That's been put forward before and has not gotten very far. I agree BTW there are plenty of artifacts that create a false ambience or atmosphere in vinyl with low frequency wander in both AM and FM.
Are art and science opposed? My CD's of Furtwangler's recordings of Wagner still capture the art.
Of course not. One would have to confront their humanity if they did. Audiophiles are not mere mortal beings.
Yeah. And really, there's nothing wrong with a little soft focus and airbrushing. Hefner built an empire on it.
se
The last performing space I did was NC15. A symphony orchestra in that space on stage and into the front of the listening area measured up to 109dba slow. That would have 129dbc peaks or more. So that is 114db dynamic range. If you understand the ear can perceive tones below the noise floor that would add another 10db or so. So just for that recording you need 21 bits. I think good practice would allow a few more for margins.
If you look at Fletcher Munson 60 db at 30 hz is equal loudness to -6 at 4000 hz that is 11 bits just to record the same perceived volume! Even with a dynamic range of 30 db you run past 16 bits.
Now if you listen to modern studio recordings, 16 bits is certainly enough. Quite often the folks who do the recordings are quite capable and utilize their medium well to get the best results possible.
A common misconception is that a symphony orchestra is louder than a rock concert. That is not quite true. A symphony takes up a lot of space producing that energy. If the total energy were compacted so you could measure it all at 1 meter from a point source then it would be around 130db. A rock concert aims for 102 db average (don't confuse that with the symphonies' peaks just mentioned) at 100 feet (or equal to 140db @ 1m). A good sound company will keep that to a maximum of 105 anywhere in the seating area.
For phase detection tests try a 6khz fundamental with a third harmonic that at 0 degrees of shift reaches it's peak at the same time as the fundamental. Then to make sure you are not hearing the level change compare say +/- 5 out to maybe +/- 60 degrees. See if you can hear the difference.
You do not need to put your head in a vise no matter how tempting the idea. Although putting your head in a vice is entirely your choice!
Yes there may be parts that meet my design goals but I haven't yet seen a system. Knowing the limits of transducers I would settle for the electronics including the recording system.
Last edited:
Digital storage
Thanks for posting the link, Joachim! Interesting paper for sure! One thing which has to be restated is that the 16-bit 44.1KHz system was settled on in the late 1970's when A>Ds and D>As and digital storage and BW were all really expensive and hard to make correctly. It was not chosen at the time due to any perception that it was perfect. I agree with one of the comments that 16 bit 44.1 KHz digital may even be overkill for general populous audio distribution (I mean look at how happy most are with low bit-rate mp3!!), but quite a bit (sorry for the punishment) more is needed in the original recording and mixdown to be truly lossless for many reasons, as we all know.
My experience doing 2-track mastering in well-designed studios agrees with comments by Ed, you can hear the noise floor (not digital black, but no signal, just dither) in a 16-bit master when the levels are set to the peaks are at or near clipping. But as Steve points out, due to cochlear desensitization, both extremes of the dynamic range cannot usually be experienced within a short time of each other.
Is the noise floor (and low-level distortion within) of a 16-bit system a problem? For me it never got in the way of function or enjoyment, but if you were designing a new storage or transmission standard, especially now in the day of inexpensive digital BW and storage, I would want it to be better than any expected use, with added margin for less-than-optimum level setting or peak limiting. From this viewpoint, the author's suggested criteria seem pertinent.
The take-home message for me in this discussion of dynamic range is it is not the limiting nor defining factor to what many consider to be a good sounding recording, although it may be with some. The relatively high level and impulse laden noise floor of some LPs I consider to be a superior listening experience proves this to me. Our search for the perhaps non-existent single factor which defines the audiophile experience must look elsewhere.
Good discussion, though! Chew on this: http://www.channld.com/purevinyl/index.html
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill
www.wxyc.org
1st on the internet
Thanks for posting the link, Joachim! Interesting paper for sure! One thing which has to be restated is that the 16-bit 44.1KHz system was settled on in the late 1970's when A>Ds and D>As and digital storage and BW were all really expensive and hard to make correctly. It was not chosen at the time due to any perception that it was perfect. I agree with one of the comments that 16 bit 44.1 KHz digital may even be overkill for general populous audio distribution (I mean look at how happy most are with low bit-rate mp3!!), but quite a bit (sorry for the punishment) more is needed in the original recording and mixdown to be truly lossless for many reasons, as we all know.
My experience doing 2-track mastering in well-designed studios agrees with comments by Ed, you can hear the noise floor (not digital black, but no signal, just dither) in a 16-bit master when the levels are set to the peaks are at or near clipping. But as Steve points out, due to cochlear desensitization, both extremes of the dynamic range cannot usually be experienced within a short time of each other.
