Bodyslam

OK, I got that off my chest, thanks everybody for humoring me. Now maybe I can make some positive contribution. (And Teresa, please don't think that I was picking on you. My beef with the SPARS Code goes way back before this thread.) I'm in agreement with everyone who has expressed some version of this thought: we'd like to see complete and accurate information about the process used to bring us these Hi-Res files, and we don't want to find out after the fact that it was upsampled from some low-res source. Personally, I'm not too interested in Hi-Res PCM files that have been sourced from an SACD, because I haven't much cared for the sound of files prepared that way. So I wish that all the download sites would make it clear when that's the source. In those cases I'd rather just listen to the SACD. Anyway that's just one obvious example. In an ideal world we'd have complete information right on the page about each title. I also know that HDTracks for instance gets their titles from many sources, HDtracks doesn't control the record labels, oftentimes the people in the labels don't really know what they have, and if HDTracks held out for this complete information they would only have half as much music on offer. What's the solution? All I can think of is we keep encouraging them in every way possible (short of class action lawsuits--get real) to provide it. Send 'em an email every time we place an order, reminding them that we care about quality. BTW the blurb I wrote to accompany the WFD files was one small attempt to provide this kind of information. Some version of that information would also apply to the other files I have provided for them--some more Concord material, nine albums from Acoustic Disc, a couple of Blind Pig releases, maybe a couple that don't come to mind right now. I think I'll write them an email next asking them to consider putting that information on the download pages for those other titles too.

Way upthread Teresa proposed a new, extended version of the SPARS Code, something that would end up something like this: "So the code could look something like: Recording: DSD 5,644.8 Mixing and Editing: DXD PCM 24/352.8 Download source: Original master file Download file: 24/88.2" The original SPARS Code was a pet peeve of mine, so I'm going to blow off steam for a moment here, and explain why I think this is a terrible idea. Of course I understand where this is coming from: more disclosure means better educated buyers and less room for confusion. However, my objection, as it was to the original SPARS Code, is that it can't help but become an implication that the equipment is what determines the worth of a recording. Nothing could be further from the truth. Assuming that you're starting with music that's worth recording in the first place, the choices made by the engineers and producers--how they hear things and what they value--has a far larger impact on the quality of the recording than any piece of gear. Why don't we propose a rating system that shows at a glance how good the recording actually sounds? Well, the problems with that are obvious, no matter how useful that would actually be in purchasing decisions. But if you can't do that, then slapping on a rating based on the gear just devalues the work of the people, and I find it offensive.

The Bill Evans Trio: Waltz For Debby 192/24 version that is offered by HDTracks is actually sampled at 192. I can say this with some certainty because I did the transfer, and handled the files personally all the way until they were submitted to HDTracks. Before you get too worked up here, consider a couple of things. You may have unrealistic expectations about the graphs you're looking at. Wgscott stated categorically that there should be signal up to 96K in there, and that would be true if those frequencies existed in the original recording. But did they? A recording made in a nightclub in 1961 to 1/4" analog tape? Where would they have come from? (I mention "made in a nightclub" because that implies to me that they didn't have the most pristine microphones, as they might in a studio. But maybe they did, who knows?) The noise band at around 38KHz is an artifact of the ATR-100 tape machine that was used in the remastering. It has long been known among people who use that machine--it's leaking in from the motor control circuits--but widely ignored because no one ever heard a problem from it. In fact lots of people in the studio world consider that their favorite analog tape machine. Who can say that the 38Khz doesn't contribute something that they like? By the way, how'd the files sound to you? Barry has mentioned that he hears a qualitative difference between the 2X rates (88 and 96) and the 4X rates (176 and 192) and I hear it pretty much the way he describes it. As we go up from 16 to 24 bits, and as we go up from the 1X rates to the 2X rates, I hear a number of specific improvements. When we get to the 4X rates done well (and here I agree again with Barry--easier said than done) it's more of a feeling that we have turned a corner and we are almost dealing with a musical experience rather than a facsimile of a musical experience. And I'll confirm that Keith Johnson has said something similar in several conversations. I'll go further: Those of us who work in digital audio understand the relationships of sample rate to frequency response, and bit depth to dynamic range. Theory says that higher sample rates allow us to record higher frequencies, and in practice that's true. But here's something that the theory doesn't account for: every time we double the sample rate (up to 4X) the bass gets better. Much better. More dimensionality, more texture, more clarity, better decay, lots of things. I'm just trying to make the point that digital audio is more complicated, and more subtle, than the first-level theory that we all learned. However, I'm not suggesting you pay extra for any 192K file if you don't think there's value to you in it. God forbid that I should be responsible for your spending the extra three bucks. I'm just trying to add another data point or two in our ongoing attempt to understand this amazing and frustrating thing we've embarked on called digital audio.

