FIDELITY POTENTIAL INDEX
|BACK 1 2 3|
I'm a long time owner of 645's and have used the 45" ribbons in a variety of ways and systems, so I'm on your newsletter list. I posted the link to your "Fidelity Potential" page to a private group of audio guys as we were roughly on the topic of MP3's and how "kids today" don't care about quality etc. I won't name names, but some of these are guys you would have heard of. One of them asked: "Could someone further indulge me on the derivation for this theoretical "Fidelity Potential?" so I thought you may want to have a whack at that. If you do I will forward to the group and keep you posted.
Personally I found it interesting, but also could not imagine how you might have quantified the "anecdotal", I think you referred to a conversion to sampling rate if I recall (I read the page a few days ago).
There is a fair amount of discussion of the method on the page and there will be more when some posts are put up. Several posts will elaborate on what I'm saying here and include their own estimates of analog format capability. So ranges are important to include.
The "anecdotal" are judgements about how far into the noise floor and ceiling one can hear. These extend the range of the hard specs for the format say vinyl. Digital has a hard, brick wall limit on dynamic range. It has no noise of it's own within that dynamic range. Analog formats have lots of noise sources and the signal to noise is less than the dynamic range. However, one can still perceive signal into the range of noise and this is what effectively extends the dynamic range of analog sources.
Yes, these are personal estimates about what sounds good and why but they affect the range of the rating not the core value.
If we could arrive at similar ranges for the effects of digital artifacts, I'd establish ranges on that basis as well. Of course, these would reduce the digital format FPIs whereas the analog FPIs are increased by the process.
Particularly it would be interesting to assign reductions for various compression schemes and for room correction methods which may increase amplitude correctness but reduce "transparency".
Let me know what you think!
If you are playing with the R45s, - do a Coaxial Ribbon LineSource - build a system with 4+ good 7" drivers in sealed enclosures and stick the R45s in front of them, hopefully using a digital crossover and re-arrange your audiophile benchmarks. This is a step up from anything you have heard guaranteed.
Your rating is an interesting project! The ratings look sensible on a wide window. I tend to disagree however on certain basic assumptions.
Sampling rate is assumed to be identical to high-level bandwidth (eg. 0 dB -10 dB). So you assume vinyl as 50’000 kHz sampling rate. This is maybe true for high level signals (maybe even worse). But, with a good low inductance cartridge and a capable stylus, like hyperelliptical, VdH I&II, Gyger I&II, micro-ridge, Paroc etc. at least 100’000 kHz sampling should be assumed. There is even proof that 75 kHz signals are traced with LPs that were cut with DMM (Direct metal mastering). I think in real life there is a wide variability in the amount of ultrasonic content on LP. But the fact is, there is considerable energy above 20 kHz available in LPs (not always the recorded signal...), And there is traceable energy up to 75 kHz. Then there is the roll-off frequency and order of roll-off in analog systems compared to digital, which makes even a (IMO wrongly) assumed “analogue sampling frequency” of 50’000 Hz audibly different to a digital one, with it’s sudden, high order drop-off vs. the more “natural” analogue r oll-off, which behaves more closely related to real-ear experiences with acoustical phenomena. Which most probably is audible in supersonic “inaudible” regions, specially when “linear phase” pre-ringing oversampling filters are involved. The problem in lining up digital and analogue systems sonically is the problem of comparing apples with oranges. High-level linearity (Freq. Resp. And distortion) vs. low-level etc.
Sonically I would not totally disagree with 320 kBs MP 3 vs. Cassettes being in a similar range of “fidelity”, still it’s my feeling that one can get (considerably) more involved in the music and the sound with a superb (and expensive) cassette tape recorder. To my ears there is a certain aspect in the sound of digital compressed formats reminding of bad main’s, sucking out some of the bounce and communicating warmth and “energy” of the music. You don’t have this with cassette, and, BTW, good cassettes register information above 15 kHz (-20 dB bandwidth), contrary to MP3. This is audible too – eg. in PRAT...
