Forgive me Computeraudiophiles, for I have sinned

[vl]

I come in late on this topic.  I would like to share with you some of my experience.

 

When I started with HiFi in the late '60s I taped LPs with a half track stereo Revox A77 MK2 modified to have 18 dB headroom above 0 VU.  I bought a second generation Sony CD player in the '80s.

 

Comparing to concert hall sound as a reference, the following is a list of SQ in decreasing order.

Concert hall sound.

Microphone feed.

CD.

R to R tape.

Analog cassette.

LP.

 

I do not like LP for its clicks and pops.  Its warp and off center wow kills the pitch of music.

 

Analog cassette are quite musical.  S/N is not great.  There is a large amount of tape modulation noise caused by the tape drive mechanicals.  A solo violin sounds fuzzy.

 

R to R tape is like audio cassette but better in S/N, W&F, modulation noise.  However if you have a good three head deck making a live recording and switch between monitoring source and tape, you will hear a very small amount of fuzziness via tape.  I picked the Revox A77 for its very low modulation noise due to its excellent transport.  This fuzziness sometimes give an impression of smoothness to the listener.  I heard one digital recording studio passed their master through a R to R tape to get the sound they wanted.

 

CD playback, when executed well is musically enjoyable.  The devil is in the details.

 

A good microphone feed from a pair of good mics and preamps does not run the risk of a digital taste.

 

Concert hall sound is not limited in dynamics, frequency response and distortion.  It comes from around the audience, not just from a pair or several speakers.

 

After many years attempting to get good digital sound in my home, I have made some observations.  Your mileage may vary.

 

Computer hardware that I used (Windows PC), USB, wired LAN, PC, sound card (internal or external) worked against the SQ of my digital audio.  I went through the usual linear PSUs, noise filtering, jitter reclocking, cables, etc.  I was not able to get rid of the digitites entirely.

 

I ended up saying goodbye to the above components.  My songs are in my PC.  They stream via the Auralic Aries using WiFi.  Wired LAN had too much noise even with a wired LAN noise filter.  The Aries is connected to the Auralic Vega DAC using AES/EBU interconnect.  USB, Toslink and coax interconnects do not work as well.  The Vega drives a pair of Benchmark power amps, with a Jensen isolation transformer between them. 

 

I picked the Benchmark amps for their exceedingly low distortion (they are feed forward amps with crossover distortion nulled out).  Most DACs have a small amount of ultrasonics at their output.  They interact with the distortion of the amp (cross over distortion and slew rate distortion) and produce distortion products in the audio range.  The Jensen isolation transformer attenuates the common mode high frequency noise.

 

Somewhere in the system that I have not described, is a Dirac Live box from miniDSP and a digital crossover implemented by miniDSP FIR filters.  This audibly improves the transient response of the system.

 

This system allows me to tell which CD was from analog tape and which was purely digital.  Compared to digital the modulation noise (even with Dolby A) in analog tape gives a small amount of haziness and loss of crispness.  Digital audio, even with the 16/44 CD format, can be very good when the recording and playback tools are designed and used properly.

[vl]

13 hours ago, Rexp said:

I've heard the Aries/vega combo, sounds typically digital (digitites), if you take out your tweaks do you agree? And if so, which tweak is most important? 

Thanks

 

I heard the Aries/Vega combo in a few systems.  In most cases the sound was competent but clinical.  That is not the characteristics of the combo alone.  Many other digital systems sound somewhat sterile and non-involving.  The combo marginally wins over a couple of areas - speed and resolution.  The Aries sounds good because it is a low noise (computer noise) streaming device with a purpose built OS.  Windows 10 is not optimized for music.  The Vega sounds good not because of its ESS DAC.  What shines is its upsampler.  Upsamplers have their sonic signatures.  I heard many high quality DACs and only two appealed to me musically - the Cambridge Audio DACMagicPlus and the Vega.  The description of their upsamplers led me to the suspicion that Weiss Engineering in Switzerland provided them.

 

Since I did not evaluate each tweak singly I can only venture a guess.  USB stands out to be a no no for audio, introducing artificial brightness and loss of details.  I went from a pair of Acurus DIA amps to the Benchmark amps.  That was a quantum jump in resolution and clarity.  Digital harshness was reduced by using all balanced analog interconnects.  I bought Benchmark balanced interconnects and speaker cables, thinking that Benchmark would not embarrass themselves by selling poor quality cables.  I was not disappointed.  Note that the balanced interconnects must pass very high frequency noise so that the common mode noise is canceled at the balanced input.  I believe the Benchmark balanced interconnects are made from star-quad cables.