This has always been among my favorites of the records I've worked on with David Grisman. I'm very glad to see it released in Hi-Res. It was recorded live to 1/2" two track analog tape at 30 ips on an Ampex ATR-100, mixed by the talented Dave Dennison. The tapes were in perfect condition when I got my hands on 'em, so no restoration was required. I did the transfer and mastering at 88.2/24 bit, then made a 16/44 version for the CD from the 88/24 at the very last step. Naturally I archived the 88/24 files, and then it was easy to make files for HD Tracks at that resolution, files which have never been at any lower resolution. If there was an LP I know nothing about it. There are now several Acoustic Disc albums available from HD Tracks at 88/24 or 96/24. I would also recommend the Vassar Clements album "Living With The Blues," the John Sebastian & David Grisman album "I'm Satisfied," the two David Grisman Quintet albums "Dawgnation" and Dawg's Groove"...well, you get the idea. There are even more than that, all good music, well recorded at Dawg Studio, and transferred from analog at Hi Res.

It's true that some top pro's post sometimes on Gearslutz. It's also true that some people post there who haven't got the slightest grasp on the fine distinctions between fecal matter and shinola. In that particular thread I saw none of the former, but plenty of the latter. It was obvious from their responses that most of them hadn't even bothered to look at this thread, and the one or two who had, had completely misunderstood what's being discussed here. There are yahoos anywhere you look on the web. You'll waste a lot less time, and make quicker progress toward your own understanding, if you develop a sense of whom to ignore.

"Barry, (the original poster, not Barry Diament)..." I could be wrong but I thought his name is Wayne." You're right, the confusion was on my end. BTW I'm a fan of your work as well.

Ralph, The point you repeat from Pure Vinyl's developer on "ghost play" has been spread around this forum, and is clearly aimed at Amarra, although I notice an effort to maintain plausible deniability on that point. I have tested Amarra thoroughly enough to be certain that there is no "ghost" making it to the digital outputs. The inference is totally bogus. If you like Pure Vinyl, fine. I'm interested too, and appreciate the user comments from yourself and others. However, repeating bogus slurs against another product does not serve the interests of this community well. I have no commercial interest in Amarra, and no ownership interest in the company. I am a user of Sonic products since 1990, and a beta tester for Amarra, if that affects your view of my opinions. As a secondary point, pay attention to Barry Diament's post about why and how decoding lossless files on the fly while playing them can possibly alter the sonic result, even if the data itself is identical.

Barry, (the original poster, not Barry Diament) I have a fair amount of experience doing digital audio with a G5, but before I can really comment, I should ask you to clarify one point. When you say, "I hope to achieve the best possible sound reproduction via the G5," do you mean that you are really shooting for ultimate quality playback, or do you mean you hope to achieve the best possible within some definition of what's practical? If the latter, what is the range of practicality you are considering? Using the Toslink connection, for instance, would put it pretty far away from "ultimate sound quality." Of course just about everybody has some limits to what they are willing to undertake--even people with a large budget have a budget. Many of the people on this site are looking for something in the "very practical and not too expensive" range. I often see them get into needless arguments with people on the "ultimate quality" quest, just because they are working within different sets of assumptions. So where are you on the continuum?

