Home |
Search |
Today's Posts |
|
Metalworking (rec.crafts.metalworking) Discuss various aspects of working with metal, such as machining, welding, metal joining, screwing, casting, hardening/tempering, blacksmithing/forging, spinning and hammer work, sheet metal work. |
Reply |
|
|
LinkBack | Thread Tools | Display Modes |
#1
|
|||
|
|||
Cryogenic treatment of brass instruments?
On another newsgroup, the subject of cryogenic treatment of brass
instruments came up, with the claim that it was "snake oil" and offered no improvement. On a trumpet newsgroup, I've seen postings by people who claimed that their horns sounded much better after cryogenic treatment. The general claim was that the treatment aligned the molecules and relieved internal stresses in the metal, or something like that. Perhaps you metal guys have some insight on cryogenic treatment as it pertains to brass, and brass instruments in particular. Any basis in fact, or pure BS? -- Jedd Haas - Artist http://www.gallerytungsten.com http://www.epsno.com |
#2
|
|||
|
|||
"Jedd Haas" wrote: (clip)Any basis in fact, or pure BS? ^^^^^^^^^^^ Has anyone ever done a blind test on this? This is the kind of thing that can be enormously influenced by the placebo effect. |
#3
|
|||
|
|||
Cryogenics is a weird thing. The only documented eveidence of any effect is
with ferrous materials. I know of no other documentation for having any molecular effects on any other material: non ferrous or plastics, etc. Having said that, though, I did a little experiment that I learned of in other newsgroups. I made two CD copies of a disc and simply froze one in a regular freezer for a few days. The frozen disc had a bit more bass and resolution. The effect seems permanent, too and this is without going to cryogenic temperatures. Perhaps it affects the thermal dyes used in a CD-R, but even in blind tests, most listeners can pick out the difference. Seems to me we need a little more research into this (probably means too much $'s, though). Oh yeah, I know a trombone player who had his instrument cryo'd. His comment was that it didn't affect the sound so much as it affected the slide action. Stu Stu |
#4
|
|||
|
|||
On Thu, 10 Mar 2005 20:34:22 GMT, audiodir wrote:
Cryogenics is a weird thing. The only documented eveidence of any effect is with ferrous materials. I know of no other documentation for having any molecular effects on any other material: non ferrous or plastics, etc. Having said that, though, I did a little experiment that I learned of in other newsgroups. I made two CD copies of a disc and simply froze one in a regular freezer for a few days. The frozen disc had a bit more bass and resolution. The effect seems permanent, too and this is without going to cryogenic temperatures. There are more than a few reasons to doubt that there is a real change here. And it's trivial to extract the song files from that CD, and compare the two disks. Perhaps it affects the thermal dyes used in a CD-R, but even in blind tests, most listeners can pick out the difference. Do you have an actual study to back that up? The problem is, it's a digital storage medium. They're either zeros or ones. The bits aren't going to be come zero-er, or one-er, just because it's been chilled, and if the bits get flipped, that's going to induce noise, not "a bit more bass and resolution". I won't even get into the error correction bits built into the format that also prevent this sort of thing from being plausible. People were convinced that green magic marker around the edge of a CD made them sound better too, but that also isn't going to change zeros into ones, and even if it could, it wouldn't do them in an audibly pleasing and predictable way. Seems to me we need a little more research into this (probably means too much $'s, though). You could do it for free, with a Mac and iTunes. Or maybe even the PC version of iTunes. Or, if you agree that a CD-R is a digital storage medium, fill it with whatever data you want, freeze one, don't freeze the other, and run an MD5 checksum of both disks to verify that there's no difference. Hard to say about the brass instruments. Without a controlled experiment, people are likely to hear what they think they want to hear, just like your frozen CD-R's. Dave Hinz |
#5
|
|||
|
|||
Actually, checksum didn't reveal any difference. I doubt if the cold
treatment did anything to the dyes, but I inserted that coment just in case someone would speculate about the dye lots. I actually beleve it is doing something to the plastic. Stu |
#6
|
|||
|
|||
On Thu, 10 Mar 2005 21:00:00 GMT, audiodir wrote:
Actually, checksum didn't reveal any difference. Well then. I doubt if the cold treatment did anything to the dyes, but I inserted that coment just in case someone would speculate about the dye lots. I actually beleve it is doing something to the plastic. If the checksum didn't change, the data didn't change. If the data didn't change, you're reading it through the plastic, from the die, so nothing changed. You've already disproven it yourself. |
#7
|
|||
|
|||
In article , Jedd Haas says...
