Home of the Squeezebox™ & Transporter® network music players.
Page 3 of 4 FirstFirst 1234 LastLast
Results 21 to 30 of 33
  1. #21
    Senior Member
    Join Date
    May 2015
    Location
    Grosse Pointe Woods, MI
    Posts
    775
    Quote Originally Posted by Archimago View Post
    Interesting history there Arny... I assume the effect must have been transparent.

    Also I presume there's an AD/DA in the device. What kind of conversion quality are we looking at with the ADD-1 back in the day for these LP's!?

    I'm sure Michael Fremer would have no trouble identifying these digitally compromised LPs .
    I can only talk about what I read from the dB article, what was SOTA at that time, and analysis of the measured evidence that I observe. Back in the day we had very little analog test gear that was accurate enough to make judegments about gear this good.

    It was indeed composed of a pair of discrete ADCs, probably sucessive-approximation, the digital delay line was probably based on shift registers implemented in TTL, and a pair of discrete (probably R2R) DACs. The brick wall filters were no doubt analog, probably made up of lots of little coils and capacitors. They were unlikely to have been minimum-phase or very accurate. The notch in the recording of low level signals suggests what were called Eliptical Filters. But it performed so much better than analog. For example it passed the keys-jangling test, and it took 30 ips half-track to do that in days when 15 ips hafl-track was the usual practice.

    As far as Mikey Fremer goes, I suspect that he knows little about this phase of analog recording which was already becoming hybrid digital. I've never seen where he called out a LP for being produced digitally in those days. It's just like the approximate 50% of SACDs and DVDs that were produced before 2006 that had low rez (including early digitial) artifacts baked into them, but nobody ever compliained, not even Mikey or the rest of the SP and other Audiophile pub gang.


    It is a coincidence of history that the golden ear writers and audiophiles were unanimous in their condemnation of early digital, and much later digital while happily listening to LPs that were produced digitally, often using worse digital converters than the earliest CD players.
    Last edited by arnyk; 2018-01-27 at 04:17.

  2. #22
    Senior Member pablolie's Avatar
    Join Date
    Feb 2006
    Location
    bay area, california.
    Posts
    1,101
    I remember the days when they would print -in the back of quite a few CDs- whether it was DDD, ADD or AAD. I don't think I ever saw vinyl stating whether it was "ADA" or something like that... :-D But I have to admit my vinyl days were over as soon as CDs came out, and that I digitized my vinyl collection as soon as some home digitization became possible.

    ...p
    ...pablo
    Server: Virtual Machine (on VMware Workstation 14 Pro) running Ubuntu 16.04 + LMS 7.9
    System: SB Touch -optical-> Benchmark DAC2HGC -AnalysisPlus Oval Copper XLR-> NAD M22 Power Amp -AnalysisPlus Black Mesh Oval-> Totem Element Fire
    Other Rooms: 2x SB Boom; 1x SB Radio; 1x SB Classic-> NAD D7050 -> Totem DreamCatcher + Velodyne Minivee Sub
    Computer audio: workstation -USB-> audioengine D1 -> Grado PS500e/Shure 1540

  3. #23
    Junior Member
    Join Date
    Nov 2014
    Location
    Everett, Wa
    Posts
    20
    I can tell you from experience exactly what one gets, a convenient way to buy digital material without buying and storing CDs. I actually use SOX to downsample and downbit these recordings to 16/44 using the default dithering setting which is triangulation dithering. Some of the worst earbleeding recordings I own are remastered hi res albums and some of the best sounding I have are redbook 16/44. I bought Nirvana's remastered high res Nevermind, I also own the 1990s CD. For compulsory reasons I listened at high res first as I always do since I paid for it and was not prepared for the nails on the chalkboard overly compressed with occasional clipping sound that was produced. The 1991 CD version sounds superior in every way.

    When I wanted a copy of Peter Frampton's Frampton Comes Alive, I bought the so called High res version. While it sounded as good as I have ever heard of this album and saved me from buying yet another CD, it was merely an upsampled 44.1k version. There was a brick wall at 22khz. everything above that is merely padded with zeros creating a huge file. I'm amazed at people swearing that upsampling something that was originally sampled at 44.1k can somehow add information that was never there to their bigger file. Sampling is done in the pre-digitization process and can never be raised above what it was originally sampled at, only lowered. See attached image and notice the brick wall at 22khz, this is a 96k file!