My ears told me on any DVD vs. SACD comparison I made (Sony SACD player, Audio Synthesis DAX Discrete) that even 24/96 PCM(DAD) sounded potentially more alive and natural than SACD. DVD-Audio sounds good too but I haven’t had the experience of totally locking into the performance with it, as was possible with optimal 24/96 DADs. Less data processing? 192 kHz / 24 Bit is promising, haven’t really heard it yet. And it’s a huge storage consumer. SACD is (for me) a theoretically impressive and brillant format which was promoted with kind of an audiophile-underground marketing hype, but which is, contrary to the hype of being “most analogfue-like”, highly feedback processed (high-order noise-shaping) and somehow in the end sounds kind of like it. It is definitely not on the level of 24/96 kHz for me, and is even a slight sonical trade-off compared to good CD. PRAT is in favour of CD, bass is heavier (hifi impressive) on SACD, and the top octave has considerably more “air” though. A further inherent problem of all these high data rate formats when burned/pressed on optical media is the considerable higher speed and motor forces involved in the process of reading DVD & Blue Ray (I think too) compared to CD. And even in CDs this problem is audible. An interesting observation when playing different data formats on my iBook and MacBooks: When you look at the processor load you see that the non-lossy compressed format ALC needs about 50% more processor work compared to AIFF or WAV. This is to my ears slightly audible on both the computer and an iPod Touch. Processor work is in the end analogue current and shapesdigital power supply noise, which in reality can not completely be blocked out by any measures IME.
You were looking for “trouble” with that rating, didn’t you ? ;-)
Note by Andrew Marshall on vinyl LP frequency response. - the quadraphonic systems may have included response up to 100kHz to manage the signal steering but they never worked well probable because at 100k, the LP playback system is simply not reliable. Also, cassette tape can go up to 20kHz with the right tape and noise reduction system.
"Civil, informed and humourous comments will be given preference". Er, quite. But the idea behind your your is rather simplistic and you have to understand may well provoke an irate response. You end up with 24/194 PCM sounding 7 times better than LP, a result so different from reality it doesn't really bear any further comment.
However, what George has to say about compression systems being impossible to measure meaningfully with steady tones (multi-tones do yield a result) yet wrecking music is very true. More amazingly they are contrived on listening tests alone - often crude ones - a point few people understand. Read what Karl-Heinz Brandenburg and the Fraunhofer Institute say. So thumbs up to George on this!
Finally, the idea of a controlled listening area freed, to an extent, from room effects, as provided by a line array is an interesting one little talked about. This also brings in ribbons, which most of us admire. Hope we can cover your products sometime.
So good luck with your Index. Time to strap on the tin hat methinks.
Noel Keywood, publisher Hi-Fi World
I knew there would be criticism but it has not turned out to be as severe as anticipated. I may be a little beaten up but not unhorsed.
The FPI is not intended to be a linear representation of the fidelity a human is capable of perceiving, but rather simply a method of putting raw capability in a numerical order. So day in, day out, the 192/24 can be counted on to sound better than the LP. But, as you say, not 7 times better or even 2 times better. Just distinctly better on most recordings.
Compression systems and the judgement of being able to listen into the analog noise floors and ceilings can skew the hard numbers. Also, from my own point of view, room correction systems flatten amplitude but result in some loss of transparency - I'm not sure at all as to how this happens or how to represent it numerically.
And the flaws in digital which effectively reduce the dynamic range or add noise/distortion are not represented in the FPI either. Maybe all in good time.
Thanks for your comment!
by Dennis Burton
CD vs. Vinyl with some other stuff thrown in.
When I was a child, the hifi debate was over small vs. large speakers (really) and transistors vs. vacuum tubes. These arguments never end because people use music systems for different reasons and also hear different things. For some people dynamics seem to be important beyond all reason, and so that is what they will notice. Others may worship at the shrine of tonal purity and musical pitch. Most of these people are quite in denial, claiming that, in fact they want it all; imaging, dynamics, vast bandwidths, seamless crossover, low coloration etc. etc. But this is not necessarily true.
Each component in a music system has a “sound”. Now we are getting simple. A reel to reel tape machine will not “sound” like a turntable, or a cassette machine-all technical limitations and considerations aside. Transistors have a sound. In the end you go with what you like best. One man’s fluid sugary midrange from tubes is trumped by another man’s huge expansive bass line from a huge Class A transistor. Or is it the other way around?
I listen to cassettes, open reel quarter track and half-track tape, MP3’s, CD, vinyl, and FM, and have heard quite acceptable results in all of them. Oh I am a blasphemer aren’t I? The ultimate source? Well it may well be reel to reel. But the machines are hard to use, tapes have to be rewound to prevent print-through, there is hiss, there can be dropouts, transport noise etc. This is not convenient. So then vinyl, only the big catch is that I simply cannot afford the equipment necessary to have that happen. I use a Linn Sondek LP12 and it is a nice turntable, but of course I would need between $30 and $60 thousand for a really good turntable and would then be off hunting for a suitable and expensive cartridge for it. Let us not discuss the drub pressings foisted on us over the years, which simply cannot be rescued by any known technology, or the fact that a cutting lathe has rumble figures you simply would not accept in a turntable of any price.
