conversion system can equally be achieved in a sophisticated over-sampled system such as the Wadia DigiMaster.»
I can follow you. But at the same time it's clear that the interpolated curve resulting from an upsampling (non-integer sample-rate converting) system isn't identical with the curve from an oversampled interpolation after smoothing by the low pass filter. Though we can't say which one is better or more adequate, there is a difference. It can also be deduced from the reviewer's (Andrew Harrison) judgement, which btw. clearly favors the «upsampling» dCS against the «oversampling» Wadia. No deciding argument, of course, due to the very different devices.
The concern with the Redbook standard that Up, Over or Down sampling miss...is that the biggest concern with all of your/my/our CD players is still where it was when we were all buying Sonic Frontier, Levinson, Classe DAC's and things like the Genisis Lens and SF Ultrajitterbug...the two areas that really have big impact on your sound, gentlemen, is jitter and the audio section of your player.
Why isn't this addressed(marketed) by CD/DAC manufactures..well it's because the cures are expensive. It is much easier to buy the latest Burr-Brown DAC...(which is a chip) and tout that, rather than do the serious structural/transport/receiver work that is the jitter end, or speak of the jfets or tubes(let alone the wire, caps, resistors..etc. that we all worry so much about in pre-amps etc.) or whatever in the audio section. Go to look at the ad/flyer on-line for some of these older DACs and see where they address jitter and the audio section.
Most of our/your most used source piece..the CD player has no better jitter than prior generation units and usually poorer audio sections..even very $ players are using op-amps for the audio signal. Kinda silly to worry about the tube vs. SS or silver vs. copper wire..etc. when the audio signal in most systems is starting in an op-amp that none of you would think much of if it was the audio section of your pre-amp!
So, my friends, up and over...is a new chip to put into a player or DAC that is easier to do and market..than the more costly..and harder to do improvements in jitter and audio signal quality.
This is interesting. Despite Wadia's previous efforts at dispelling the marketing hype of "upsampling/oversampling", they now have embraced it. (See quoted text below)
Or have they? They use the words "upsampling" in describing their process, but they don't "upsample" to 96kHz or 192kHz (non-integer or "asynchronous" up/oversampling). They actually use a 63X (~2.8MHz) oversampling. Is Wadia being disengenuous now or are those (mfg's and reviewers) who have been touting "upsampling" as something new and fundamentally different the ones who have been disengenuous?
I actually applaud Wadia in this. They have made efforts to inform the public that up/oversampling are the same, but it clearly has not worked. Plan B: Give up trying to point out the marketing hype and dilute it by embracing the terminology. If every mfg uses the term "upsampling", then is there any marketing advantage by touting it? Not a bad plan I suppose.
BTW Martian, Wadia use a 12th order polynomial spline curve fitting to interpolate the extra data points.
Wadia Technical Bulletin, March 2001: Up-sampling and Wadia Technology
«BTW Martian, Wadia use a 12th order polynomial spline curve fitting to interpolate the extra data points.»
Sounds good. I don't know exactly what it is (just know their «spline» low-pass filter), but it sounds like the famous sine-function interpolation which I've always associated with Wadia and also my Theta Pro basic II, BTW.
(Wadia:) «...asynchronous interpolation discards the original CD data and can actually reduce the clarity and detail of the original recording.»
I would'nt expect any other statement from them ;-) But obviously it CAN also occasionally open new sonic horizons... [/euphoria off]