Sunday night I decided it was time to play some vinyl. It had been over a month since I'd last played a vinyl record...one was still sitting on the platter waiting for side 2 to be recorded. Things are often like that here, and this time I have the excuse of the Holidays.
Cleaning the stylus with my usual device there wasn't the usual predictable noise, but instead just one coincidentally timed full power pop from my system. I checked all the levels and if anything they were lower than normal. Then there was silence when I tried to clean the stylus again.
I tried playing records, even just lowering the stylus to the record. At first I though I had killed my cartridge. Perhaps I should never have trusted a "sticky" stylus cleaner, especially one in a metal box.
Once in awhile, there WOULD be another pop, typically when "cleaning" the stylus or lowering it to the record. But then after that there would be silence (and I was amazed at how little stylus chatter there was too).
I tried re-adjusting the tracking force, rebalancing the arm. Then I tried removing the damping tape from the back of the arm where it contacts the cueing lifter. Perhaps the tape had swelled so the stylus wasn't actually making contact. I currently have difficulty seeing very well close up so I wasn't sure.
Anyway, after removing the tape in the back, of course I had to re-balance the arm and set and check the tracking force all over again.
I was in greatest fear that I had damaged the cantilever with my stylus cleaning device.
But finally it dawned on me that it probably wasn't the cartridge. With the best reading glasses I could find, the stylus looked OK, I just wasn't used to seeing it in the cartridge wrapped up in tape.
Then I suspected the preamp. I tried playing other inputs and they didn't seem to work either. Notably the Denon 9000 on the bottom.
The next morning I removed the new pile of equipment beside the rack so I could remove the connections from the preamp so I could replace it with the other XSP-1 in the bedroom which is mostly unused nowadays. I also had to clear my bench of the Adcom 535 I'd been checking out (update: for awhile, I believed my conclusion that new filter caps were needed was shown to be in error while measuring the other channel. There seemed to be only some 60 Hz noise in the distortion residual. But going back to the first channel, I could now see clearly the "distortion" was that same 60 Hz plus a larger amount of 180 Hz plus some actual 2nd and 3rd order distortion. So it's only one channel that needs the caps...good thing as I only ordered one pair which has now arrived, and I didn't want to spend even more money on this amp).
But, on the bench, I could clearly see that the Emotiva XSP-1 was working fine. Something else was wrong with my living room system when playing records.
It occurred to me that the Denon could have gone bad too. Or something else which had been nearly unthinkable...
The Lavry AD10, after 12 years of continuous use, always being left on, had died. (In future I'm tending not to leave things ON always as I think it shortens their lifespan.)
The Lavry "seemed" to work in that you could set all the settings. But level lights weren't responding to any input. Nor was there any signal being output. I tried switching the clock to Internal and that didn't help.
Currently the Lavry AD10 is the presumed latest failure, and it's on my bench for checking when I get around to it. Perhaps the problem was actually a cable connection and not the ADC after all.
So now, at least until I get the Lavry fixed (or proven working again) I had to rig up a completely different analog setup. I have to convert "analog" sources to digital to even hear them because my system depends on digital processing for EQ and crossovers.
Actually something that I had long considered worthy of trying anyway!
The output of the Emotiva is now going straight into the Tascam DA-3000, which is now dedicated to that purpose. It plays on my audio system when I press "record" (which is actually "record pause") on the Tascam. So I'm using the Tascam AD converter, rather than the Lavry AD 10 converter, to hear analog sources, and possibly record them.
The Tascam uses later generation analog to digital chips, and is itself quite high end featuring 192kHz maximum PCM and even 5.6mb DSD. (Tascam is no longer selling this in the USA???)
Though I imagine a real record label would use something even more special, like a Lavry Gold or Weiss AD converter for PCM anyway.
But the Lavry has always seemed perfect, and the peak level holding lights are perfect for finding the best recording level for any vinyl record.
Why do so few devices have peak hold lights like that??? The Tascam surely needs them.
Also the Lavry may use a slightly higher grade of earlier AD chip. It may have better AD engineering, and it may gain from being a "dedicated" AD converter as opposed to the converter section of a recorder. The Lavry "Dynamic Range" spec is a few dB's better than the AD-3000 "S/N" spec, but they are following different standards so it's not really clear which is better that way. The Lavry has even lower distortion specified at 0.0009% as compared to 0.003%. Either is as good as my preamp.