 

The Jensen isolation transformer also makes a clear difference in SQ - cleaner and more dynamic sound, due to it low pass filtering effect and common mode noise reduction.  Here is the link:

https://www.parts-express.com/jensen-iso-max-pi-2xx-dual-channel-xlr-line-input-isolator--246-0124

Use this model only as its bandwidth is limited to just above the audio range, filtering some of the higher frequency noise.  I use one such transformer wherever there is a pair of analog input and output to be connected together.

 

As for digital interconnects, Toslink does not pass noise but it has too much jitter.  Coax does not have common mode noise rejection so it is bright sounding.  AES/EBU has common mode rejection and it sound most like live sound. 

 

If any of your digital components use an ASRC, be cautious.  It behaves like a very basic upsampler and will not process a poorly mastered CD properly.  On the other hand once this poorly mastered CD is processed by the upsampler in the Vega, it will sound good in contrast.

[vl]

1 hour ago, Fitzcaraldo215 said:

I agree with your conclusions, as far as they go with prevailing commercial media.  I have dabbled and experimented with many of those same alternatives.  

 

If concert hall sound is your reference standard, and I quite agree it should be, then the next step beyond what you have considered is discretely recorded Mch sound.  To me, it is, without question, the closest approach in reproduction yet to concert hall sound, and not by just a little.  Equipment choices are always important, but breaking the sonic, spatial barriers imposed by 2-channel stereo are even more important if you wish to get closer to concert hall sound.

 

I agree.

[vl]

12 hours ago, tmtomh said:

 

 

 

In the analog days, it was quite easy to assemble a hi-fi component system.  So long as the individual components were good good sonic quality, the system usually sounded good.  This was true for LP and tape based sources.  At the worst we got some hum from ground loops.

 

It is different for digital audio.  Digital audio components are noise sensitive, from other digital components, the AC, the computer, etc.  Interconnects and AC cords react with noise differently so they sound different.  Then there are preamps and amps.  Some are very sensitive to the ultrasonic noise which is present from most digital gear.  It is due to high order harmonic distortion or slew rate induced distortion interacting with the ultrasonic noise in the digital gear, but not in analog gear.

 

We are into digital audio for more than 30 years but the way we pick a preamp or amp have remain more or less the same.  Digital audio requires amps and preamps of different behavior.  It requires noise isolation that is still not well understood by most manufacturers or audiophiles.  That is one reason why LP and tape still get such great support.  For those realize the potential of digital audio, we do not need to turn back to LP or tape.

[vl]

8 minutes ago, elcorso said:

 

Of course the early CD players were unlistenable,  but the same with a great majority of CDs of that time. But wait, about three years ago I bought some whose origin was Norway. Extraordinary because they were of extraordinary recordings also, from the AAD period. Quite the opposite DDD Sony Classics of the same era ...!

 

As today I have no complaints with recording and digital playback, but I have not listened to analog recording and playback for a long time. Maybe not to suffer? ?

 

Roch

 

Most CD players were not good sounding until the mid '90.  One exception was the Sony flagship CDP101.  It was head and shoulders above other CD players of those years and musical sounding.

 

Surprisingly the early CDs, even the ones from 1983 to 85, are actually quite good sounding.  It took about 30 years for me to get them to sound like master tapes or better.  

[vl]

1 hour ago, mansr said:

Very funny. I have one of those. It sounds terrible. Horrendous distortion of high frequencies.

 

I am sure your observation was accurate.  I did my audition via a good pair of head phones.  It matter what is down stream of the CD player.

[vl]

59 minutes ago, marce said:

rubbish, we have the best quality replay systems available these days with digital...

As to noise, most people involved in electronics understands noise, like you said its been with us 30 years... things have moved on. Interestingly from the DAC output its all analogue anyway, the digital part is just the storage medium.

 

Did the early CDs sound good in your system in the '80s?  Do they sound better now?

[vl]

6 hours ago, mansr said:

Your headphones may well be better than mine, but that's irrelevant if, as is the case, mine sound OK with other sources.

 

It would appear then the sample I heard was quite good in SQ.  I was comparing Sony's second generation CD players with the CDP101 and the CDP101 had more clarity and details.  I ended up buying a second gen Sony CD player that came close but not as good as the first gen CDP101, at half the price.  Even this second gen player was better sounding than my LPs (Thorens TD125, SME3009 Mk2, ADC25) and Revox A77 Mk2 (half track stereo).