My very first CD player cost less than the phono cartridge in my turntable. I knew then that the CD was inevitable. But more than that, at $175 that Philips CD player smoked anything and everything in a turntable at that price range. It’s not about money you say; it is art, purity, subtlety, and poise. Oh, I suppose so, but personally I enjoy music. I have heard an MP3 mastering from a 45RPM 7” single of the Chiffons singing He’s So Fine coming out of the hideous 3” speaker of my kitchen clock radio and been thoroughly delighted. The audio mangling was scarcely describable, with multipath from the FM into the bargain and the fridge itself adding noise to the background. Music, as with Art, demands that we bring something to the table or we are simply being entertained for good or bad. The whole idea of HIFI is to enhance the “transportation” of that music. With listening, I am not talking surrender, I am talking engagement. That is not the HIFI bit, that is your bit.
Does a CD sound better than a vinyl record? Well that depends. If it is Classical Orchestral music recorded in a great hall, then I suspect that Half Track Reel To Reel or vinyl is the winner. If the music is Ultra Chilled Down Tempo Electronica then why use vinyl when all the samples are 16 bit anyways-CD wins there. And that is just two examples folks. Some people love songs and listen to the words, others adore virtuosity and listen obsessively to the playing, ignoring whatever words there might be. Still others surrender to melody and yet others to rhythmic structure. Still others love tempo and arrangement. Some thrill at production values and sonic volcanoes. And there are still others and others and others, but they will all tell you they want it all, and none of them (or you!) can pretend to be a final judge of anything we all might or might not care to actually hear.
Years ago at the dawn of digital, I listened to a test involving a Classical vocal quintet in a very large room singing. It was a recording studio and we had two Neumann microphones optimally placed and we used a pair of very expensive microphone preamps and simultaneously routed the signal to a Dat recorder, which was then a new recordable digital medium, also the signal went to a very large Sony PCM reel to reel, and
finally also to an Otari Studio Half Track analogue machine. All three recordings “sounded” different. Most noticeable however was that the analogue half track recording, although having a bit of noticeable hiss provided a very interestingly pleasant insight into the harmonic structure of the blending of the vocals as presented by the room itself. This was absent from both digital recordings, which presented as comparatively “dry”, although I hasten to add, both of them sounded wonderful with the Sony sounding superior to the Dat. We could not actually speculate as to what might be happening except to surmise that perhaps 16 bit 44.1 KHz was insufficient to render the complexities of the harmonic structure. Does this mean that the analogue was better?
Well, to whom? I mean, the choir members were listening to their performance and couldn’t understand our obsession with some tiny detail of the sound. They thought it all sounded great. The choirmaster loved the room information and chose the analogue. The Sony reel-to-reel mastering machine was worth about $110k at the time and the Dat was $799 and the analogue half-track was about $34K. For the studio, the money certainly mattered, but quality is number one. But was the absolute silence of the digital to be opted for over resolution of harmonic structure? Which best was best? They ended up buying the Sony, but they never sold the Otari. They figured that for multi track recordings where lots of blending and stacking of tracks happens, then the edge goes to the digital because it does this very well. Co-incident pair purist Classical recordings were offered the choice between analogue and digital and for most of them, they either ignored the faint hiss or never heard it and went analogue. Which is best?
OK, so you’re saying that someone who loves all this stuff has to be the judge, the final arbitrator so that for everyone else there is a reliable standard to follow, and if we can’t agree on or figure out what is best, then we aren’t worth our salt as experts. If asked any expert on anything is sure of herself or himself. So if I do not know wines, I can ask an expert and know that the given advice is based on vast experience and therefore worthy. But hey! It is the same for them-there is no best. It just depends on exactly what you want, who you are, where you are and maybe a few hundred other things.
There is the story of the Buddha walking through a small village, and as he was passing the Butcher he happened to overhear a customer ask the butcher which of his cuts of meat was the best. The butcher replied that they were all the best. At these words the Buddha became enlightened.
HOME | PRODUCTS | FEEDBACK | NEW TO HIGH END AUDIO? | EXPERT ADVICE | PURCHASING | COMPANY | CONTACT
© 2000 - 2005 Newform Research
Site by: VY Web Design