Roughly speaking, both are at the same level, which you might call something like "prosumer" or "pentultimate". So they would be worth comparing perhaps.
My initial impression is that the DA-3000 converters, running at 24/96 (which is my standard for "all that's necessary") sound a bit more transparent than the Lavry, but not as smooth and relaxing as the Lavry, and the Lavry is very transparent also (not unlike a good preamp, I've always said). So it's almost a draw but it still seems I might like the Lavry better, plus it's peak reading meters.
To continue to have a separate digital path for both playing the FM tuner as background music, AND recording special FM programs, I needed to set up my second digital recorder, something I'd wanted to do anyway. I removed my old Marantz 480 recorder from the living room system when I got the DA-3000 because the DA-3000 was clearly better in every way. I was hoping to sell the Marantz (I see they're going for as little as $280, so hardly worth bothering). Then I got the idea I could put the Marantz back up so that each recorder would have a dedicated purpose. The Marantz would be dedicated to recording FM, and the DA-3000 would be dedicated for recording vinyl. That way I wouldn't have to be constantly changing the settings (Sample rate, SRC, and input source).
So this was already turning into a two day project, with all the items (and cables!) that needed to be removed from the rack, and possibly repurposed into a new arrangement and connections.
I also needed to replace the Denon 9000, now also presumed dead (no apparent audio output) with my second unit, which (on last testing) had been able to play DVD-Audio discs. I had been saving the second unit for that purpose, AND because the first unit I got was always the one that I thought might sound best on CD's and HDCD's, even though it was unable to play DVD-Audio discs anymore, after $150 worth of repairs done by a Denon center didn't help (they said I'd start having to replace boards like the power supply board).
So, I'd wanted to exchange these units as well, which will enable me to make another sampling of the DVD-Audio disk Supernatural by Santana, which seems to sound the most magic when played through Denon DVD players, with the DVD-9000 being the ultimate king of the hill-but never recorded before by me (because I usually had my non-DVD-Audio playing Denon 9000 hooked up).
But now, before before the Second DVD-9000 dies, I also wanted to address a long lingering question about HDCD. It has been claimed by many audio engineers who have studied the matter (but not Pacific Microsonics, which no longer exists, nor Keith Johnson) that no HDCD decoder uses the "transient filter" that is toggled on many HDCD recordings. It was apparently used when the recording was being made, and one of the reasons the PM One and Two recorders were so good. But no decoder was ever set up to do that (according to these audio engineers, including the late Charlie Hansen) because it had already been patented by Ed Meitner, then working at Museatex (but not otherwise used, to my knowledge, and that company no longer exists either IIRC, so it appears the patent did nothing but prevent application of this idea where it would have been most useful).
https://www.head-fi.org/threads/hdcd-list.65414
(Ed Meitner appeared previously in my audio experience in the early 1980's. His amps were being demonstrated at a meeting of the San Diego Audio Society at the same time as the Apogee 1 ohm speakers. Someone hooked them up without authorization, and the Meitner amps fried. Someone who might have been Ed Meitner or his rep scolded someone for hooking up something wrong, I never got the details. We were supposed to have been using the Krells on the 1 ohm Apogees.)
Here is the definitive HydrogenAudio (Objectophiles) table of which discs use which HDCD features.
I retained hope for several years now that this story about HDCD was wrong. HDCD's just sounded so much magically better than other discs. I was sure it wasn't just the "peak expansion." I truly believed in usefulness of the variable Transient Filter.
And they sounded best of all on the Denon DVD-9000 player, which is the only player that has proven to me it does the "peak unlimiting" thing as PMI would have wanted. It puts out more than just 2 volts, I determined, reaching nearly 3.8v peak on some HDCD's.
Other players seem to implement Peak Unlimiting by lowering the average level, and only having peaks reach the full output level, which for most disc players is nominally 2V.
Back in the day the DVD-9000 was being sold, a friend of mine complained (either about that exact player, or perhaps another) that HDCD was a gimmick where they raised the level twice as high. Well, he said, I can just raise the level twice as high on my preamp. It's not fair!