The general claim was that the treatment aligned the molecules and relieved internal stresses in the metal, or something like that. Of course it made them 'sound better.' They paid a lot of money for that to happen, right? So they'll sound 'better' afterwards because all that money went in there for *something*. Jim -- ================================================== please reply to: JRR(zero) at pkmfgvm4 (dot) vnet (dot) ibm (dot) com ================================================== |
#8
|
|||
|
|||
I tend to think the only difference would be that their lips would
freeze to the 'treated' horns. -- Fred R ________________ Drop TROU to email. |
#9
|
|||
|
|||
jim rozen writes: In article , Jedd Haas says... The general claim was that the treatment aligned the molecules and relieved internal stresses in the metal, or something like that. Of course it made them 'sound better.' They paid a lot of money for that to happen, right? So they'll sound 'better' afterwards because all that money went in there for *something*. Jim I want to get into the business of making $200 wooden knobs to go on the amp volume controls for the audio freaks. Beats working for a living... Gregm |
#10
|
|||
|
|||
"audiodir" skrev i en meddelelse news:ib2Yd.77923$uc.67663@trnddc08... Cryogenics is a weird thing. The only documented eveidence of any effect is with ferrous materials. I know of no other documentation for having any molecular effects on any other material: non ferrous or plastics, etc. I bought some cryo-treated strings for my bassguitar some years ago... They were just as expensive as my usual brand ( DR ) but lasted only half the time before the sound went "dull".. Having said that, though, I did a little experiment that I learned of in other newsgroups. I made two CD copies of a disc and simply froze one in a regular freezer for a few days. The frozen disc had a bit more bass and resolution. The effect seems permanent, too and this is without going to cryogenic temperatures. Perhaps it affects the thermal dyes used in a CD-R, but even in blind tests, most listeners can pick out the difference. Seems to me we need a little more research into this (probably means too much $'s, though). Errh.. Did you use directional cables* in your listening setup ? I would really love an explanation of how freezing a CDR should change the quality of the sound, when played in a CD player.. You see.. CDs are not like old LPs... On a CD the music is stored in a digital format, much like on a floppydisc.. The way the data is stored is a string of bits ( basically 1/0 or On/Off ).. These bits shouldnt change a whole lot, even if you freeze the CD....... LPs are a completely different animal, basically printing the movement of the speakercone in the vinyl disc directly.. Oh yeah, I know a trombone player who had his instrument cryo'd. His comment was that it didn't affect the sound so much as it affected the slide action. One effect it may have is that it most likely cleans or atleast loosens a lot of the "gunk" that sticks to the brass-- That, in turn, could affect the sound of the instrument.. /peter ( still wiping coffee off the keyboard after reading the CD story ) |
#11
|
|||
|
|||
Actually, I think the weird thing here is what people in audio will
think they can hear. Don't do a blind test; do a double blind test. Then the truth comes out. Steve audiodir wrote: Cryogenics is a weird thing. The only documented eveidence of any effect is with ferrous materials. I know of no other documentation for having any molecular effects on any other material: non ferrous or plastics, etc. Having said that, though, I did a little experiment that I learned of in other newsgroups. I made two CD copies of a disc and simply froze one in a regular freezer for a few days. The frozen disc had a bit more bass and resolution. The effect seems permanent, too and this is without going to cryogenic temperatures. Perhaps it affects the thermal dyes used in a CD-R, but even in blind tests, most listeners can pick out the difference. Seems to me we need a little more research into this (probably means too much $'s, though). Oh yeah, I know a trombone player who had his instrument cryo'd. His comment was that it didn't affect the sound so much as it affected the slide action. Stu Stu |
#12
|
|||
|
|||
I have found that if I freeze a book before I read it, it becomes more
meaningful. |
#13
|
|||
|
|||
"Steve Smith" wrote in message
... Actually, I think the weird thing here is what people in audio will think they can hear. Don't do a blind test; do a double blind test. Then the truth comes out. Let me put it this way. "DBT" is strictly taboo on certain audio forums. Seriously! Actual moderated groups, cannot seriously mention the subject. Tim -- "California is the breakfast state: fruits, nuts and flakes." Website: http://webpages.charter.net/dawill/tmoranwms |
#14
|
|||
|
|||
"Leo Lichtman" writes:
I have found that if I freeze a book before I read it, it becomes more meaningful. I tried freezing vodka, but I don't think it makes a better martini. Gregm |
#15
|
|||
|
|||
Actually this is a common misconception. The actual reading of the disc
is a totally analogue process. The reflected optical signal from the pits on the CD is not a strict "one" or "zero". The signal varies within a lower range which we call a "zero", and a higher range which we call a "one". There is a "guard" band between these two signal levels. It is only the digital circuitry that comes after the disc reading process that causes "binning" of the two signals into either a "one" or a "zero". It may not happen very often, but it is possible to introduce errors at the analogue reading stage. How this may or may not affect the sound I'll leave for others to debate. However, recall that many moons ago jitter was ridiculed and totally dismissed in the press as being an implausable cause for any affect on the sound. If I recall, it was one of these so- called nut case audiophiles that originally discovered and measured this effect. I do believe that low jitter circuitry is now pretty much standard in any decent quality digital equipment. Cheers. Do you have an actual study to back that up? The problem is, it's a digital storage medium. They're either zeros or ones. |
#16
|
|||
|
|||