    I will tell you that many recordings DO sound better due to more careful remastering by someone who actually can hear but when I convert them to Redbook, they retain every bit of the improvement. I keep the original hi res downloads in an archive folder and all my music files are on a fault tolerant raid server using one of the few filesystems that protect against bit rot, a COW (copy on write) filesystem that does self checksumming and can restore itself due to the copy on write data. To give a visual example of what bit rot can do, think of when you visit a website and only part of an image shows up and the rest is grayed out, that my friends is bit rot of an image file. You do not want it EVER on any file.

    Trying to reproduce non musical noise above 22khz if your speakers and amp etc could even reproduce it if anything might be cruel to certain pets you may have, there may be some musical information there but it's effects within our hearing range have already been captured once. No human in the last 100 years has ever heard very far beyond 20k just as none of us can see infrared and ultraviolet. Do certain lower harmonics of certain instruments exist in the audible range, sure they do but those lower harmonics we can hear in our audible range were already captured by the recording equipment as I touched on above, trying to reproduce it twice, once before the recording and once again on playback is not what I would consider proper reproduction. In many cases, your tweeter as well as amplifier, preamp and such have a good chance of creating problems above 20khz since they are designed to behave in the audible range and are not designed to double as an RF amp or transducer in the case of the tweeter. Throughout my life, the poor harsh quality of some recordings has always been in the upper midrange and in my late teens as well as now, it is the same and can drive me up a wall while cursing the deaf engineer who subjected us to that. Some of the best tweeters on the market that I would unhesitatingly use in my speaker designs, have a breakup node well above 20khz to the point that I would use a notch filter outside our hearing range just to kill it unless it somehow keeps pests away without bothering me or pets. There are benefits to rolling off sharply above 20khz. Many are fooled by the engineering representation of a sampled sine wave that shows a stair step effect, this effect does not exist in reproduction. It has never been seen on a scope or otherwise. The assumption is that the higher sampling will make this non existent stair step effect more fine grained and thus closer to an analog sine wave, nope, one only needs the sampling to be twice that of the desired top frequency to reproduce a perfect waveform. Digital audio was worked on for nearly a century before we got it.

    As far as 24bit depth, well it harms nothing but a dithered 16 bit recording can easily produce a clean -105db signal. 96db is not the limit on these recordings, even at that, the worst sounding digital recordings are usually compressed to have about 20db dynamic range at the top of the scale and are are overdriven above 0db producing clipping. You don't need even four bits of depth to produce some of these awful recordings. With digital, distortion is extremely low below 0db but increases horribly above 0db. It is not like the analog days when it was common and preferred to have an occasional signal hit +3 or +6db on peaks only since good tape had headroom and not much signal to noise so you had to fully utilize what you had. the heavy compressed and marginal overdriving effect used on far too many modern recordings are a relic of the perceived loudness effect some think still sells even though it has not worked since AM radio. In the late 70s FM Stations began trying this, it just sounded awful, not louder as it was with AM and now people are trying this with digital, enough already! High sampling and larger bit depth will solve none of the ills created by this technique, just bigger file size.

    I know this is long and sounding like a rant. I will say that I did find some places to download redbook quality FLAC files for certain artists such as dead Can Dance. I finished out my collection of their music only to see that whoever converted them to FLAC created files that have tons of clipping. My three CDs from the 1990s of their material when ripped by me to never go above -1db in peaks have none of these problems. There just is no need for something that can produce over 100db of dynamic range to have clipping in it nor is there an excuse. The attached analysis shows The Fatal Impact from the eponymous Dead Can Dance album that I purchased as a download. The red sections are where the clipping occurred. Dead Can Dance recordings usually are recordings worthy of judging equipment by, not these downloads who were converted by someone who does not know or does not care what they are doing, possibly both.

