gerbache

In practice, interpolation will result in a slightly different signal than a non-interpolated one. As to whether it's a better signal than the original, that's up to the listener to decide, but it will be slightly different, depending on the method of interpolation. With any PCM encoding of a signal, there's a set amount of noise caused by the number of bits of quantization. As you increase the number of bits, this level of noise gets lower and lower, until, especially with these newer 24-bit ADCs, the amount of noise added from the quantization is below the noise floor of the analog components backing it up. In other words, the quantizer isn't adding any appreciable noise on its own. Now, as I mentioned, the number of bits basically sets the amount of noise present. The sampling frequency will basically set the frequency at which that noise appears, because the actual value of the noise will be different at each sampling point. When we use these really high speed, 96 kHz or 192 kHz ADCs, the frequency of the noise is largely pushed out to the realm of 40+ kHz, which is inaudible to humans, so as was pointed out, we don't notice it anymore. The biggest benefit of this is that any ADC (and consequently, DAC) that performs at this rate is going to have very little noise present at any frequency that matters to us. By upsampling, we're attempting to change the signal so that we can take advantage of these benefits. It's entirely possible, depending on the manner in which the upsampling takes place, that it could cause an audible benefit that you don't get with a non-upsampled signal, even though the input signal is the same. With a lot of these new ADCs, the limiting factor in things like frequency range and SNR is in the analog electronics around them, so it's not quite as important to have a super ACD.

I should have read Max's post before making that response..his explanation of how upsampling works is actually spot-on, and much simpler than what I said. In a nutshell, an upsampler smoothes out the spaces between each sample; if it does a particularly good job of it, it could potentially have a benefit on the sound output.

A quick pedantic note on the way upsampling works: Even though you're right, no new information is added to the signal when you upsample, it doesn't just double each value. It performs an interpolation between the two values, such that rather than being the same as the value preceeding it, it'll be in between the previous and next value. I haven't looked much into the technical method of how they upsample in these players, but they're probably doing some relatively sophisticated methods of interpolation that may have some impact on the sound at the output. It's also worth pointing out that many upsamplers out there nowadays aren't increasing the frequency in integer multiples, so it has to provide a different method of mapping the new values, which I will almost guarantee is some form of interpolation. In any case, upsampling can actually have more of an effect on the sound than simply doubling the sampling frequency, even without any "real" extra information. Whether this is an improvement or a detriment to the sound would depend on how they do the interpolation, but you're right that upsampling doesn't really create any more data from the original source than was encoded on the CD.

Canon's lower end models get really good reviews. From my experience, they're really well made and easy to use. The Rebel cameras are nice, and nearly as feature-filled as the next line up (the EOS series, if I remember correctly), which basically buys you a more sturdy body build. One nice thing about Canon is that their lenses are in a really broad range of price, from the cheapie stuff they give you with the camera all the way up to the huge, professional lenses you see at sports events. One lens I really like (if you're willing to sacrifice zoom capability) is their 50 mm, f1.8 lens, which can be found for around $50, I think. Super sharp pictures, if a little cheap feeling. Nikon's lens lineup is pretty similar to Canon, since they are both in consumer as well as professional cameras, but the Nikon lenses seem to go a fair amount more expensive than the Canon. If you want to go a cheaper, but solid way, and she is willing to learn to use a non-autofocus camera (it's not that hard, although I admit that I have an autofocus body myself, in addition to my old manual), you can find older bodies of Nikon, Canon, or Minolta for pretty cheap on Ebay, and can usually snag a body plus an assortment of lenses for that kind of price range. I like the older SLR bodies, personally. My old Minolta is built like a tank, and has survived more torture than I figured would be possible. These old cameras will take pictures that are at least as good as, if not better than the newer models, plus a quick way to learn the ropes of SLR photography is to have to do it a few times with a fully manual camera.

---------------- On 6/26/2004 1:08:53 PM Marvel wrote: If the source and load are equal impedance, you will have the maximum power transfer. If the headphone output were 50mw, the amp (if the same impedance as the phone amp) would only get 25mw. As the load impedance increases, the efficiency will go up, but the power transferred will actually decrease. IOW, you probably won't get too high a signal to the power amp, considering the high impedance of the amplifier input. This is a standard textbook way of looking at impedance matching, but in general, amplifiers like this don't follow this rule. Normally, you want your amplifier output impedance to be considerably lower than the load impedance. In addition, you can find headphones that range anywhere from 32 to 600 ohms, so it would be next to impossible to design a headphone amplifier such that it is "matched" to the headphones. It turns out that if you look at the way these amplifier circuits work, you actually generate maximum power when your output impedance of the amplifier is as low as possible. That being said, some years ago, a "standard" was designed proclaiming that all headphones should be able to expect a 120 ohm output impedance regardless of the impedance of the headphones themselves. I'm not sure how often amplifier manufacturers follow this standard, but nonetheless, many headphones are designed with this in mind.