It seemed like ultimately other manufacturers agreed, and that while some earlier HDCD players may have raised the actual output level to implement Peak Expand, others didn't, they lowered the average level to make Peak Expand possible.
So few other HDCD players do the Peak Expansion in the way that the Denon DVD-9000 does.
Anyway, as I became more and more aware of the stories regarding the Transient Filter flag not actually being used, I was still convinced it was important, because otherwise how could the Fresh Aire HDCD discs from Mannheim Steamroller sound so specially good when played through the Denon DVD-9000.
So good I was convinced that resampling through the DVD-9000 was better than playing the original CD bit stream (which needed no other decoding than the Transient Flag anyway). I was beginning to believe there was just something special about this high priced and limited production player.
The DVD-9000 literature made note of the "new generation" HDCD decoder, along with the new AL 24 Plus digital filters. IT DOES NOT HAVE A PMI 100 DECODER, as such. It is the first Denon that supports HDCD that does not actually have the PMI chip. Denon paid for those things when they were still being made. The Denon 9000 included DVD-Audio, but NOT SACD. It was Denon's last chance to display technologies alternative to SACD. The Denon manual, IIRC, notes licensing to Museatex but doesn't say why.
After just a few months, the DVD-9000 was discontinued, and replaced (for a while) by the 5900, which uses sigma delta chips and supports SACD. Sony would not apparently license DSD to Denon for the DVD-9000 which used the very different Burr Brown 1704 resistor based chips. This also may have affected what Denon was able to do for HDCD.
So it could be, I have long wondered, that Charlie Hanson and others who claim the switchable digital filter thing was never implemented were simply unaware of a small number of exceptions, if they exist, which might have included the Denon DVD-9000.
Not knowing people who would know, another way is to try it, as I've long wanted to try.
So I started recording a series of HDCD's that I know use the transient flag, but no others, on the DVD-9000 through the XSP-1 and directly into the DA-3000. These are the Fresh Aire series on HDCD from Fresh Aire 1 through Fresh Aire 8. They all use only the Transient Filter flag from HDCD and no others. (According to Charlie Hanson, these shouldn't be considered HDCD as they don't require HDCD decoding.)
The Fresh Aire recordings are not even included on the standard list of HDCD recordings, but they say HDCD on their boxes and were sold as such (by Audio Advisor at one point). They light the HDCD light if a player has it. But they do nothing but transient filter decoding which I reconfirmed with a little ffmpeg script I still have called hdcd-check:
rm junk.flac 2>/dev/null
for name in "$@"
do
ffmpeg -i "$name" -af hdcd junk.flac
rm junk.flac 2>/dev/null
done
At the bottom of the output you get lines like this for all the Fresh Aire discs:
[Parsed_hdcd_0 @ 0x7fdd67e02000] HDCD detected: yes, peak_extend: never enabled, max_gain_adj: 0.0 dB, transient_filter: detected, detectable errors: 0
Not knowing the correct level, and also believing the 9000 might benefit from some warm up time, I played several different Fresh Aire discs at increasing levels, until I zeroed in on the +0.5dB level as being correct for these, causing the highest non-clipping level. (If theses discs had used Peak Unlimiting, it would be something like 2.5dB lower.) I ultimately ended with my best recording of Fresh Aire 8, which reached a maximum level of -0.114dB with the Emotiva gain at +0.5dB. I saved this exactly as recorded AND as releveled (by Audacity) to -0.02dB for audible comparison with the original digital, which has peak level of 0dB (Audacity won't Amplify because it is already at 0dB).
Once again, as I have matched the levels ever closer, I found the differences if any increasingly hard to detect. The non-decoded HDCD played direct from the digital sounds basically the same as the output of the Denon DVD-9000 with "HDCD decoding" being resampled on the DA-3000.
While I was doing this, I still thought how much more wonderful the DVD-9000 decoding sounded, but now that I have both the original and the decoded copy I can say they are basically impossible to tell apart, so far.
Strangely, however, Roon finds a difference and applies different levels of automatic leveling (which I do on an album basis). It finds my HDCD version to require 0dB level change, but -2.7dB for the original digital.