    Of note is that the 1999 remaster of Roxy Music's Avalon is agreed on by many as the best sounding version of this beautiful album, may have to buy as CD and rip yourself. All the High Res downloads mentioned here by me are from HD Tracks and is likely the only practical way to get a remastered lossless download even though most are High Res. Many do sound great because of the remastering itself. Another side note is that analog LPs converted to CD sound exactly like the LP complete with the pre post groove echo that gives the false impression of more depth and air as well as the audible effects of wow, flutter, rumble, tangential tracking errors, skating, surface noise, tonearm microphonics and hum if audible not to mention the higher distortion of the inner tracks of the LP due to less molecules of vinyl per second passing the needle and least of all, don't forget the only 45db of stereo separation and 8 bit equivalent dynamic range at best. I lived that for 25 years of my life and miss none of that and certainly do not miss playing one side of one record at a time. Heck, I don't even miss digging CDs out of their jewel cases. I started moving to server storage around 1999 when Flac was well on it's way, then the iPod completely derailed the progress being made on the server/streamer method that we thankfully and sanely resumed. Seriously, did we need a 3000 dollar docking station for the lossy mediocre sound from these devices? What were people thinking? I DO miss the huge 12" album covers in my hand while dreaming of being a rock star, the latter really added to the experience. Apple ruined that too with the iPod, bastards! they really have the marketing savvy to dumb down the public and convince them that the sows ear they bought really is silk. Who else could make a fashion statement from polycarbonate which is all the iPod really was?

    Top Image: Spectrum of 96k download Frampton Comes Alive obviously originally sampled at 44k. Waste of file space.
    Bottom Image: Dead Can Dance DB showing clipping.
    Name:  Screenshot_20180205_183831.jpg
Views: 107
Size:  198.4 KBName:  Screenshot_20180205_161958.png
Views: 107
Size:  147.1 KB

  4. #24
    Junior Member
    Join Date
    Nov 2014
    Location
    Everett, Wa
    Posts
    20
    Quote Originally Posted by pablolie View Post
    I remember the days when they would print -in the back of quite a few CDs- whether it was DDD, ADD or AAD. I don't think I ever saw vinyl stating whether it was "ADA" or something like that... :-D But I have to admit my vinyl days were over as soon as CDs came out, and that I digitized my vinyl collection as soon as some home digitization became possible.

    ...p
    Some of the best sounding vinyl I had was digitally mastered. Most new vinyl is made from CDs but nobody ever talks about what amplifier runs the cutting heads. In the old days the best vinyl was cut using McIntosh 250 watt tube amps. Still, like you if there is no digital version I would at least digitize the vinyl which is better than nothing. I preferred the carefully mastered high quality DDD cds. At first they simply threw the analog masters that were mastered for vinyl onto CD and they sounded bright and harsh. Few people can appreciate just how well pure digital can sound if done properly. Some still think digital is not continuous and analog is. There are no gaps. There was an episode of Ghost whisperer where they propped up this myth by giving the boy a turntable and he said, "wow, it does sound better". The mom of the character said, "that's because it's continuous". I almost screamed and wrote an angry letter to the writer. I'm amazed at how misunderstood digital audio is by even people who should know better by their background. By the same token I would never return to analog TV or video. The codecs are just getting too good and efficient.

  5. #25
    Senior Member ralphpnj's Avatar
    Join Date
    Mar 2007
    Location
    Northern NJ, USA
    Posts
    2,646
    Quote Originally Posted by Davesworld View Post
    ....I know this is long and sounding like a rant....
    Well I for one enjoyed every bit of your wonderful post. You covered all of the points about digital audio that are most often misrepresented and backed up your statements with good solid scientific proof. Unfortunately science in the Age of Trump is down on one knee and hanging onto the ropes for dear life.

    One little thing that you missed is the inability of humans to hear nano and pico second distortions, e.g. digital jitter.
    Living Rm: Transporter-SimAudio pre/power amps-Vandersteen 3A Sign. & sub
    Home Theater: Touch-Marantz HTR-Energy Veritas 2.1 & Linn sub
    Computer Rm: Touch-Headroom Desktop w/DAC-Aragon amp-Energy Veritas 2.1 & Energy sub
    Bedroom: Touch-HR Desktop w/DAC-Audio Refinement amp-Energy Veritas 2.0
    Guest Rm: Duet-Sony soundbar
    Garage: SB3-JVC compact system
    Controls: iPeng; SB Controller; Moose & Muso
    Server: LMS 7.9 on dedicated windows 10 computer w/2 Drobos
    Last.fm

  6. #26
    Hi,

    A very interesting read, thank you to everyone who's clearly put time and thought into the research.