---------------- On 5/26/2004 6:32:08 PM garymd wrote: You should hear the original mixes of Anthem Of The Sun and Aoxomoxoa. Very different. More spacial if you will. They're hard to find since they were pulled from the shelves very early on but I have a DJ friend with both. Very cool stuff. Do you know anything about the white cover vinyl release of Anthem of the Sun? It's a different mix that really has a different sound from the widely released version. My dad has a copy of it, but I've only heard it once, so I don't really remember all that much about it. Think that might be the mix you're talking about? I need to borrow it from him and make a copy of it for both of us...

Live/Dead is a great intro to their live shows, especially from the earlier days. In the Dick's Picks series, probably my favorite is Volume 4. Don't remember which year it was from, but it was recorded at the Fillmore East and has some of their best live stuff I've heard. The entire Dick's Picks series is well worth a listen, though, and is really pretty reasonably priced when you consider the amount of music they all include. That being said, I'm a huge fan of live Dead music. Some people aren't at all fans of the Dead, but will really get into the more accessible stuff like Workingman's Dead. You really just have to give it a chance, though. I will have to say that their first few albums (Anthem of the Sun, Aoxomoxoa, for example) tend to not be quite as accessible as their later ones. At that time they were trying to emulate their live sound in the studio, and it just doesn't work quite as well. I really like their first few, but I can easily see how it wouldn't be the best intro into the Dead.

They're making home networking solutions now which use the AC power lines of the house as a cheap, easy way to connect all the machines. The way I see it, yeah, it's a really simple way to get the problem solved, but there's enough noise on the line without me adding to it. Speaking of noise, I have a light in my new apartment right now that has been going nuts, looking like the voltage is fluctuating WAY up and down. Off to go see if I can find the cause for this one...

I'll throw in another recommendation for the Hsu VTF-2. Great little sub for 2 channel only use. I'm running it with my heresies right now. Most of the time I don't even notice that I have a sub in the room, unless for whatever reason, I have to turn it off. Suddenly, the whole system just seems to lose something. It seems to reach down as far as I could ever want for two channel purposed, and definitely far enough for jazz/classical. I'd highly recommend it, plus the price is pretty nice (right around $500, I believe). Mine is one of the original designs, so I can't comment exactly on the new model, but I can't imagine that it's any worse, and probably a whole lot better.

---------------- On 4/22/2004 7:53:33 AM mdeneen wrote: I agree with Ray on the economics, and I think there is merit to the headphone sans room acoustics arguement, although I would never want to do all my listening through 'phones. Ray, do your Sennheisers suffer the "constantly intermittant connnectors" problem? Perhaps it's me, but every pair of Senns' I have owned, within weeks the tiny connectors go intermittant. I think their cable is the WORST design I have ever seen. Those connectors are crap for one, and the "cowboy hat" style dual-cord under your neck has got to be the most uncomfortable style possible. Great 'phones, somewhat cheesy design, IMO. mdeneen ---------------- Don't they use that style of wire because you're supposed to be able to change them? My dad has a set, and I vaguely remember seeing something in the book that came with them that you can buy other wires to "upgrade" the ones that are built in. Not that this makes the dual-cord style any more comfortable, but at least it gives an opportunity to get rid of a staticy set. I don't think dad's are staticy, but then I haven't heard his in forever now. I suppose if you're paying the money that Sennheiser costs, you ought to be able to expect a set that doesn't need to have new wires added on, though. Anybody have any other favorite headphones? I might be in the market for some in the near future, but Sennheiser are the only top notch ones that I'm familiar with (and can possibly afford).