"C.A. Decker" writes: Actually this is a common misconception. The actual reading of the disc is a totally analogue process. The reflected optical signal from the pits on the CD is not a strict "one" or "zero". The signal varies within a lower range which we call a "zero", and a higher range which we call a "one". There is a "guard" band between these two signal levels. It is only the digital circuitry that comes after the disc reading process that causes "binning" of the two signals into either a "one" or a "zero". It may not happen very often, but it is possible to introduce errors at the analogue reading stage. How this may or may not affect the sound I'll leave for others to debate. However, recall that many moons ago jitter was ridiculed and totally dismissed in the press as being an implausable cause for any affect on the sound. If I recall, it was one of these so- called nut case audiophiles that originally discovered and measured this effect. I do believe that low jitter circuitry is now pretty much standard in any decent quality digital equipment. Cheers. Any bits flipping because of analog effects is an error reading the medium. It happens, which is why there are ecc and checksums. Their job is to detect and/or recover from the error- the same kind of thing is done in more expensive computer memory to help recover from cosmic rays flipping bits here and there, among other things. However, how the audio nuts go from arbitrary bits flipping between states to an organized effect like "better bass" instead of a simple increase in random noise is very curious to say the least. Particularly when such claims are not backed up by before and after disc images showing the beneficial effects and organized data tables showing the effect as a function of time-at-temperature. Gregm |
#17
|
|||
|
|||
The instruments usually sound a lot better when warmed up if the musicians
also got the cryogenic treatment at the same time. Must be the thawing effect... kind of spring-like. :-) cheers T.Alan "Jedd Haas" wrote in message ... On another newsgroup, the subject of cryogenic treatment of brass instruments came up, with the claim that it was "snake oil" and offered no improvement. On a trumpet newsgroup, I've seen postings by people who claimed that their horns sounded much better after cryogenic treatment. The general claim was that the treatment aligned the molecules and relieved internal stresses in the metal, or something like that. Perhaps you metal guys have some insight on cryogenic treatment as it pertains to brass, and brass instruments in particular. Any basis in fact, or pure BS? -- Jedd Haas - Artist http://www.gallerytungsten.com http://www.epsno.com |
#18
|
|||
|
|||
"Greg Menke" wrote in message
... | "Leo Lichtman" writes: | | I have found that if I freeze a book before I read it, it becomes more | meaningful. | | I tried freezing vodka, but I don't think it makes a better martini. | | Gregm No, but if you mix your screwdriver as follows it's even better: 1/3 ice 1/3 orange juice concentrate, still frozen 1/3 vodka from a freezer. Instead of a screwdriver it's called a thigh spreader. Jimmy Buffet had some similar comments about the concept, but I can't quote them exactly. Something like "I can't feel a thing!" where the response is "That's the plan, baby!" That's as much of that cryo crap as I'll ever vouch for! |
#20
|
|||
|
|||
I thought that in the case of audio media the sections of the disk
normally used on data CDs for error checking and correction data were instaed used to store extra audio info hence no error correction on audio CDs. Greg Menke wrote: "C.A. Decker" writes: Actually this is a common misconception. The actual reading of the disc is a totally analogue process. The reflected optical signal from the pits on the CD is not a strict "one" or "zero". The signal varies within a lower range which we call a "zero", and a higher range which we call a "one". There is a "guard" band between these two signal levels. It is only the digital circuitry that comes after the disc reading process that causes "binning" of the two signals into either a "one" or a "zero". It may not happen very often, but it is possible to introduce errors at the analogue reading stage. How this may or may not affect the sound I'll leave for others to debate. However, recall that many moons ago jitter was ridiculed and totally dismissed in the press as being an implausable cause for any affect on the sound. If I recall, it was one of these so- called nut case audiophiles that originally discovered and measured this effect. I do believe that low jitter circuitry is now pretty much standard in any decent quality digital equipment. Cheers. Any bits flipping because of analog effects is an error reading the medium. It happens, which is why there are ecc and checksums. Their job is to detect and/or recover from the error- the same kind of thing is done in more expensive computer memory to help recover from cosmic rays flipping bits here and there, among other things. However, how the audio nuts go from arbitrary bits flipping between states to an organized effect like "better bass" instead of a simple increase in random noise is very curious to say the least. Particularly when such claims are not backed up by before and after disc images showing the beneficial effects and organized data tables showing the effect as a function of time-at-temperature. Gregm |
#21
|
|||
|
|||
"carl mciver" writes: "Greg Menke" wrote in message ... | "Leo Lichtman" writes: | | I have found that if I freeze a book before I read it, it becomes more | meaningful. | | I tried freezing vodka, but I don't think it makes a better martini. | | Gregm No, but if you mix your screwdriver as follows it's even better: 1/3 ice 1/3 orange juice concentrate, still frozen 1/3 vodka from a freezer. Instead of a screwdriver it's called a thigh spreader. Jimmy Buffet had some similar comments about the concept, but I can't quote them exactly. Something like "I can't feel a thing!" where the response is "That's the plan, baby!" That's as much of that cryo crap as I'll ever vouch for! Gah- I like to keep my OJ for breakfast and the vodka for other times. OTOH, my brother makes a nice Raspberry Wheat beer (but flavored with real Raspberry concentrate), which is affectionately known as "liquid panty remover". Though its best served a bit warmer than cryo temps, unlike Budwiser and the rest of that ilk. Gregm |
#22
|
|||
|
|||
David Billington writes: I thought that in the case of audio media the sections of the disk normally used on data CDs for error checking and correction data were instaed used to store extra audio info hence no error correction on audio CDs. Even if thats so, then why would the alleged bit flipping happen to the digitized audio in such an organized way as to improve the sound instead of introducing noise? I think you should run the test as a previous poster suggested, take an arbitrary test disc, copy off its image, freeze the disc & confirm that it sounds better, then read the image again and run a md5 against both images. If the checksums are the same, you are imagining things. If they are not, then a per-track investiagtion would be the next step to find out where the errors occurred and how many occurred. Its just a matter of repetition to track down which samples have flipped bits & evaluate what the impact on the reproduced sound is. The image reading and comparing is easier in Linux/bsd/Solaris and probably MacOS, but I imagine Windows can do it if you track down the software. Gregm |
#23
|
|||
|
|||
In article , Greg Menke says...
