Hi-Rez Downloads....

[cradeldorf]

The only thing I listen to is music videos I download from youtube. Feel free to lambase me if you must but I find them plesant to listen to for the most part, yes there are some crappy sounding ones but a lot of them have nice analog sound.

Edited October 20, 2015 by cradeldorf

[artto]

 

There are usually, especially in the analog/LP days, different "masters" used for the LP that have RIAA correction. Many of the early CDs were made with these "masters" which why they sound so thin & tinny.

 

This is a myth...in my experience, apparently to cover for some exceptionally bad mastering EQ practices that occurred in the 1980s on CDs.

 

Of all the CDs that I've remastered to date, I've not found one that actually had a problem with incorrect or "forgotten" encoding RIAA curve still left on the CD. One would easily recognize the presence of the RIAA curve: it's 40 dB end-to-end. 

 

 

Chris

 

 

It is NOT a myth!!

 

Just the record, I am the first person, guaranteed by Sony, to take delivery of a Sony CDP-101 player and the 14 original CD releases in the U.S. In fact, I had a CD-101 and a few (European) CD available for use in my system over a weekend for a photo shoot for Stereo Review Magazine months before the player & CDs were released.

 

Those first 14 CD sounded awful. The most offensive one was Billy Joel The Nylon Curtain. After 5-10 minutes of playing at 75-80dB my ears would be ringing it was so brash. Some time later producer Phil Ramone came across an article about CD and how they used the "original masters". He said he wondered how that could be since he had the original Billy Joel master in his personal vault and no one had asked him about using it. Sony/CBS had apparently used the LP master, sort of. Actually they made a copy of it, sent that to their L.A. facilities where it was supposed to go to Japan. Well, they made a copy of it and sent that to Japan where they too made yet another copy and used that for mastering the CD.

 

This kind of CD "mastering" went on for quite some time until the manufacturers finally figured out what was wrong. That is one of the reasons why we see so many "remastered" CD of the "same" recording (not even audiophile re-releases). CBS went so far as to completely re-catalog and even change graphics on the re-released iniital 14 CD.

 

My current amplifier will only allow me to boost bass +10dB. On those early CD +10dB is not nearly enough. And cutting the treble by the same amount doesn't make sound quite right either.

 

And then of course there is the issue of "tweaking". Anyone who has been involved with recording/mastering for a few decades, especially going back to the LP days, will know that the mastering engineer/producer will "push" the treble and/or midrange on the LP master to compensate for the losses and deficiencies inherent in LP and LP playback. So even without any RIAA curve being involved, we still have something that can be quite exaggerated when used for a flat frequency response medium like CD.

 

The only "original master" is not something that is usually used for LP/CD/tape production of the commercial release. The "original master" is just that, the original recording and it is something that has all the information that the artist/producer wants available for the "recording". After that another 'mix down" (or many mix downs) usually occurs where things are left out, balance adjusted, etc. Believe me, you wouldn't really want to hear the "original" master.

 

That being said, I seriously doubt Hi-rez download sites like HDTracks are remastering anything, and certainly not using "original" master recordings. That's just some crap that came out of the marketing department.

Edited October 20, 2015 by artto

[Travis In Austin]

It is well documented that in early days of cd that many, many cds were mastered from the "vinyl master tape" which is the tape created from the outputs of amps that go to the cutting heads. There tapes were made so that if a new master had to be cut, which was frequent, they had an exact cooy to cut another master.

It was frequently the only master quickly available in the mad rush to get back catalog on CD.

The lables with multiple pressing plants all did this, Capitol, Atlantic, Columbia, etc.

Travis

[Chris A]

This kind of CD "mastering" went on for quite some time until the manufacturers finally figured out what was wrong.

 

I assume that you're talking about this particular Billy Joel CD disc number: http://www.discogs.com/Billy-Joel-The-Nylon-Curtain/release/5614291.