It looks like it does a different type of upsampling. The 795 upsamples to 24 bit, 176.4 kHz, while the 963 upsamples to 24 bit, 192 kHz. Honestly, I'm not sure if this makes much of a difference, but 176.4 is just a multiple of standard cd audio, so it is probably a less sophisticated type of upsampling. As for whether it makes a big difference or not..well, I don't know about that. Honestly, I can't hear a huge difference between the upsampled and non-upsampled audio on my 963. The product specs for the 963 are definitely more upscale than the 795, with stuff like separate video and audio sections and a more impressive sounding video decoder, but then if the product specs weren't much more impressive sounding, I'd really wonder about Philips, considering the 963 costs twice as much. I'd be really curious to hear if someone tries to compare the two. I like my 963, but the convenience of a 5 disc changer is really nice to have, as well.

Sadly, THD is not really a good measurement of sound quality. Pretty much any amplifier you buy these days can be made with an impressively low figure for it, so many people cease to use it as a guideline. If you look at some of the tube amplifiers that are very well regarded, the THD figure can be up in the several percent range. In other words, if it sounds good, I wouldn't worry too much about what the manufacturer quotes for THD.

*****es Brew is a tough one to listen to if you aren't a fan of that type of music. It seems to be an album that people either really like or just realy don't get. I wouldn't base you're entire opinion on Miles Davis on that one, anyway..

If you aren't opposed to using a subwoofer, you can fix the big problem with the Heresies pretty easily. I'm running a Hsu VTF-2 with mine, and it blends well enough that it really sounds like the Heresies just suddenly reach down to the 25 Hz region. What's really amazing is when I turn the mains off and just listen to the subwoofer's output. The thing is barely on most of the time, but the speakers sure sound a lot bigger! Granted, I could probably have gotten the bass I wanted by just getting the Forte's in the first place, but with the subwoofer, I do get bass down a bit lower than Forte's can manage, so it's all a tradeoff.

---------------- On 3/23/2004 1:38:06 PM Griffinator wrote: Fine. DVD-A is better because it uses multibit delta-sigma modulators that are capable of employing dithering and optional noise shaping to remove digitization artifacts from the stream. A true 1-bit delta-sigma cannot employ such technology - how exactly do you dither one bit? You can't. That's why I stated that "1-bit delta-sigma sucks" Not that d-s converters are a bad technology, but rather that single-bit d-s converters are a bad technology. No one, not Burr-Brown, not anyone, uses 1-bit converters anymore. It's oversampling and dithering, both of which have made a positive, not negative, contribution to the fidelity of digital audio. For Sony to declare that these functions are a bad thing? I guess they're hoping we've all forgotten how utterly bad the first CD players sounded. ---------------- I haven't heard Sony declare these things bad. They're still doing plenty of noise shaping and such on the DSD signals. DSD doesn't just involve a straight 1-bit delta-sigma modulation. Sure, the final result of the whole deal is down to a 1-bit message, but there's a lot more to the technology than just a simple delta-sigma modulator...

Now attacking the DSD technology on the basis of financial motive to change the studios over, that I can agree with. The first few posts you made about the technology had nothing about the business aspects, but were attacking the technology. The technology itself is fine. I'm still inclined to believe that DSD may be a superior technology in the long run, simply because it theoretically can involve less conversions in the digital domain. That being said, I can't argue that right now, there may not be much motive for the studios to convert their equipment from PCM to DSD. This does not mean that the technology is inherently flawed, though. It simply means that there is not enough motivation at this present moment to warrant the conversion.