Is the noise floor (and low-level distortion within) of a 16-bit system a problem? For me it never got in the way of function or enjoyment, but if you were designing a new storage or transmission standard, especially now in the day of inexpensive digital BW and storage, I would want it to be better than any expected use, with added margin for less-than-optimum level setting or peak limiting. From this viewpoint, the author's suggested criteria seem pertinent.
The take-home message for me in this discussion of dynamic range is it is not the limiting nor defining factor to what many consider to be a good sounding recording, although it may be with some. The relatively high level and impulse laden noise floor of some LPs I consider to be a superior listening experience proves this to me. Our search for the perhaps non-existent single factor which defines the audiophile experience must look elsewhere.
Good discussion, though! Chew on this: http://www.channld.com/purevinyl/index.html
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill
www.wxyc.org
1st on the internet
Last edited:
etc...
Yeah, it's impossible to satisfy objectivists.
But it would attract more people less expert and more humble, like those who follow NP forum without posing too many questions.
A Neumann U47 has 23dBA self noise and .5% Max SPL of 117dB and lots of great recordings have been made with it. You have rolled these head room numbers before. I don't follow the reasoning and don't know if it makes any sense. I've dithered 24 bit recordings down to 16 and can't hear the difference YM obviously varies.
Last edited:
Have you tried this? I was wondering if the zero phase rumble filters were causal. Great LP choice Liz Fraser is great.
Splitting Hairs....
The question is what objective test methods are appropriate to define really fine differences ?.
Dave.
If the recordings are good enough resolution and the test gear is good enough resolution to discern fine differences, then subjective opinions are moot are they not ?.
The question is what objective test methods are appropriate to define really fine differences ?.
Dave.
For some reason everybody glorifies digital. Digital, in its ideal form may be next to perfect, but over the last 43 years that I have listened to digital, I have NOT heard PERFECT digital yet. Just at this year's CES, one company that I work for, had a pair of $150,000 loudspeakers, and an equal amount of digital and analog playback electronics on display at the Venetian. I had just purchased, for myself, an SACD of 'The Raven' by Rebecca Pigeon having previously heard it on vinyl. The vinyl version was sold out, so I got the SACD. We put it in the system and played it back. Sounded pretty good. Then, as I was leaving the room, the same song from the same singer came on, being put into the system earlier, so I sat back down and listened, again. It was different in a subtle way, I think slightly better, but then I turned to my boss and said: "It's not as good as vinyl". It is just a fact, at least a professional opinion, and I will stand by it. Now, IF we can do better with digital as to make it AS GOOD as vinyl, then we can put away the vinyl. Now this is just MY opinion, the listening opinion of a retired guy with fairly good hearing (tested by doctors). What can a 20 year old woman hear? What about a young guy? I could easily hear 15K test tones when I was young, I even found that I could resolve 24K when I was in my 20's, as a tone, not just a feeling. No more, of course.
IF I can still hear the difference between vinyl and digital at 69 years old, there is still work to do. IF you, Scott, CANNOT hear where more than 16 bits is necessary at 60 years old, maybe you have a young daughter or even a granddaughter that you should ask whether they can hear any background noise, rather than trust your own hearing as the reference standard.
IF I can still hear the difference between vinyl and digital at 69 years old, there is still work to do. IF you, Scott, CANNOT hear where more than 16 bits is necessary at 60 years old, maybe you have a young daughter or even a granddaughter that you should ask whether they can hear any background noise, rather than trust your own hearing as the reference standard.
Last edited:
Yes there are even good recordings made when the 16 bits was really 9 & 7. Some people love MP3 recordings. I would not expect an emotional appeal kind of argument from you.
The U47 is a 1949 design. In many recordings more than one microphone is used and that has the effect when properly done of extending the dynamic range. As noise is measured broadband and signals can be perceived below that the s/n of even a U47 may effectively exceed 16 bits. So that brings up the question, "Is dither as effective in allowing signals below the noise floor to be perceived as normal human hearing can extract from more common analog noise?" As I have no experience with it I will not offer an OPINION.
I have seen many 24 bit recordings that were 22 real bits or less, but I will assume you knew better than to use a faked 24 bit recording. As in any experiment there are always issues as to if you were really just looking at what you thought you were. The last audio A/B switch I measured had distortion -110 db re the 1.2 volt rms test signal. So the test switch was only good to 18 bits! That is the best switch I have measured so far!
We have the technology to do better. When a bit of research allows us to do so without spending silly amounts of money I see no reason not to.