 

My issue with CD players before the mid '90s was the upper two octaves not being open enough.  I stuck with Sony and they never sounded harsh.  I only had two CD players, a Sony second gen and a Sony third gen (4x oversampling), which was a little more open in the treble.  Then I moved on to a Sony preamp with their last 1-bit DAC.  It sounded better.  But it was not until after the turn of the century when I moved to a full digital system (digital preamp with room correction and three stereo digital amps) did I hear the open treble.  In some ways some earlier CD players were easier to integrate into a system with good SQ.

[vl]

25 minutes ago, sandyk said:

 

 I agree with mansr on this one. I had one of the first commercially available Sony 101s days before the discs were available.  

 My Marantz CD65 and a Nakamichi CD player were far  less harsh sounding than my original Sony player..

 With improvements in the PSU area and replacement of poor tolerance ladder resistors, the Marantz CD65 was then WAY better than the original Sony and many later players.

 

What I liked about the clarity and details in the CDP101 may be perceived as harshness by others.  This CD player had one trump card.  It could handle intersample overload without clipping.  

 

In the early CD days there was a CD (Philips, Widor Organ Symphonies) that produced a clipping like sound the first time I heard it from an FM station.  I bought the CD and played it on my second gen Sony CD player and got the same clipping sound.  I then went to a very nice audio store and listened to practically all their CD players.  Only the CDP101 played that track without clipping and with clarity.  Later on I learned it was intersample overload.  This Sony CDP handled it well.  

[vl]

1 hour ago, daverich4 said:

 

I had a CDP101 and still have the copy of Thriller that was one of the five free cd’s that came with it. It’s so bright and fatiguing that it’s unlistenable. 

 

Do other CDs sound overly bright as well on the CDP101?

[vl]

4 minutes ago, daverich4 said:

 

HAD a CDP101. Thirty years ago. The copy of Thriller that came with it is the same age. It as well as most of the other CD’s I bought back then are pretty much unlistenable. 

 

Do you still have these old CDs?  How do they sound now on modern digital playback system?

 

I remember that at least half of the classical CDs I bought before 1986 were not good sounding but some were very good.  These same bad CDs now sound normal in my current system.  The blame is partly on the mastering of these CDs and partly on  the inability of the CD player to handle these CDs.

[vl]

1 minute ago, daverich4 said:

 

I do. Unlistenable. 

 

Would you be able to post one track for me to listen to?

[vl]

2 minutes ago, fas42 said:

The trap that is so easy to fall into is to 'tune' a rig so that the 'right' CDs come across well, and others sound ordinary, or even awful. A well sorted setup can "handle anything", and all recorded music is then available to be fully enjoyed, with no excuses needed ...

 

I agree.

[vl]

3 hours ago, daverich4 said:

 

I guess I just don’t “grok” how your magic system can make the original CD version of Thriller sound good...

 

My experience tells me that it is feasible.  In my case most of the harshness/brightness of early CDs came from interaction of the CD with the noise and distortion of my playback gear.  Such interaction is dependent on frequency and loudness and it is not possible to use EQ to reduce it.  By painstakingly removing ultrasonic noise and reducing the noise and distortion of the playback chain, I was able to reduce the interaction (a form of intermodulation), making these CDs sound like normal recordings.  

 

For example to control noise from the AC, each AC powered audio component is powered from its own balanced AC isolation transformer.  These transformers also serve the purpose of isolating the noise from one digital audio component to the other.  Using one balanced AC transformer for several audio components is not good enough for my system.

[vl]

5 hours ago, Teresa said:

 

So you can listen to any CD no matter how poorly it was engineered, good for you. I can't, in fact I'm not fond of CDs or 16/44.1kHz PCM, even if well recorded. I prefer well engineered recordings from audiophile and boutique recording companies who take great pains to make the most natural sounding recordings possible. And I prefer them in DSD or high resolution PCM.

 

 

 

Not all CDs are created equal in SQ.  That is why given a digital playback system, some CDs sound good and some less good.  A good playback system will demonstrate a noticeable SQ difference among CDs.  An exceptionally good playback system will show much less difference among CDs.  

 

One of the key factors in good SQ is the upsampler.  A good upsampler has to have a good algorithm and ample computation power.  If the upsampler is not good enough, the system can still sound very good when playing back hires tracks, because the upsampler is not needed to do a hard job in playing hires tracks.  