 

I still haven't found a single instance of RIAA encoding curve retention in a CD to date. Listening to a CD with an RIAA encoding curve impressed into it would be so bad to listen to that it would be impossible to listen to for more than a few seconds--in my experience of hearing records played back without a phono preamp to do RIAA decoding EQ.  The RIAA curve's 40 dB equalization width from 20 Hz to 20 kHz is extreme. 

 

What I've actually found is that this handful of the earliest CD releases from perhaps one manufacturer (Columbia) is used to make excuses to cover up for the really bad mastering jobs done over the remainder of the 1980s.  I find it to be a poor bet that a mix up of mastering sources (RIAA lathe copies) is the cause of poor sound in any one CD that most anyone would likely have in their possession.  That's the mythology that I was referring to.  My findings to date actually contradict that notion. 

 

One of the best CDs that I own or have heard is the 1982 release from Elektra-Musician of Tom Scott's Desire.  This CD was produced before the big 1983 market push and is still one of my favorite test discs.

 

The good thing is that all the bad mastering EQ can be reversed without audible artifact.  The fact that all records have RIAA encoding and decoding, and it still sounds okay, is something of a testimonial to the ability of EQ being used to correct the really terrible mastering jobs that many CDs have. 

 

YMMV.

 

Chris

Edited October 21, 2015 by Chris A

[Travis In Austin]

Ok, here is the process.

Multitrack tape for tracking instruments and vocals. Generally speaking, two tracks per instrument, and two tracks per boice. That alows person mixint the tape to pan intruments and voices left and right. Limiting and compression can, and if often, utilized as part of tracking. EQ is in the entire tracking chain. Microphones are selected for their sound, so are amps for electric instruments. Sometimes no amp/microphone is used and they go DI because of that sound. High pass filters are used at various points for each instrument with an 8 to 12 db slope. They typically use anout 40 Hz for an electric bass. This eliminates some are all of the line noise if done properly, along with an inaudible noise floor.

Particular studios are used because of their sound (room EQ). Mixing boards have a signature sound, or at least perceived that way by artists. Famous boards have been bought from old studios and placed in new studios.

The multitrack tape is the Original Multitrack tape. Relatively very few still exist. There is, usually, one multitrack tape for each song. After a band becomes big, they start hanging onto those multitrack tapes and they sit in vaults. Most are lost forever. If an original multitrack tape exists it can be remixed, if it doesn't exist, it cannot be remixed.

For stereo, that multitrack tape is mixed down to a two-track "mix down" tape. In reality it is actually mixed down to an intermediate mix down tape, where it can be copied onto the "final mix down" tape. Compression, mixing and EQ can be added at the mix down level. The EQ can be active, right from the mixing board, and/or passive by what preamps are used, what recorder is used, and what tape is used.

The individual song tracks on the final mix down tape are"raw", they end abruptly, they are not sound level balanced, they are not in the order they will necessarily be on the finished tape. They typically have a 1K test tone at the beginning of each track. In a lot of cases there was a primary recorder and a second recorder was going which was recording the backup "safety tape."

Those 2 track final mix down tapes, usually 1 for each side of an lp, which now have at least two levels of limiting, compression and EQ is sent for "mastering" either internally, or to an external mastering facility.

That 2 track mix down tape then goes through the "mastering" process which, at a minimum, includes sequencing of the songs, making them have a uniform loudness level so that so you should not have to turn some songs up, and some down as you listen, fade outs, or abrupt endings. That process is done in all cases. This can be done by recording to a seperate tape, or by cutting and splicing, or both. Each song is then evaluated for what it needs in terms of EQ, limiting or compression. If it is going onto vinyl that can be done in one or two steps. They can apply any of those things and record it to another tape, or they can apply those things, "on the fly" as they are cutting the lacquer with the cutting lathe.

Whichever they do it, they put that tape on a repro deck and run the outputs into a preamp with RIAA EQ (which is referred to as "de-emphasis", not encoding or decoding) and that goes into the amps for the cutting head. There are also two outputs from those amps that were hooked up to a recorder that recorded the signal coming out of those amps. This was done at all of the larger studios. You can only make so many stampers from that original lacquer. If the lp was a huge hit, like multi-platinum, they can cut another lacquer from that tape and pressings will sound relatively uniform regardless of which plant it was pressed in, or which lacquer was used to make the stamper.