Yes that's true, but the issue was that you would send me two 'subtly different' pieces of music. We should at least make certain that they are that. That you are able to discern between the two pieces, in a controlled test with, say, 98% confidence. It's rather useless to measure two pieces for differences you can't hear (for this discussion).
As to the 2nd question, I don't know. It surely depends on the differences, but we don't have an accepted standard to define them. Suppose you tell me that you find that one piece sounds 'more defined', and I measure lower THD, does that correlate? If you say one piece sounds more woolly, and I measure more xtalk, does that correlate? There are *some* factors I believe that correlate, for instance 'warm sounding' often correlates with increased even harmonics. But it's all a pretty loose correlation.
jan didden
Digital Domain RIAA De-emphasis
Scott,
I had not tried it, but had considered it for LP library acqusition at WXYC. This particular package (Channel D PureVinyl) seems to be one of the more serious attempts to do it with fidelity. If I get a few minutes I may DL the demo.
For serious listening at home, I cannot see an advantage (other than the lure of lazy-ars couch potato remote control track access) to ripping vinyl. I do not worry about 'wearing the vinyl out', I have a more than a few LPs, so despite frequent LP listening each one gets played about once per five to ten years, so the fidelity on my LPs will outlast me. I decided a while ago to not transfer my vinyl, because when I listen I am AT the stereo, and I can play the vinyl directly. This gives a minimalist playback topology without various and sundry (and unnecessary) A>D and D>A conversions, no matter how 'transparent' they may be. Plus, Henry Cow dislikes being digitized, it is another attempt to repress the proletariat! I dare not incur the wrath of Dagmar Krauss!!!
For background music (during a party, for instance) I have a music server with a lot of CD rips feeding an Aux in on the stereo.
I am sure you understand the compromise involved in an analog RIAA preamp much better than I, but I can't think of a reason why the sound quality of digital RIAA, even optimally done would be better than a high-end RIAA preamp. Vanishingly close, maybe. Better...?
The RIAA curve taxes both the noise floor and headroom limitations of analog phono preamps, but the best designs make these problems pretty much non-audible. Besides; using a flat preamp before an A>D merely moves the headroom and noise floor optimization problems to the A>D front-end, necessitating accurate level setting for each transfer, which is perhaps more problematic for the user.
I know that there are eq and filter designs possible in the digital domain that cannot be easily executed in the analog world, so maybe for filtering and correction of LPs with problems, this approach has merit? I expect to get a bit more personal insight into the noise spectrum of analog RIAA systems when I finish that Excel table...I'll let you know when I do.
Onward into the fog...
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill
www.wxyc.org
1st on the internet
Scott,
I had not tried it, but had considered it for LP library acqusition at WXYC. This particular package (Channel D PureVinyl) seems to be one of the more serious attempts to do it with fidelity. If I get a few minutes I may DL the demo.
For serious listening at home, I cannot see an advantage (other than the lure of lazy-ars couch potato remote control track access) to ripping vinyl. I do not worry about 'wearing the vinyl out', I have a more than a few LPs, so despite frequent LP listening each one gets played about once per five to ten years, so the fidelity on my LPs will outlast me. I decided a while ago to not transfer my vinyl, because when I listen I am AT the stereo, and I can play the vinyl directly. This gives a minimalist playback topology without various and sundry (and unnecessary) A>D and D>A conversions, no matter how 'transparent' they may be. Plus, Henry Cow dislikes being digitized, it is another attempt to repress the proletariat! I dare not incur the wrath of Dagmar Krauss!!!
For background music (during a party, for instance) I have a music server with a lot of CD rips feeding an Aux in on the stereo.
I am sure you understand the compromise involved in an analog RIAA preamp much better than I, but I can't think of a reason why the sound quality of digital RIAA, even optimally done would be better than a high-end RIAA preamp. Vanishingly close, maybe. Better...?
The RIAA curve taxes both the noise floor and headroom limitations of analog phono preamps, but the best designs make these problems pretty much non-audible. Besides; using a flat preamp before an A>D merely moves the headroom and noise floor optimization problems to the A>D front-end, necessitating accurate level setting for each transfer, which is perhaps more problematic for the user.
I know that there are eq and filter designs possible in the digital domain that cannot be easily executed in the analog world, so maybe for filtering and correction of LPs with problems, this approach has merit? I expect to get a bit more personal insight into the noise spectrum of analog RIAA systems when I finish that Excel table...I'll let you know when I do.
Onward into the fog...
Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill
www.wxyc.org
1st on the internet
Last edited:
- Status
- Not open for further replies.