[vl]

42 minutes ago, bachish said:

 

A 4416 recording, properly downsampled and dithered, should be indistinguishable from the high res version, especially on a high end system.

 

 

 

 

Sound Keeper's web site has tracks of different resolutions for free downloads.  Listen to them.  With my system the 16/44 tracks sounded very similar to hires tracks.

 

http://www.soundkeeperrecordings.com/format.htm

[vl]

26 minutes ago, daverich4 said:

 

Well, I for one would like some evidence of that because that sure hasn’t been my experience. 

 

A capable system does not resurrect a bad recording.  A capable system will make some a badly mastered CD sound better, if the bad characteristics are brightness, harshness, etc.

[vl]

11 hours ago, bachish said:

 

Thanks for the link. I'II check it out. I have done this sort of thing many times and over the years, due to certain experiences, I have developed a skepticism of my own ears.  

 

Oneof my hobbies is recording - I'm a classical musician. Anyway, when you have the experience of fiddling around with processing settings and being convinced you hear a difference for the better, only to find out the bypass button is pushed, you learn pretty quick about expectation bias and how powerful an effect it is on the brain.

 

 

 

Try the tracks from Lift.  There are enough musical sounds in the recording for a listening test.  Soundkeeper did an excellent job in the recording.  They do not edit or post process their tracks.  What they captured in real time is what we hear.  They take extreme effort to decimate the hires tracks to the CD format.  The Lift CD is one that sounds good on just about any digital playback system.  This is a good example of what the CD format is capable of.  

 

On my system it is not hard to hear the minute difference.  However most of my audiophile friends preferred the CD sound!  Please share with us what you hear.  I shall provide more details later as I do not what to influence anyone who wants to do the comparison.

 

On my system the Soundkeeper CD and hires tracks are definitely better than tape in clarity.

[vl]

5 hours ago, Rexp said:

I think, given you preferred cd to analog sources in the 80's you are not going to be a reliable reference for tape lovers. 

 

My reference is the concert hall sound.  Analog sound in the 80's was limited to LP and tape (both compact cassette and 1/4 inch).  From a musical stand point I preferred the early CDs to LPs for lack of clicks and pops, off-center and warp wow.  My CDs, when played by my CD player (Sony) were also more musical than the then compact cassette and the limited number of commercial 1/4 inch 7.5 ips tapes.

 

[vl]

44 minutes ago, bachish said:

 

 

 

This paper was presented in May 2010. The study was likely done prior to the year 2010, which puts the technology at approximately a decade old. Not only is the study a decade old but the converters were meh...I'd be curious to see if these results are repeatable today with top converters and the best modern SRC. Personally, I seriously doubt it would be. But I'm willing to change my mind if double blind tests show otherwise.  And besides people, they weren't even picking the 88.2 files when they detected differences! They chose the 44.1 files!

 

 

 

 

I once had an audition of the Dragon Boat track from Lift from Soundkeeper, in both 16/44 and 24/96.  There were five people excluding me.  All five considered themselves to be audiophiles.  The test was a blind test.  One of the five preferred the 24/96 track.  The rest liked the 16/44 track better!  Perhaps some people were accustomed to the sound of 16/44.  

 

To my ears the difference between the two tracks was small but audible.  I preferred the hires track.  The playback system was good enough to allow the listener to tell if the recording for the CD was made with analog tape or digital.

[vl]

19 minutes ago, sandyk said:

 

 How many C.A. members are capable of doing that, considering that many still find Computer Audio lacking in comparison with a good CD player ? Audio files played from System Memory have the potential to crap all over most affordable CD players.

 

 I wonder if Bachish was able to do that using high quality software such as JRiver for example ?

 Not ALL software players sound the same when playing the exact same material either.

 

After more than ten years of experimentation with computer based (windows) audio, I came to the realization that Windows PCs and good audio did not mix.  I abandoned using the PC for any part of my audio system related to sound.  Today my PC's function is to store the music tracks.  These tracks are streamed via WiFi to the Auralic Aries streamer, which is a purpose built computer doing the streaming and the management of the music library.  It does an excellent job of that.  

 

As good as the Aries is, it needs fine tuning.  For example I use the 5.8 GHz WiFi band exclusively for this streaming.  Even with one streamer operating, the choice of 20 MHz or 80 MHz bandwidth for WiFi imparted a different sonic signature, though both have data throughput far in excess of what I use for streaming.