I have found that if I freeze a book before I read it, it becomes more meaningful. I tried freezing vodka, but I don't think it makes a better martini. Vodka ice cubes. You put them *in* the martini. Jim -- ================================================== please reply to: JRR(zero) at pkmfgvm4 (dot) vnet (dot) ibm (dot) com ================================================== |
#24
|
|||
|
|||
Greg Menke wrote: David Billington writes: I thought that in the case of audio media the sections of the disk normally used on data CDs for error checking and correction data were instaed used to store extra audio info hence no error correction on audio CDs. Even if thats so, then why would the alleged bit flipping happen to the digitized audio in such an organized way as to improve the sound instead of introducing noise? I think you should run the test as a previous poster suggested, take an arbitrary test disc, copy off its image, freeze the disc & confirm that it sounds better, then read the image again and run a md5 against both images. If the checksums are the same, you are imagining things. If they are not, then a per-track investiagtion would be the next step to find out where the errors occurred and how many occurred. Its just a matter of repetition to track down which samples have flipped bits & evaluate what the impact on the reproduced sound is. The image reading and comparing is easier in Linux/bsd/Solaris and probably MacOS, but I imagine Windows can do it if you track down the software. Gregm I had a quick look and my comment was not entirely correct. On a data disc you get a 512 byte sector followed by about 150 bytes IIRC of ECC data, if the ECC can't correct it you get that nasty situation were the data is corrupted and possibly irrecoverable. The audio CD differs in the data layout but still has means of some ECC although the error handling strategies still play the CD trying to minimise the impact of errors on the reproduced audio. From this I could see that some method of uniformly tweeking the CD might have an effect on the audio output. I agree though that if you run a MD5 on the disc image and they are the same then I can't see how the freezing has effected anything. |
#25
|
|||
|
|||
Greg Menke wrote:
"C.A. Decker" writes: Actually this is a common misconception. The actual reading of the disc is a totally analogue process. The reflected optical signal from the pits on the CD is not a strict "one" or "zero". The signal varies within a lower range which we call a "zero", and a higher range which we call a "one". There is a "guard" band between these two signal levels. It is only the digital circuitry that comes after the disc reading process that causes "binning" of the two signals into either a "one" or a "zero". It may not happen very often, but it is possible to introduce errors at the analogue reading stage. How this may or may not affect the sound I'll leave for others to debate. However, recall that many moons ago jitter was ridiculed and totally dismissed in the press as being an implausable cause for any affect on the sound. If I recall, it was one of these so- called nut case audiophiles that originally discovered and measured this effect. I do believe that low jitter circuitry is now pretty much standard in any decent quality digital equipment. Cheers. Any bits flipping because of analog effects is an error reading the medium. It happens, which is why there are ecc and checksums. Their job is to detect and/or recover from the error- the same kind of thing is done in more expensive computer memory to help recover from cosmic rays flipping bits here and there, among other things. However, how the audio nuts go from arbitrary bits flipping between states to an organized effect like "better bass" instead of a simple increase in random noise is very curious to say the least. A puzzler, but I can think of a few possibilities. Unlike data, the error correction for music is CDs is not absolute - if there is an error the circuitry will try and play around it - and the method used to approximate that might well involve smoothing the sound out over a certain interval of time with interpolated values, causing better bass and worse treble response; or if not actually improving the bass and/or overall quality, causing a relatively better bass/treble quality ratio. In the studio I believe it is also an occasional practice to add noise at innaudible level to ameliorate distortion caused by the notes being exact fractions of the sampling ratio, and some non-linearities in the overall response, causing and a slight increase in both actual and perceived quality. Note that these are pure speculation, and probably wrong. Any effect, if there is an actual effect, may well be caused by other processes. They are just to show that it might be possible. Like CA Decker's Jitter story, his sort of thing happens in musical reproduction technology - for instance when "they" said "44.1 kHz is enough, you won't get better fidelity with a higher sampling rate" and justified it as mathematically proven under Nyquist theory, they were subtly wrong - it's fine and true for the reproduction of constant tones, but the theory falls down when reproducing changing notes, which after all are what music is composed of, and when some implicit assumptions about your reproduction equipment are untrue in practice. I'm not buying silver speaker cables though! More, a general improvement in faithfulness to the original causing "better bass" is much easier to imagine possible reasons for. -- Peter Fairbrother The new moon is rising the axe of the thunder is broken as never was not since the flood nor yet since the world began |
#26
|
|||
|
|||
In article ,
Peter Fairbrother wrote: I'm not buying silver speaker cables though! Of course not. You'll want the Phacetious Audio quicksilver liquid speaker connectors, where mercury is pumped through hoses to carry your speaker signal. The sound is "liquid" and "flowing" in subtle audiophile ways that mere engineers can't quantify. Pumping the mercury, rather than letting it sit still in the hose adds subtle "motion" to the sound. As if that wasn't reason enough to spend $50,000 on a set of ten foot speaker cables, you'll have a hazardous waste site when your dog chews on the hoses. What could be better? Sign up now. Get your deposits in and be among the phirst to be phleeced... -- Cats, Coffee, Chocolate...vices to live by |