---------------- On 3/23/2004 12:01:21 PM Griffinator wrote: You're missing the point. Take a read here. http://www.analogzone.com/acqt0310.pdf The delta-sigma modulator by itself is a terrible converter device. It was only the advent of oversampling and noise shaping that made it feasable for use as a quality audio device. DSD is a de-evolution of the process. They declare that by reverting to the single-bit setup, but using an incredibly fast (2.8Mhz) sample rate, that they will somehow produce a better final signal. Even Sony's published data is misleading - http://www.proav.de/data/DSD.html Not exactly. They're just saying that by simply using the already oversampled and noise shaped delta-sigma modulated output, they can avoid the extra decimation steps required to encode this into 24-bit PCM. I'm not missing the point that straight delta-sigma modulation is not acceptable, I'm just saying that DSD is eliminating some of the steps required. Did you look at any of the information on the TI website. Reading through the data sheet for the latest 24-bit encoder chip, they're quoting a dynamic range for -both- DSD and PCM as approx. 112 dB. This is because, while theoretically both can manage more, you cannot achieve this in the real world with current technology. This is also due to the fact that both the PCM and the DSD outputs for these chips start out life in the same way. The theory behind both technologies is solid. From the article: The main problem with standard PCM technology is that it requires brick wall filters to block frequencies above 20kHz... Absolutely incorrect. The brickwall filters Sony refers to here are strictly for the redbook CD specification, not DVD. 24/96 DVD spec does not engage these filters until well up into the 40Khz range, with much more gentle sloping, which reduces the impact on the actual signal. This site isn't hosted by Sony, nor does it say that they wrote it anywhere that I am seeing. Even if Sony had something to do with it, we all know that this sort of thing is written by people in marketing, most of whom are trying to find the most impressive way to say things to people who do not know the technical specifics. I'm well aware of the fact that DVD spec audio doesn't require the same brick wall filters, but for that matter, neither does DSD. DSD systems only consider the change in amplitude from sample to sample 2.8 million times per second. Thus, they can record the signal in any number of relative steps rather than having to determine which of a fixed number of amplitude values best describes the signal at any given sample point. This is where the inconsistencies start showing up. DSD is still subject to implementing oversampling in order to accomplish its feat, which means that it's really not a true 1-bit process. The high samplerate is supposed to compensate for the lack of noise shaping, but there is still the matter of inaccurate sampling. DSD can only describe the amplitude of a given waveform in terms of +/- up to 6dB from the last waveform. 24 bit PCM, on the other hand, is capable of describing the same waveform with individual word lengths anywhere within a 144dB range. You tell me - is the relativity (and non-linearity) of DSD a superior alternative? That 24-bit PCM started out its life from the exact same place as a DSD source. In many chips, both are subjected to similar noise reduction methods until they do the final conversion to the output. Plus, you aren't going to get a true 144 dB range. The best analog electronics cannot approach a true 144 dB range, so the effective range is all that we care about, and from the data sheets on the chips I've looked at, both DSD and PCM are achieving the SAME effective range. More deception: Such a high sample rate allows the system to record frequencies far beyond the scope of human hearing. The result is a frequency response of up to 100kHz and a dynamic range greater than 120 dB. First off, Sony's own specifications for SACD demand a brickwall filter at 50Khz. So much for DSD not needing them, and so much for that 100Khz frequency response. Oh - and by the way - they just verified my declaration about dynamic range. 24 bit PCM (DVD-V and DVD-A spec) is capable of 144dB dynamic range. DSD is capable of a mere 120dB. Well, going pedantic here, even if Sony calls for a brickwall filter at 50 kHz, it's a moot point when comparing it to 24/96 PCM, because at a rate of 96 kHz, the PCM will ALSO require a brickwall filter at 48 kHz. As for the dynamic range, I addressed that above, and the fact of the matter is that you will not see an effective dynamic range of much over 112 dB in typical use anyway because of the limits of the analog electronics. Despite the connection between SACD and DSD, DSD is an independent format that can be used without the SACD itself. There are already some DSD recorders (stereo and multitrack) on the market and it seems that it will become the best choice for top-quality audio recordings. Now here's the dirt. I was, about this time last year, looking into purchasing DSD AD converters for my studio, so I could produce SACD masters. I talked with George Massenburg over on his forum about DSD and the feasability of converting DSD to PCM for non-SACD applications. He and several others on that forum echoed the same sentiment: Such a conversion would be disastrous, because the DSD signal is a noisy, non-linear one to begin with. A better choice would be to record high-resolution PCM (in the order of 24/192 or above) and convert to DSD should you choose to produce SACD masters. Now why would people who are directly involved in the technology (George has a large number of hi-res multichannel productions to his credit, in both SACD and DVD-A) say something like that if it wasn't true? Does he have a vendetta against Sony/Phillips? Highly doubt it. He owns several top-shelf Sony digital consoles. There's plenty of other engineers and studios who -do- like DSD, so I think finding specific examples on either side is a waste of our time. Really, I think both technologies are sufficient for our current state of the art in sound reproduction. Final point? When Sony has to pit SACD against CD, not DVD, to demonstrate its superiority, there's a problem. When they feel compelled to distort both their own specifications and those of PCM in order to reinforce that superiority, something is rotten in Denmark. Sony is comparing SACD against CD because that's the current leading format. I haven't seen -them- distort their specifications any more than you've distorted the specs of PCM. Both camps are trying to make themselves look as good as possible. Besides, I'm pulling data sheets, which are largely written by engineers for engineers. I really don't think you can say that TI is distorting the specs on these chips, but right now, BOTH DSD and PCM are achieving almost identical specs out of them. I think the real message this sends is that both technologies are fine, and the real war will not be over the technical features, but rather the marketing successes of all the parties involved. Let me just say, to clarify once more, I'm not trying to say that DSD is superior to DVD spec PCM. I'm simply trying to refute the statement that it is an inherantly inferior technology. Implemented properly, BOTH can be successful, in my opinion.