The tape that was made from the cutring head amps is typically called the "EQ Limited Master Tape." This tape was almost always retained by the record company so that could be ready to pump out more albums quickly if they needed to. The 2 track mix down tape that left the studio was returned in accordance with the directions on where to return it, if they thought to include instructions. If not, it could end up anywhere, and if there were instructions, they could still end up anywhere. There are numerous stories about whete 2 track mix down tapes have been "discovered."

In the early days of CD, they needed the mix down tapes, and they needed them fast to pump out product. Remember the AAD designation on those early CDs? Well frequently the only tape available was the EQ Limited Master Tapenwith the RIAA EQ.

I have never heard of a cd being released that came from an RIAA EQ Master Tape without at least some attempt to boost the bass and cut back the treble. I am sure that is possible. However, there are many articles written about the early CD mastering guys at the major labels, Atlantic in particular, where they had equipment that was limited in how much bass they could put back in, along with other technical issues they had to confront.

I have seen one or two of Chris' Audicty pre-eq graphs where it looked to me that it tracked a RIAA EQ, but to a much, much lesser extent, like the emphasis to the bass being applied was 5 db too short to get it back to "normal."

"Remastered from the original Master Tapes" is label hype. The "master tapes" can be any of the 2 track tapes that are created in the process from the 2 track mix down tape forward. They tend to specify now which tapes they used to remaster. Beatles remastered, the ones from about 2010 had a lot of documentation about the process. What tapes, the A/D process, that some stereo mixes were lost forever, etc.

This process is pretty much the same for CD except instead of going through RIAA EQ it had to go from tape onto a umatic video recorder that had the A/D converter, and that was sent to have a glass madter made from which the CDs were produced.

Travis

[Chris A]

Thanks Travis. 

 

The processes that you describe are not very far from what I imagined, believe it or not.  The whole industry was (and is, to some extent via retention of old processes) geared for vinyl until 1982-83 ushered in the CD (44.1 kHz, 16 bit).  It still took a number of years (like 5-7) for all recording studios and mastering shops to convert to digital.  In that time, records were still the main manufacturing targeted process, it seems. 

 

Seeing how much noise and real damage to the original music tracks occurs to get the music onto the records in order to keep playback needles from jumping out of the grooves or print-through from obliterating quiet portions of the grooves from sudden following sforzandos--is a Rube Goldberg process to this engineer's eyes.  It's amazing to me that it worked as well as it did--and never worked very well in terms of fidelity.

 

It also illuminates why digital (in the form of CDs) took over the marketplace so rapidly from 1982-1990.  CD production facilities are cleanrooms compared to the dirty, low-tech "corner car repair" record plants. The change in required infrastructure investment however, didn't dissuade the industry from responding to consumer demand for a better product.  It's not difficult to see why.  A chart showing the plummeting sales for LPs, and by 1992, CD production had achieved a volume sales replacement of LPs, and by 2007 digital download sales had replaced CDs. (In the table below, "MC" = "music cassettes".)

 

 

Also note the decrease in fidelity of records and CDs over that same period, as measured by declining average and increasing standard deviation of album dynamic range:

 

[Travis In Austin]

Here is an article, I am not sure I have seen it before on Klipsch forums, but it sounds familiar. It explains why CD quality is enough, and going up can actually degrade quality.

I can only understand about every other point he is trying to make (if that) and so I cannot judge whether there is any merit to his arguments or not. The author is well respected by his peers, that much I could find.

https://xiph.org/~xiphmont/demo/neil-young.html

Chris and Arrto have both posted that they find that sound quality for them is derived more significantly through either remastering, or all digital signal path (DSD using HDMI connection) rather than the bit depth or sampling rate. That article support those impressions, if I am reading it right.

[Travis In Austin]

Another article from C/Net, talks more about pro/con of High Res.

http://www.cnet.com/news/sound-bite-despite-ponos-promise-experts-pan-hd-audio/

[Chris A]

There are many good points made in the first article, and the second link you provided is a basically a rehash of the first.