#27
|
|||
|
|||
(somebody wrote, but this guy deleted the attributes for while top-posting,)
Do you have an actual study to back that up? The problem is, it's a digital storage medium. They're either zeros or ones. On Fri, 11 Mar 2005 02:49:45 GMT, C.A. Decker wrote: Actually this is a common misconception. What, that a digital disk is digital? The actual reading of the disc is a totally analogue process. The reflected optical signal from the pits on the CD is not a strict "one" or "zero". The signal varies within a lower range which we call a "zero", and a higher range which we call a "one". There is a "guard" band between these two signal levels. Yes. And if the bit is above the upper limit, it's a one. If it's below the lower limit, it's a zero. If it's in the middle, it's an _error_, and the error correction bits tell your reader how to correct it. It is only the digital circuitry that comes after the disc reading process that causes "binning" of the two signals into either a "one" or a "zero". Right, so far. It may not happen very often, but it is possible to introduce errors at the analogue reading stage. Of course it is. Scratches, bad pits, and so on. That's why the correction bits are there. How this may or may not affect the sound I'll leave for others to debate. If you're taking signal and turning it into something else, it's called _noise_. However, recall that many moons ago jitter was ridiculed and totally dismissed in the press as being an implausable cause for any affect on the sound. If I recall, it was one of these so- called nut case audiophiles that originally discovered and measured this effect. I do believe that low jitter circuitry is now pretty much standard in any decent quality digital equipment. Cheers. A lot of really stupid ideas have been ridiculed as well. I think this is one thing that has always turned me off from the "high end audiophile" people - they can't quantify their theories, and try to justify them as good because some other theory turned out to be good when it was initially ridiculed. For every person thought to be a fool who turns out to be a genius, there are 999 who really, really were just fools. |
#28
|
|||
|
|||
On 11 Mar 2005 07:51:55 -0500, Greg Menke wrote:
The image reading and comparing is easier in Linux/bsd/Solaris and probably MacOS, Yes, MacOS is a FreeBSD system, so all of the tools you'd expect on any other Unix box are there, and all of whatever you'd want to build builds just fine. |
#29
|
|||
|
|||
Greg Menke wrote:
I tried freezing vodka, but I don't think it makes a better martini. That's because a better Martini is made with gin. -G |
#30
|
|||
|
|||
Dave Hinz writes: On 11 Mar 2005 07:51:55 -0500, Greg Menke wrote: The image reading and comparing is easier in Linux/bsd/Solaris and probably MacOS, Yes, MacOS is a FreeBSD system, so all of the tools you'd expect on any other Unix box are there, and all of whatever you'd want to build builds just fine. I've had the occasional problem compiling things on OS X, different ranlib defaults for instance, but its loads easier to manage than Windows. Gregm |
#31
|
|||
|
|||
Peter Fairbrother writes: Any bits flipping because of analog effects is an error reading the medium. It happens, which is why there are ecc and checksums. Their job is to detect and/or recover from the error- the same kind of thing is done in more expensive computer memory to help recover from cosmic rays flipping bits here and there, among other things. However, how the audio nuts go from arbitrary bits flipping between states to an organized effect like "better bass" instead of a simple increase in random noise is very curious to say the least. A puzzler, but I can think of a few possibilities. Unlike data, the error correction for music is CDs is not absolute - if there is an error the circuitry will try and play around it - and the method used to approximate that might well involve smoothing the sound out over a certain interval of time with interpolated values, causing better bass and worse treble response; or if not actually improving the bass and/or overall quality, causing a relatively better bass/treble quality ratio. In the studio I believe it is also an occasional practice to add noise at innaudible level to ameliorate distortion caused by the notes being exact fractions of the sampling ratio, and some non-linearities in the overall response, causing and a slight increase in both actual and perceived quality. Note that these are pure speculation, and probably wrong. Any effect, if there is an actual effect, may well be caused by other processes. They are just to show that it might be possible. The errored blocks of data are then missing and interpolated, which inevitably introduces noise- sort of like a fleck of dust on the LP groove. Its not just bass frequencies that are magically affected, the lost samples literally stop the speaker cone from reproducing the waveform for the duration of the error. At that point you're presumably relying on the ballistic performance of the cones to "smooth over" the lost samples and damp the sharp transitions from signal to error to signal, which doesn't sound like a good way to improve "quality ratios", whatever they are. Its effectively a decrease in the signal/noise ratio. The ECC codes are essentially data checksums, computed such that some amount of missing or corrupt data can be recovered given the surrounding data and the ECC code itself. Once the algorithm's capacity is exceeded (because of too many errored bits in the associated data block), then all you'll get out is data corrupted to some unknown extent. So from a reproduction standpoint, you get perfect data from the CD as the number of mis-read bits coming off the CD read head increases up to the point where the bit-error-rate exceeds the capacity of the ECC algorithm, then corrupt data starts coming out. By "perfect", I mean the bits read off the CD are the same as written. So, the CD format provides a limited degree of error recovery built into the format of the audio samples- it alway reads the same bit-stream until too many errors show up for the ECC to deal with. I look forward to your detailed proposition documenting how bit flips that get past the ECC algorithms improve sound, with emphasis on how CD audio formats make it possible. I'm also interested in how bit flips on dvd's and digital audio tapes improve the reproduction quality but do not corrupt things like software images or compressed data. Gregm |