I don't have anything in the way of details to share, but I can tell you that he is still very actively developing it.

There's lots of good info in this thread, including most of Nyall Mellor's post, but I hate to see this common oversimplification repeated: "2) 16bit dynamic range is 96db. Most systems have max dynamic range of maybe 60db, world class ones will approach 90db. This is because a) they have a high noise floor maybe 40db and b) they have inefficient speakers with limited max SPL level" In fact we hear very far below that supposed 40 dB noise floor. To get a good understanding of dynamic range I highly recommend the AES papers by Louis Fielder, especially the first one titled "Dynamic Range Requirement For Subjectively Noise-Free Reproduction of Music," available from aes.org. There's a lot in there, so don't even consider this a synopsis, merely a taste: Fielder established that people can consistently hear down to 4 dB SPL, even in noisy rooms. Most of that effect is understandable when you look at the frequency spectrum of room noise: it's almost all low frequencies, mostly very low frequencies. That's usually the case with system electronic noise too. The noise floor at mid frequencies is way way below that 40 dB figure. Then a little of the effect probably is explained by our hearing's ability to separate correlated sound from uncorrelated noise. To bring this back on topic, I have found that having a quiet room adds to my enjoyment of music playback. But my current server runs on a G5, which is pretty noisy. That's why I keep it in the next room.

Xenophilic opines: "artificlally augment the sound" : The folks who market the Amarra software have been extremely evasive in describing what Amarra does or does not do. It's not my exclusive opinion that the sound of music played through the Amarra software exhibits a sense of expanded soundstage that is oddly similar to what can be achieved through DSP. There's no way to validate that without more openness from Amarra. I am sure that my hypothesis is reasonable, and it is easily tested, but the test is actively evaded by the software marketers." Regardless of how well Sonic has or hasn't explained what their software does, your conclusion is not supported by logic or by experiment. In my line of work we have been quite interested in knowing which products change the audio data and which do not. Using bit-comparison techniques we have satisfied ourselves that Sonic products are capable of storing, retrieving and passing data without changing bits (except of course when we want them to) for the twenty years that we have been using them. We have checked Amarra, and found that it also passes the audio without changing the bits (unless you intentionally invoke a sound-changing process.) I am not a part of the company, just a customer for twenty years.

I haven't done any serious listening to the Model 4.

The box has quite a lot of DSP power built in. If you buy it from Sonic (under either name) it gets some extra firmware that allows some of Sonic's proprietary processing, such as EQ, to be done in the box, taking the DSP load off of the host computer. The Amarra 4 and the Sonic 304 don't contain the mic preamps. They sell one which is the exact equivalent of a ULN-8, including the mic preamps, called the 305. Other than that, the only difference is in the faceplate. If you think one faceplate sounds better than the other...

We had a couple of ULN-8's in the studio for two years through the entire beta period, right next to the Pacific Microsonics Model Two's. I think the ULN-8, or the Sonic version, either one is a terrific performing piece. However I never thought it was as transparent as the Model Two. It does sound noticeably better with an external linear power supply. Don't expect a two channel version of it. Due to the economics of manufacture it would be MORE expensive to make a two-channel version just by leaving stuff out of the eight channel version. Just look at it for what it is; either it's worth it to you or it isn't. For all it does, and how well it does it, I think it's a screamin' deal.

Peter, you are discussing what happens when you go up from 16/44 in either sample rate or bit depth. But why should we assume that 16/44 has an optimal relationship between time and amplitude? Remember, that standard was chosen as a balance of commercial concerns and what was actually possible at the time, or what looked like it would be possible soon. So what is the optimal relation between sample rate and amplitude, in order that we are not wasting resolution on either axis?