 

I believe that most engineers, like the author of the first linked article above are particular victims of the "Streetlight Effect".  This subject area is actually a well-known issue in my particular profession (system design/complexity and how people make decisions).  There is a tendency to "know" some effect, property, dysfunction or limitation in a particular technical area, and to extend that knowledge by  inductive logic to conclude that, "well, I know this effect, so it must be applicable to every case that occurs". 

 

The fallacy, of course, is that there are special cases and more complex properties of every technical domain whose effects we haven't identified yet that makes that inductive logic by the engineer/physicist invalid. 

 

In the case of the bit depth, I believe that there is something to the quanta size--the size of the bits converted into sound on the musical decays--that is audible, but that has nothing to do with threshold noise level--dynamic range.  Higher bit depth sounds more "solid" on decays, but there is a limit, and for me, that limit is somewhere less than 24 bits.  [This effect is likely wrapped up with the music recording, mixing, and editing processes, which typically introduces clipping, but deeper bit depth allows more dynamic range into which the recording engineers can fit their raw recordings before mastering. However, there is something to say about the sound of higher re-sampled analog masters that may contradict that notion.]

 

In similar fashion, there may be issues with the author's coverage of all measurable human perception when talking about sample rates, but I feel most of these effects are due to the recording, mixing, mastering, and music file production processes, and really have little to do with the flat out "can you hear differences in music produced with higher sampling rates?".  Unfortunately, it seems like everyone is focused on sampling rate and not on bit depth, i.e., another example of the "streetlight effect".

 

The issues having to do with the repeating/folded lower sidebands on ultrasonic modulation of 192 kHz or even 96 kHz PCM files with closely-spaced (in frequency) high amplitude noise spectra--which is more like FM distortion than AM distortion that is implied in the "nonlinearities" discussion--that are real, but one way to mitigate this is the use of effective anti-aliasing filters on the recording side, and paying a little attention to the nonlinearities above 20 kHz down the line--which apparently no one is really paying much attention to right now. 

 

This is particularly true of DSD, i.e., "SACD" files, that have a great deal of HF ultrasonic noise embedded into them at the 1-bit delta-sigma ADC step.  DSD is actually inferior to PCM in every way but one: up til recently, there were no DSD editing tools available to "remaster" the EQ and compression of the recorded DSD files prior to issuing them on SACD, so they sound much better than the accompanying PCM files on hybrid SACDs which are, as a strong rule, edited.  Nature abhors a vacuum, and recording executives can't keep their fingers off the merchandise when they should.

 

I have quite a few anecdotes of these type of effects--the streetlight effect and others in terms of decision making that come to mind. One in particular is the audibility of loudspeaker modulation distortion.  Geddes, Allison and Villchur have all written articles on the inaudibility of it--Allison and Villchur in the 1970s, and Geddes in the 2000s.  PWK also wrote about it, quantifying at % modulation distortion level at which it is audible.  Allison and Villchur ignored PWK's papers or pooh-poohed them, and Geddes didn't even acknowledge PWK's papers.  Later, when presented more information, Geddes declared a mea culpa that he couldn't test his subjects using sound louder than 80 dB (which is a major factor in hearing modulation distortion in humans as it turns out).  Now Geddes says that it "may" be audible, but he doesn't want to redesign his Summa Cum Laude loudspeakers to have horn-loaded bass bins because "it's too hard". 

 

YMMV.

 

Chris

Edited October 22, 2015 by Chris A

[MyOwn]

This is very interesting

High-Resolution Audio Demystified

[bracurrie]

This is very interesting

High-Resolution Audio Demystified

Very interesting.  Going to have to replay it again to understand, but there are a lot of my beliefs that are going to be rethought after this.

Thanks for posting Steve.

[Wolfbane]

Quote from the Cnet article mentioned above: "The problem is taking the HD audio philosophy too far, as some with HD audio gear and music to sell do. "To the extent they're selling high-quality, well mastered recordings, they're not silly," Xiph's Montgomery said. "When they attribute the superiority of anything they're selling to the higher resolution, that is indeed silly"."