#32
|
|||
|
|||
I came across this site which give some background to the technology
behind the audio CD. Quite a bit of effort goes into trying to make sure you don't hear the result of even fairly severe CD damage. Greg Menke wrote: Peter Fairbrother writes: Any bits flipping because of analog effects is an error reading the medium. It happens, which is why there are ecc and checksums. Their job is to detect and/or recover from the error- the same kind of thing is done in more expensive computer memory to help recover from cosmic rays flipping bits here and there, among other things. However, how the audio nuts go from arbitrary bits flipping between states to an organized effect like "better bass" instead of a simple increase in random noise is very curious to say the least. A puzzler, but I can think of a few possibilities. Unlike data, the error correction for music is CDs is not absolute - if there is an error the circuitry will try and play around it - and the method used to approximate that might well involve smoothing the sound out over a certain interval of time with interpolated values, causing better bass and worse treble response; or if not actually improving the bass and/or overall quality, causing a relatively better bass/treble quality ratio. In the studio I believe it is also an occasional practice to add noise at innaudible level to ameliorate distortion caused by the notes being exact fractions of the sampling ratio, and some non-linearities in the overall response, causing and a slight increase in both actual and perceived quality. Note that these are pure speculation, and probably wrong. Any effect, if there is an actual effect, may well be caused by other processes. They are just to show that it might be possible. The errored blocks of data are then missing and interpolated, which inevitably introduces noise- sort of like a fleck of dust on the LP groove. Its not just bass frequencies that are magically affected, the lost samples literally stop the speaker cone from reproducing the waveform for the duration of the error. At that point you're presumably relying on the ballistic performance of the cones to "smooth over" the lost samples and damp the sharp transitions from signal to error to signal, which doesn't sound like a good way to improve "quality ratios", whatever they are. Its effectively a decrease in the signal/noise ratio. The ECC codes are essentially data checksums, computed such that some amount of missing or corrupt data can be recovered given the surrounding data and the ECC code itself. Once the algorithm's capacity is exceeded (because of too many errored bits in the associated data block), then all you'll get out is data corrupted to some unknown extent. So from a reproduction standpoint, you get perfect data from the CD as the number of mis-read bits coming off the CD read head increases up to the point where the bit-error-rate exceeds the capacity of the ECC algorithm, then corrupt data starts coming out. By "perfect", I mean the bits read off the CD are the same as written. So, the CD format provides a limited degree of error recovery built into the format of the audio samples- it alway reads the same bit-stream until too many errors show up for the ECC to deal with. I look forward to your detailed proposition documenting how bit flips that get past the ECC algorithms improve sound, with emphasis on how CD audio formats make it possible. I'm also interested in how bit flips on dvd's and digital audio tapes improve the reproduction quality but do not corrupt things like software images or compressed data. Gregm |
#33
|
|||
|
|||
Sorry let me try that again.
I came across this site which give some background to the technology behind the audio CD. Quite a bit of effort goes into trying to make sure you don't hear the result of even fairly severe CD damage. http://www.ee.washington.edu/consele...udio2/95x7.htm Greg Menke wrote: Peter Fairbrother writes: Any bits flipping because of analog effects is an error reading the medium. It happens, which is why there are ecc and checksums. Their job is to detect and/or recover from the error- the same kind of thing is done in more expensive computer memory to help recover from cosmic rays flipping bits here and there, among other things. However, how the audio nuts go from arbitrary bits flipping between states to an organized effect like "better bass" instead of a simple increase in random noise is very curious to say the least. A puzzler, but I can think of a few possibilities. Unlike data, the error correction for music is CDs is not absolute - if there is an error the circuitry will try and play around it - and the method used to approximate that might well involve smoothing the sound out over a certain interval of time with interpolated values, causing better bass and worse treble response; or if not actually improving the bass and/or overall quality, causing a relatively better bass/treble quality ratio. In the studio I believe it is also an occasional practice to add noise at innaudible level to ameliorate distortion caused by the notes being exact fractions of the sampling ratio, and some non-linearities in the overall response, causing and a slight increase in both actual and perceived quality. Note that these are pure speculation, and probably wrong. Any effect, if there is an actual effect, may well be caused by other processes. They are just to show that it might be possible. The errored blocks of data are then missing and interpolated, which inevitably introduces noise- sort of like a fleck of dust on the LP groove. Its not just bass frequencies that are magically affected, the lost samples literally stop the speaker cone from reproducing the waveform for the duration of the error. At that point you're presumably relying on the ballistic performance of the cones to "smooth over" the lost samples and damp the sharp transitions from signal to error to signal, which doesn't sound like a good way to improve "quality ratios", whatever they are. Its effectively a decrease in the signal/noise ratio. The ECC codes are essentially data checksums, computed such that some