    A couple of questions from my rather naive (and very computer science orientated background). I ripped my CD collection to FLAC many years ago using Max on a Mac (http://sbooth.org/Max/). I can't remember the exact settings, but would have probably been very close to the defaults as shipped with the application. My understanding being the software digitally extracted the 1s and 0s from the CD and then converted these to a FLAC file. In such a setup, how can you produce a bad FLAC version with clipping not present on the original as mentioned by Davesworld?

    Secondly, where do people recomend downloading lossless files from if not HDTracks? Is 7digital regarded as a better source (I'm mainly thinking about new releases as opposed to old recordings re-issued for the digital age).

    Cheers
    LMS Server: Mac Mini with music library on Drobo 5D
    Living Room: Raspberry Pi + AlloBoss -> Onkyo A 9010 -> KEF LS50
    Study: Raspberry Pi -> Audio Engine D1 -> Audio Addon Pro T3
    Kitchen: Squeezebox Radio
    Bedroom: SB3 -> B&W Zeppelin

  7. #27
    Junior Member
    Join Date
    Nov 2014
    Location
    Everett, Wa
    Posts
    20
    Quote Originally Posted by odw199 View Post
    Hi,

    A very interesting read, thank you to everyone who's clearly put time and thought into the research.

    A couple of questions from my rather naive (and very computer science orientated background). I ripped my CD collection to FLAC many years ago using Max on a Mac (http://sbooth.org/Max/). I can't remember the exact settings, but would have probably been very close to the defaults as shipped with the application. My understanding being the software digitally extracted the 1s and 0s from the CD and then converted these to a FLAC file. In such a setup, how can you produce a bad FLAC version with clipping not present on the original as mentioned by Davesworld?

    Secondly, where do people recomend downloading lossless files from if not HDTracks? Is 7digital regarded as a better source (I'm mainly thinking about new releases as opposed to old recordings re-issued for the digital age).

    Cheers
    I've not purchased from 7digital yet but they do have some material I am interested in. I am reminded that I need to occasionally seek out which lossless downloads are available from where.

  8. #28
    Junior Member
    Join Date
    Nov 2014
    Location
    Everett, Wa
    Posts
    20
    Quote Originally Posted by ralphpnj View Post
    Well I for one enjoyed every bit of your wonderful post. You covered all of the points about digital audio that are most often misrepresented and backed up your statements with good solid scientific proof. Unfortunately science in the Age of Trump is down on one knee and hanging onto the ropes for dear life.

    One little thing that you missed is the inability of humans to hear nano and pico second distortions, e.g. digital jitter.
    I hadn't considered that but I will say that a networked transport is likely to fare very good compared to a disc transport in this regard. Any good DAC should be responsible for clocking accuracy in it's end these days making it a moot point. I do prefer jitter to be under 10ms though given a choice. I'm sure someone has done tests where they induce varying degrees of jitter while test subjects listen. It would be neat to test DACs buffering and re-clocking ability by throwing a heavily jittered stream into it. Some say only re-clocking on the DAC end is necessary without buffering first. If DACs aren't doing this they should be to put the notion to rest. It takes hours if not days of listening to something to know if fatigue sets in. This is especially true with speakers, at first comparison the one with more harmonic distortion will give the impression of more detail but this new found detail is not supposed to be there. Third harmonics are the most annoying. Some like the coloration of even harmonic distortion and you would have to wait for them to die before pulling the single ended triode amp for their hands.

  9. #29
    Senior Member
    Join Date
    Apr 2005
    Location
    Watford, UK
    Posts
    1,850
    Quote Originally Posted by Davesworld View Post
    I know this is long and sounding like a rant.
    Well - yes, it does sound like a rant.
    And while most of your points are basically true, it seems a bit like a stream of consciousness diatribe.
    To a non-believer, it will sound like a bunch of opinions stated as fact, and that's not going to convince anyone.

    I'd like to address a few points.

    Quote Originally Posted by Davesworld View Post
    Many are fooled by the engineering representation of a sampled sine wave that shows a stair step effect, this effect does not exist in reproduction. It has never been seen on a scope or otherwise.
    For an in-depth explanation of the reasons why (and lots of other digital audio fundamental truths), see Monty's excellent video: https://xiph.org/video/vid2.shtml - this is the sort of hard evidence you need to present to people who believe otherwise.

