Back to DSD and distortion, here is one clue as to where I might have heard this.
http://www.thewelltemperedcomputer.com/KB/DSD.htm
The author mentions distortion several times but not harmonic distortion.
Back to DSD and distortion, here is one clue as to where I might have heard this.
http://www.thewelltemperedcomputer.com/KB/DSD.htm
The author mentions distortion several times but not harmonic distortion.
Thereās a lot of loose speculation there and (probably accidentally) misleading statements. There are many kinds of distortion, the question is which kinds interfere with our hearing (and appreciating) the music we all like. Noise is a distortion: noise correlated with the signal is very noticeable, noise uncorrelated with the signal is quite benign, the ear/brain is really good at ignoring noise it hears in the real worldā¦ Phrases like āā¦above 20 kHz the music signal will start to be drowning in the noise.ā are misleading even taking the Stereophile graph as described. Note that the signal to noise ratio is still better than 50dB at 90k, but for PCM at 176.4k or 192k the signal to noise ratio is 0 dB. Iād much rather have a 50dB S/N ratio there than no signal at all. But the graph from Stereophile is misleading: itās a graph of the noise thru a one third octave swept filter (people shouldnāt take graphs and leave behind the captioningā¦) So, for example, the level of the noise in the DSD signal being displayed at 20k is being affected by the signal at 30k, etcā¦ Similarly the Hawksford graph is wrong in the caption claims that itās showing a noise reference line for 20/44.1, but 20 bits is approx. -120dB S/N not -150dB S/N as shown, similarly a 44.1k sample rate doesnāt represent a signal at 200k as shown: it doesnāt go above 22.05k. Another misleading phrase is āAll these strong high frequency content might fry your tweeters!ā I donāt know about you but I wouldnāt want speakers that canāt handle a HF signal thatās more than a factor of 100 lower than their rated signal handling just an octave lower. Or to put it another way, the DSD noise is much lower than a perfectly legal PCM signal at a 176.4 or 192k sample rate.
Iāll stop now
P.S. Since weāre in a thread about tape, just imagine the ultrasonic frequency spike at the tape bias frequency - itās quite a bit louder than the DSD noise and itās a pure tone, ā¦
wglenn, I have not found exactly what I remember but I keep going back to this thread discussing a paper by the Swiss engineer Daniel Weiss:
http://www.realhd-audio.com/?p=1505
and this mind-numbing thread on a discussion of ātrueā DSD DACs vs those that first convert the stream to PCM in response to Charlie Hansenās (Ayre) PCM vs DSD Comparison;
http://www.audiostream.com/content/ayres-pcm-dsd-comparison#k0Exqsr7cZkPciVZ.97
But I think my confusion stems from this statement:
I canāt help but wonder if whatās happening here is like tube amps and analogue playback (eg. vinyl). Objectively the DSD conversion adds distortion but the anomalies are not perceived as objectionable and in some material, the added noise and imprecision actually makes it sound less āsterileā, āclinicalā, more ārealā (conversely being in an anechoic chamber is disturbingly unreal due to the profound silence). It would make sense to me that some people could prefer DSD64 over DSD128 upconversion since DSD64 will give you more of that distortion. Even though the noise is ultrasonic in nature as measured off the DAC, nonlinearities in the playback system like your headphones and speakers (perhaps certain amps as well) could create audible intermodulation. Maybe for certain music, this could be especially beneficial.
which is excerpted from the following article showing hi-res PCM and DSD64 and DSD 128 to be identical in many measurements.
Since these articles are a couple of years old, are they still valid and does the use of an FPGA and Tedās magic in the DS change the playing field? Andā¦I recorded vinyl directly to tape and compared it via headphones to analogue from my CD player. Without the DS in the chain, some of the magic was gone, therefore the magic IS in the DS.
I think the euphonic argument about DSDās noise comes from the cognitive dissonance that some have that DSD could sound better. People seize on the most obvious difference between FR measurements between DSD and PCM and then āpost hoc ergo propter hocā. They (usually) donāt look at the phase response of DSD vs PCM and take note that DSD preserves the waveform shape better, and in particular that transients are better preserved. IMO itās this about DSD that matters the most for sound quality and the extra noise isnāt really a feature, but more of an unavoidable (but mostly benign) side effect of the noise shaping in DSD.
The modulation argument doesnāt hold water either: many of us have speakers flat to 50 or 80k and amps and preamps that have at least that bandwidth. DSD compared to PCM is perhaps even better on these systems not worse.