amount of missing or corrupt data can be recovered given the surrounding data and the ECC code itself. Once the algorithm's capacity is exceeded (because of too many errored bits in the associated data block), then all you'll get out is data corrupted to some unknown extent. So from a reproduction standpoint, you get perfect data from the CD as the number of mis-read bits coming off the CD read head increases up to the point where the bit-error-rate exceeds the capacity of the ECC algorithm, then corrupt data starts coming out. By "perfect", I mean the bits read off the CD are the same as written. So, the CD format provides a limited degree of error recovery built into the format of the audio samples- it alway reads the same bit-stream until too many errors show up for the ECC to deal with. I look forward to your detailed proposition documenting how bit flips that get past the ECC algorithms improve sound, with emphasis on how CD audio formats make it possible. I'm also interested in how bit flips on dvd's and digital audio tapes improve the reproduction quality but do not corrupt things like software images or compressed data. Gregm |
#34
|
|||
|
|||
In article ,
David Billington wrote: I thought that in the case of audio media the sections of the disk normally used on data CDs for error checking and correction data were instaed used to store extra audio info hence no error correction on audio CDs. Actually, consider it the other way around: 1) The standard for CDs was for Audio *first*. 2) The audio CDs have a sector size larger than the 2048 byte size used for data CDS, and follow that with an additional chunk for The ECC (Error Checking and Correction code). 3) When used for *data* purposes, the difference between the audio and data sector sizes are used for an *additional* layer of ECC, so the data off the CD is *doubly* checked. So -- from your point of view, there is less error checking for audio purposes, but there still is quite a bit. even with audio. If it can't actually *fix* an error, it will replace the bad sample with an average of the adjacent samples, which helps hide the error, at least. So -- in any case, greater errors on the raw CD result in less perfect audio reproduction, but not the glaring errors to be found with a flipped bit (which would probably come out as a loud "click" on playback without that substitution.) Enjoy, DoN. -- Email: | Voice (all times): (703) 938-4564 (too) near Washington D.C. | http://www.d-and-d.com/dnichols/DoN.html --- Black Holes are where God is dividing by zero --- |
#35
|
|||
|
|||
Dave Hinz wrote:
The actual reading of the disc is a totally analogue process. The reflected optical signal from the pits on the CD is not a strict "one" or "zero". The signal varies within a lower range which we call a "zero", and a higher range which we call a "one". There is a "guard" band between these two signal levels. Yes. And if the bit is above the upper limit, it's a one. If it's below the lower limit, it's a zero. If it's in the middle, it's an _error_, and the error correction bits tell your reader how to correct it. The misconception mentioned earlier being that the change from "one" to "zero" is instantaneous; it takes time to make that transition, and that's an analogue thing. Like people have pointed out; loss of data is not a problem, but since in SP/DIF (the protocol behind the CD) the clock, which governs the data rate, is embedded into the data itself, the rate at which those transitions occur is extremely important, and any ambivalence as to where a transition takes place potentially forces the system to second-guess itself. This is called jitter, is usually measured in nanoseconds, and it sounds like, well, an old cheap CD player. I suppose that getting those pesky little holes in the CD surface to have better defined edges might make CDs sound better. However, I seriously doubt that freezing them will do that ( although I'm open to scientific explanation) Hans -- This is a non-profit organization; we didn't plan it that way, but it is ===================================== (remove uppercase trap, and double the number to reply) |
#36
|
|||
|
|||
Greg Menke wrote:
Peter Fairbrother writes: Any bits flipping because of analog effects is an error reading the medium. It happens, which is why there are ecc and checksums. Their job is to detect and/or recover from the error- the same kind of thing is done in more expensive computer memory to help recover from cosmic rays flipping bits here and there, among other things. However, how the audio nuts go from arbitrary bits flipping between states to an organized effect like "better bass" instead of a simple increase in random noise is very curious to say the least. A puzzler, but I can think of a few possibilities. Unlike data, the error correction for music is CDs is not absolute - if there is an error the circuitry will try and play around it - and the method used to approximate that might well involve smoothing the sound out over a certain interval of time with interpolated values, causing better bass and worse treble response; or if not actually improving the bass and/or overall quality, causing a relatively better bass/treble quality ratio. In the studio I believe it is also an occasional practice to add noise at innaudible level to ameliorate distortion caused by the notes being exact fractions of the sampling ratio, and some non-linearities in the overall response, causing and a slight increase in both actual and perceived quality. Note that these are pure speculation, and probably wrong. Any effect, if there is an actual effect, may well be caused by other processes. They are just to show that it might be possible. The errored blocks of data are then missing and interpolated, which inevitably introduces noise- sort of like a fleck of dust on the LP groove. Its not just bass frequencies that are magically affected, the lost samples literally stop the speaker cone from reproducing the waveform for the duration of the error. I don't understand what you mean. If you mean the cone stops, it doesn't. If you mean the cone stops getting a signal, it doesn't. What does happen is the signal sent to the amp may (may, not will) be different to the intended digital signal. At that point you're presumably relying on the ballistic performance of the cones to "smooth over" the lost