This is probably an artifact of the re-recording somehow, but I'll have to figure it out before reading conclusions about the audibility. Unless I can control the actual playback level and ensure all versions are dBa or dBc matched to within 0.1dB, no conclusions can be made. The peak level matching I've done apparently works differently. Perhaps it's because my resampling at 96kHz realizes some of the "intersample overs" which are hidden in the original digital. And also in ways like that the Lavry might be different too. I'm finding a difference between the phono-to-CD level change using the Lavry and using the DA-3000. For some reason, the two digitizers show different relative phono to CD peaks.
Bu at this point, I'm also closer to completely believing Charlie Hanson in these regards. Even though the level matching doesn't seem to match, there is no special magic that the HDCD transient filter decoding seems to add. But I really want to test some of the earlier recordings too.
Charlie Hanson loved HDCD nonetheless because the original encoding converters (PMI One and Two) were so good, and because he believe even though the dynamic range improvement is little better than one bit at most, it's still important.
Like me, Charlie Hanson also hated DSD and MQA. But unlike me, he was a believer in some flavor of linear phase reconstruction filter (his own) instead of straightforward minimum phase. His flavor added significant ultrasonic noise, though not as much as NOS or DSD.
I don't think there is much need for amplifiers with no feedback, which was ultimately his motto.
So in spite of his acclaim, I still consider him sort of tweaky, and perhaps not fully to be trusted for that reason, but perhaps sometimes right anyway, and less connected to the big wheels than many.
An insider/outsider or something like that. So I sort of like him a lot anyway (while thinking he may have been wrong in several key areas).
I thought the idea and design of Pono were not a bad idea, and often considered buying one, but I'd have had to rig it into my system which means basically resampling... And I'm not interested in headphones. And they never got that much support from the music industry which of course really killed it.
I would have been interested in a high resolution front end device with AES connections. Ultimately I got Roon on my Oppo, at least that has coax digital, and through that I get Qobuz for high resolution which it has in some things.
It sounds like Charlie had a very interesting life and career in audio. And he was a healthy active guy and everything. But he died from the complications of a bicycle accident in 2006 after suffering from it increasingly for 11 years.
You can't tell any suffering from his online persona. He seemed to be bright and cheerfully talking about what HDCD does and doesn't do (even that, something barely connected to his business) up to his last days. I had been reading the HDCD column and was shocked when it was written that he had passed away.
It would be very funny in a way (for me) if the Transient Filter flag actually does nothing. Because I believed in it so long.
It shows the power of belief. Belief can make things special and magical, when they are just ordinary or nothing at all.
I'm suspicious this happens a lot to other audiophiles, but not me. For one thing, I don't really believe in much of anything anyway, anything outside of mainstream audio engineering, in which audio reproduction is fairly well described by its lack of distortion and frequency response errors.
So when it comes to other things, things that Objectophiles wouldn't believe are audible, I have mostly if not entirely disbelief.
This was not at all true for HDCD. I really believed in it, and that the Transient Filter was a kind of precursor to SACD that Sony wanted killed because it was nearly as effective and yet far simpler. One filter for transients and one filter the rest of the time, I thought. That way you could get the perfect transients that were a (usually unheralded) feature of SACD and DSD and NOS type filters, but just when they are needed so as not to be getting a lot of noise and grunge too.
That belief made HDCD's played on HDCD decoding players always specially magic to me. I believe I wrote about it here.
That's a huge upside perhaps from mere "belief" if it is a false belief, as it now appears.
The downside is that I was afraid to play HDCD material any other way, including on my Roon/Oppo based system which I use to listen to most music now.
HDCD's were a special category that I didn't want to listen to unless properly decoded.
So I spent hours, and more hours, always making sure that in one way or another I could properly decode HDCD, transient filter operation and all. I was convinced that software converters were inadequate because few could actually model the effect of different reconstruction filters.
And I avoided listening to my HDCD's in any other way. If the DVD-9000 was not connected as a digital converter (which is often true because the extra Sonos ZP80 I use to send it digital is often removed as being one to many things when I need more room in the rack), I'd avoid listening to any HDCD's.
So while the power of belief can be magical, it can also produce fear, the well known audiophila nervosa.
No comments:
Post a Comment