 

I purchased the HiRes DSD version of the Nat King Cole Story for $44 from an online web site. The SACD version of this album runs ~$60 to $100+ new. Does it sound better than any of my Nat King Cole albums in conventional CD format? Hell yes!!

 

Will someone else hear a difference? Who knows.   

[Chris A]

He's the only guy that I know of that is actively campaigning the music recording industry to "get honest" about the subject.  Right now, I believe that you need hip waders whenever you read and hear any information from anyone else about "hi rez".  Mark Waldrep is the only guy that I'd recommend listening to on this subject.  I recommend reading on his blogs, starting pretty much at the beginning and progressing at least through the first year or so.  You'll learn what you need to know about "better-recorded music downloads".

 

But I have to add a little caveat to that recommendation: unless you have a listening room that is anechoic to exterior ambient noise (i.e., its noise floor is like a good recording studio) you'll be hard pressed to hear the dynamic range arguments that even he makes. 

 

When it comes to bit depth and sound decays, this is where the greater number of bits per digital word is going to show up as a difference in listening to tracks in my experience--but that's basically it.  I'm talking about music that was originally recorded in 24/96 or 24/192 format, or its equivalent in DSD format--not older analog masters that are re-sampled up to 24/96 or higher, as he mentions.

 

It's been my experience that 96 kHz is more than sufficient in terms of reproduced frequency bandwidth--any greater sampling rates are basically inaudible in my experience on my setup, in fact 192 kHz starts to create its own issues technically.

 

YMMV

 

Chris

Edited March 25, 2016 by Chris A

[derrickdj1]

A lot of this comfirms a listening test I did between HD music file and good to excellent mp3's.  I conclude that many of the HD files sounded no better than a well done mp3 at 256 or 320.

[MyOwn]

I got very lucky as I did not invest in LP's or CD's back in my earlier days. If I paid a premium for my current downloads than at least it wasn't a double whammy for nothing....Going forward I will find a used CD, rip it to AIFF and be done...Chris,  I wish I had the time to do the things you are doing :)

[busht4169]

Hey all, I am late to the debate.  Lots of good tips in here.  May I recommend Native DSD.  One must have a pretty good DAC in order to use most of the DSD/DXD content, but they are committed to preserving quality.  Not using methods to enhance something.  We all have good ideas I think.  Perhaps someone other than me has used this

 

nativedsd.com

[Chris A]

Chris, I wish I had the time to do the things you are doing

 

Me, too. Time passes fast.

 

Chris

[Rhoddog]

What do you guys think about tidal ?

[Chris A]

I haven't heard it-- but the audio formats look pretty good: https://support.tidal.com/hc/en-us/articles/201314741-How-good-is-the-sound-quality-on-TIDAL-

 

When you compare the Tidal second- and third-level quality to YouTube audio,  http://www.h3xed.com/web-and-internet/youtube-audio-quality-bitrate-240p-360p-480p-720p-1080p, you have better quality.  If you only want stereo, this looks to be a good alternative if you don't mind the monthly price for stereo only vs. something like Netflix, which brings 100x more bandwidth with video, but even less bitrate than Tidal on audio (64 or 96 kbps) which I can tell you is a big issue with Netflix streaming (that almost no one is talking about). 

 

I'd like the ability to get quality 5.1 audio on an audio streaming service, however.  The difference is pretty big.

 

Chris

[Wolfbane]

Hey all, I am late to the debate.  Lots of good tips in here.  May I recommend Native DSD.  One must have a pretty good DAC in order to use most of the DSD/DXD content, but they are committed to preserving quality.  Not using methods to enhance something.  We all have good ideas I think.  Perhaps someone other than me has used this

 

nativedsd.com

 

A number of us have DAC's with Hi-Rez capability and have down loaded DSD/DXD files. In my experience some of the available downloads sound great, some are no better than the non-SACD DVD's they already own.