Now there's a topic I'll agree with that rather scares me. I see the music industry moving away from physical media altogether. While this doesn't necessarily mean a negative in terms of sound quality, the collector in me really hates it. I rather enjoy putting on a good album, regardless of the format, and flipping through the liner notes/album cover. Plus there's just something satisfying about walking into your music room and seeing a few hundred albums. Sadly, it seems that I'm in the minority in terms of the general population with this view, though.

Ok, so I posted too soon. Here's a better page for the curious: http://focus.ti.com/docs/search/paramsearch.jhtml?familyId=581&tfsection=param_table&templateId=5275&showAssociated=false This is a list of all the current audio specific ADC chips from Burr-Brown (TI now). Every one of them uses a "precision delta-sigma modulator". Considering that these all start out as delta-sigma anyway, it makes sense to me that by eliminating all the processing to make multi-bit PCM and just using the delta-sigma output directly (or at least with a little filtering), we may be able to eliminate some processing. That's the theory behind DSD. I'm sure this is not the "ultimate" in digital audio technology, as something else will surely come along in the future that will replace it, but it seems to me like it's the right direction to proceed for the moment.

A quick search on the TI (Burr-Brown) website shows that all of their current production 24-bit ADC chips use a delta-sigma architecture. Considering that Burr-Brown chips are highly touted by many a manufacturer, I would think they'd at least be pretty close to the cutting edge, so it appears that someone is using delta-sigma still. For reference, this is the website I checked: http://focus.ti.com/docs/search/vparamsearch.jhtml?searchId=25822&familyId=390&tfsection=param_table&templateId=4&showAssociated=false It may be that there are better ways of encoding PCM audio than delta-sigma, but this doesn't mean that delta-sigma is a bad technology in itself. After checking the other bit rates of their chips, they are using different technologies for rates other than 24 bits, but there still appear to be delta-sigma chips in all of them, so I think it's safe to say that this technology is still alive and well.

As for the long range archive abilities of DSD, how is it even possibly unstable? DSD itself is NOT a physical archival format, any more than simply saying that analogue or PCM is an archival format. I can store any one of these in many different physical configurations, each of which will be more or less stable long range. How can PCM be any more stable an archival format than DSD? If they're both stuck on the same physical media, they'll both be exactly as stable for the long term.

1-bit technology is plenty for providing all the dynamic range you want. PCM can be (and often is) encoded first using a 1-bit delta-sigma converter, then is processed into being a 16, 20, or whatever bit rate signal. The theory behind DSD is just as solid as the theory behind PCM, it's just that it's a different technology. Both still suffer from their respective problems, but both are still acceptable methods of conversion. This link has some information about delta-sigma converters that might be interesting: http://www.beis.de/Elektronik/DeltaSigma/DeltaSigma.html

Suppe is some fun stuff if you're into marches. Anybody know good recordings of it? I'm only familiar with it because we played them a lot in my high school band days. I'd really like to hear someone other than myself playing them...

building an amplifier is not necessarily a difficult proposition, but it does take some equipment that you may not have, especially in order to do it easily. For instance, you'll need a soldering iron and a multimeter as an absolute minimum. Life will be much, much easier if you have a power supply and oscilloscope, though, because it is difficult to demonstrate the fruits of your efforts without either. The power supply is not necessarily essential, if you were wanting to assemble your own, but if you intend to actually hook up speakers to an amplifier for a demo, I'd be sure to have an oscilloscope to test with once before you try it with speakers. Honestly, unless you're required to actually build something, I'd probably stick with doing some sort of a demonstration or talk about the theory behind them. Even with some experience, unless you purchase some sort of kit, building an amplifier can be a somewhat challenging task.

Exactly. For the purposes of music, any coupling caps in the amp will block all the DC present. In fact, pretty much all of this discussion is particular to watching square waves on a scope, but it does help to explain some of what we see when we're watching square waves. In theory, the lower frequency waves will probably be rolled off slightly, just like we've seen from Craig's first post, and at the high frequencies, they should look just like a high pass filter, which means that the sharp edges will be rounded out, without any "ringing," where they bounce above and below the wave. Again, this is just like how the square waves looked from the first post in this thread.