samples and damp the sharp transitions from signal to error to signal, No, you are relying on the interpolation circuitry. This "guesses" a value for the signal and inserts into the stream going to the d/a converter. The aural effect depends to a great extent on the actual circuitry of the player, and this is done in many different ways. A simple interpolator might half the signal between the before and the after samples, or add the difference between s-2 and s-1 to s-1, but a good predictive interpolator can involve really very complex DSP circuitry (or software). which doesn't sound like a good way to improve "quality ratios", whatever they are. Its effectively a decrease in the signal/noise ratio. A decrease in treble s/n without a decrease in bass s/n might well be perceptually seen as an improvement in bass response. This could come about because eg the interpolation circuitry was better at interpolating bass frequencies. The ECC codes are essentially data checksums, computed such that some amount of missing or corrupt data can be recovered given the surrounding data and the ECC code itself. Once the algorithm's capacity is exceeded (because of too many errored bits in the associated data block), then all you'll get out is data corrupted to some unknown extent. No, what you get is a "data corrupted" signal, and the active interpolation circuitry takes over and provides the next sample to the d/a converter. So from a reproduction standpoint, you get perfect data from the CD as the number of mis-read bits coming off the CD read head increases up to the point where the bit-error-rate exceeds the capacity of the ECC algorithm, then corrupt data starts coming out. By "perfect", I mean the bits read off the CD are the same as written. So, the CD format provides a limited degree of error recovery built into the format of the audio samples- it alway reads the same bit-stream until too many errors show up for the ECC to deal with. To wander slightly for a moment, jitter is not always relevant nowadays - most modern players will store several seconds of data in ram in case of temporary loss of datastream due to mechanical shock. The feed to the d/a converter comes from the ram store, not the reading in process. Some of the better players will try to reread any corrupted blocks - computer "ripper" software especially does this, that's why it can take a good ripper a long time to read a badly-scratched disk - but the ECC on music cd's is such that unless both the CD and the player are top quality (and usually not even then) the MD5's of the data stream going to the d/a converter in two seperate playings of the same CD on the same player will not match. You simply do not get 100% accurate bit-for-bit reproduction or copying of music CD's, at least not without multiple reading. That's why people use (or used to use) expensive CD players for audio, and CD caddies, and so on. You can get that kind of accurate copying, but you have to record the music as data files, with their extra ECC. I look forward to your detailed proposition documenting how bit flips that get past the ECC algorithms They don't exist - and iirc all single bit flips are detectable and correctable anyway. The ECC is done in such a way that a "data corrupt, interpolate instead" signal replaces any putative bit flips - the ECC detects that the signal is corrupt, and that it doesn't have the ability to correct it. The interpolator then comes into action. If the signal corruption is in the upper half, it is quite likely that the interpolator will come up with a signal identical to the original signal. It's a bit late now, I may write more tomorrow. -- Peter Fairbrother |
#37
|
|||
|
|||
On Sat, 12 Mar 2005 03:35:14 +0100, Hans van Dongen wrote:
Dave Hinz wrote: Yes. And if the bit is above the upper limit, it's a one. If it's below the lower limit, it's a zero. If it's in the middle, it's an _error_, and the error correction bits tell your reader how to correct it. The misconception mentioned earlier being that the change from "one" to "zero" is instantaneous; it takes time to make that transition, and that's an analogue thing. Right, but let's keep this in mind - the OP has agreed that an MD5 checksum of the frozen, and the unfrozen CD, were the same. Theoretical this, that, and the other aside, all of that analog and interpretation layer isn't relevant, when the digital result of all those spinning, reflecting, jittering, and whatever else-ing end up giving you exactly the same data. If you've got the same data, it's going to sound _exactly the same_, because it is. |
#38
|
|||
|
|||
I never saw the original post on this one, just replies about CDs.
Brass instruments do not respond to Cryogenic treatment... rather, they do respond but somewhat like a "Baptist Fish" The changes do take place but disappear almost as soon as the instrument thaws out. Save your money and take some extra lessons instead. LB |
#39
|
|||
|
|||
Leonard & Peggy Brown wrote:
I never saw the original post on this one, just replies about CDs. Brass instruments do not respond to Cryogenic treatment... rather, they do respond but somewhat like a "Baptist Fish" The changes do take place but disappear almost as soon as the instrument thaws out. Save your money and take some extra lessons instead. Persons of a more engineering or even metalworking bent _do_ cryogenically treat several metals to change their properties - why not brass? But why do they cryogenically treat them in the first place? Does it work with brass (or perhaps even monel, which somne "brass" instruments are madeof?)? I don't know, so I can't answer, but they do seem to do it for a reason. Anyone? Maybe it's like cryogenically treating CD's? -- Peter Fairbrother |
#40
|
|||
|
|||
"Peter Fairbrother" wrote in message ... But why do they cryogenically treat them in the first place? Does it work with brass (or perhaps even monel, which somne "brass" instruments are madeof?)? As far as I know monel is only used in the valves. Randy |
Reply |
|
Thread Tools | Search this Thread |
Display Modes | |
|
|
Similar Threads | ||||
Thread | Forum | |||
Fast Brass Antiquing | Woodworking | |||
brass musical note source? | Metalworking | |||
melting brass padlocks | Metalworking | |||
Brass drill bit (for lead) | Metalworking | |||
Alternate treatment for depression | Electronics Repair |