BTW I donāt intend to imply that DSD is necessary to preserve transients - high enough resolution PCM can, but it would require a lot more processing and bandwidth than DSD for the same low distortion of waveform shape. I chose DSD simply because the āhardā work for me was an excellent output stage (which is simple with DSD) and the rest was simply math/programming.
Stevie B said Even though the noise is ultrasonic in nature as measured off the DAC, nonlinearities in the playback system like your headphones and speakers (perhaps certain amps as well) could create audible intermodulation. Maybe for certain music, this could be especially beneficial.Interesting concept. IM distortion and difference tones are real, but I have trouble accepting the headphones and speakers are reproducing ultrasonic noises in an euphoric fashion to lead some to prefer DSD.
āā¦but I have trouble accepting the headphones and speakers are reproducing ultrasonic noises in an euphoric fashion to lead some to prefer DSD.ā
Hehe, yep. Certainly it would fall to harmonic distortion and so far itās not happening.
āThey (usually) donāt look at the phase response of DSD vs PCM and take note that DSD preserves the waveform shape better, and in particular that transients are better preserved.ā -Ted
Glad to hear my suspicions about phase info confirmed. This is what really gobsmacks me about DSD and why I often prefer the DSD version of a track to the PCM version. If the source for both is an analog master, I think that the phase relationships sound more ācorrectā to me. I find it hard to put into words but Iāve thought lots about phase this year after learning so much from Galen and his cable designs. Is this the āthingā that analog gets right that PCM misses out on, phase preservation (relative phase relationships)? Is this why master tapes sound so incredible?
Master tapes sounds great, but so do digital masters.
Master tapes exhibit a great deal more phase distortion than digital. For example, analog equalizers introduce a great deal of phase shift. Digital equalizers can either mimic this behavior or avoid it (linear shift) depending on oneās choice of digital equalization.
Then there is the tape recorder itself and all its analog circuity: capacitors, transformers, inductors, record and playback heads, etc. All of these goodies introduce phase smearing, especially capacitors and inductors. The phase shift is greater at high frequencies. Plus, If the azimuth is off, there is phase shift between tracks with tape - which can be particularly significant with 24-track tape - this phase shift is heard as the shifted and unshifted tracks interact. High frequency phase shift can sound more open as it creates a widening effect.
Tape emulators mimic all of these characteristics, along with the natural compression of tape. Many like this sound even though it is in reality less accurate than digital. This is what people hear as āwarmer,ā more analog like.
Itās just darn hard to sort out. On further thought, Reference Recordings PCM has that āphase coherentā sound that Iām thinking of. I also just got the DSD transfer of Kind of Blue and so far, I think that I like the Hi Res PCM remaster better. I havenāt spent a lot of time with the DSD version yet so this is just a casual first impression. I do wish that Rudy VG would share more of his knowledge. Clearly a brilliant ear and technician but stingy on explaining how he did it. Interestingly, he was an Optometrist who continued to practice to fund his recording habit, an amateur with extraordinary talent (and talented musician friends!).
Back to bewilderment over formatsā¦
It is easy once we remember that the room and microphone placement determine 95% of the quality of the recording. The rest is just fluff to obsess over.
Iām learninā, Iiiiām learninā This obsession is getting in the way of my other obsession but they do go hand in hand, donāt they.
As itās fine to measure , for me in the end itās how it sounds. Our music is complex and our brains all seem to be wired a bit different too. I wonder how much of preference is just wiring and not really true preference . Meaning as I say we all learn to hear different. Some seem to hear so much more than others. Kind of like wine or food if one is into it there is so many more layers to our senses . I eat out a a lot and can tell the variations in the food in the same place , itās easy. I also know home cooking done right blows away most outside eating. Food tasting has layers like our manic does and most definitely booze too. I can easily tell macallen 12/15/18 . Same with JW SCOTHCH
It could be that in our race to remove as much as possible from the reproduction chain and get ever closer to the source, we have forgotten the value of so many wonderful things of the past that made the sound we rememberā¦
http://www.rollingstone.com/music/features/accessories-mysteries-of-the-black-box-revealed-19770908
Sorry, couldnāt resist.
Ah, the good old days.
Note the reference, in 1977, to the heavy dynamic compression necessary to get music onto tape.
Same for vinyl. For me itās more choice and what we feel is real.