    Quote Originally Posted by Davesworld View Post
    The Fatal Impact from the eponymous Dead Can Dance album that I purchased as a download. The red sections are where the clipping occurred.
    Your screen shot showing the clipping is from Audacity, which is known to have a flawed clipping detection algorithm. PCM samples of N bits range in value from -(2^(n-1)) to +(2^(n-1)-1). A solitary sample at -(2^(n-1)) is NOT necessarily a clip, but Audacity will tell you it is. I'm not saying that the Dead Can Dance file doesn't have clipping, but the Audacity screen shot is not proof that it does.

    Quote Originally Posted by Davesworld View Post
    I do prefer jitter to be under 10ms though given a choice.
    I sincerely hope that is a typo and you meant to say 10ns!

    Quote Originally Posted by Davesworld View Post
    I'm sure someone has done tests where they induce varying degrees of jitter while test subjects listen.
    Indeed they have:

    Benjamin & Gannon.
    Theoretical and audible effects of jitter on digital audio quality.
    105th AES Convention, 1998
    Jitter added to digital signal between transport and DAC with a hardware device.
    Conclusions: uncorrelated jitter inaudible below 10nS rms on pure tones; uncorrelated jitter inaudible below 20nS rms on music signal

    Ashihara, Kiryu et al.
    Detection threshold for distortions due to jitter on digital audio.
    Acoust. Sci. & Tech. 26, 1 (2005)
    Jitter simulated in the digital domain.
    Conclusions: uncorrelated jitter inaudible below 250nS on music signal.
    Transporter -> ATC SCM100A

  10. #30
    Junior Member
    Join Date
    Nov 2014
    Location
    Everett, Wa
    Posts
    20
    Quote Originally Posted by cliveb View Post
    Well - yes, it does sound like a rant.
    And while most of your points are basically true, it seems a bit like a stream of consciousness diatribe.
    To a non-believer, it will sound like a bunch of opinions stated as fact, and that's not going to convince anyone.
    Fair enough. I'll pay close attention to how I present information.


    Quote Originally Posted by cliveb View Post
    For an in-depth explanation of the reasons why (and lots of other digital audio fundamental truths), see Monty's excellent video: https://xiph.org/video/vid2.shtml - this is the sort of hard evidence you need to present to people who believe otherwise.
    I've seen this but should have linked it since it clears up many misunderstandings and myths.

    Quote Originally Posted by cliveb View Post
    Your screen shot showing the clipping is from Audacity, which is known to have a flawed clipping detection algorithm. PCM samples of N bits range in value from -(2^(n-1)) to +(2^(n-1)-1). A solitary sample at -(2^(n-1)) is NOT necessarily a clip, but Audacity will tell you it is. I'm not saying that the Dead Can Dance file doesn't have clipping, but the Audacity screen shot is not proof that it does.
    Yes, I just tried that same file using Ocenaudio and while it does not give a visual it did list dozens of clipped areas, over 40 on the right channel but I am open to more tools as I find them.

    Quote Originally Posted by cliveb View Post
    I sincerely hope that is a typo and you meant to say 10ns!
    Maybe should have but in the last ten years I have spent a good deal of time with VOIP and what affects MOS (Mean Opinion Score) and it's not critiqued in the ns, most equipment has an adaptive jitter buffer which will add delay if it becomes large since you are buffering 20ms or more in extreme cases, only the best trained listeners can notice a delay under 200ms though but this is for another topic in another forum.

    Quote Originally Posted by cliveb View Post
    Indeed they have:

    Benjamin & Gannon.
    Theoretical and audible effects of jitter on digital audio quality.
    105th AES Convention, 1998
    Jitter added to digital signal between transport and DAC with a hardware device.
    Conclusions: uncorrelated jitter inaudible below 10nS rms on pure tones; uncorrelated jitter inaudible below 20nS rms on music signal

    Ashihara, Kiryu et al.
    Detection threshold for distortions due to jitter on digital audio.
    Acoust. Sci. & Tech. 26, 1 (2005)
    Jitter simulated in the digital domain.
    Conclusions: uncorrelated jitter inaudible below 250nS on music signal.
    Ok now I see why you thought nanoseconds should have been mentioned by me, this is very interesting indeed! Thanks!
    Last edited by Davesworld; 2018-02-07 at 17:18.

Posting Permissions

  • You may not post new threads
  • You may not post replies
  • You may not post attachments
  • You may not edit your posts
  •