Scott, I saw it as a distraction, but I got triggered by Thorsten's remark that he has done this stuff for a quadrillion decennia and that microphonics was the problem he had solved before the advent of daylight, etc. Which it was bloody not. It gets tiring.
That was my playfull side, on a more serious note: the ESR/DA and where does it come from, this is where it all becomes interesting to me.
ESR and DA might be distinguishable after all, but I like to visualise the mechanisms at work before coming to definite conclusions, and I am still in the middle of it. Part of where I am right now:
-ESR is at least partly caused by the impendance of the plates and the way the are connected to the leads. Mica's are completely different with respect to construction from foils.
-How to distinguish: Johnson noise for the ESR part would have a positive and linear temperature correlation. For DA, my gut feeling is that this might be more complicated. Perhaps SY knows if measurements have been done on DA of dielectrics as a function of temperature. Must have been done somewhere, someplace. If they don't correlate straight, there would be a way to differentiate between the two effects.
vac
Scott: I'll email you my cell number in case you don't have it. When you get here, I'll draw you some pictures of what I *think* is going on mechanistically.
Vac: Yes, DA vs temperature has been measured; Morgan Jones sent me a copy of an interesting review. I'll see if I can dig it up and give you a cite. Bottom line: DA vs T depends on the type of dielectric, especially when it's a polymer getting near the Tg. And either there's a whole lot of literature on the Johnson noise from DA or there's none, depending on which post you read here.
Vac: Yes, DA vs temperature has been measured; Morgan Jones sent me a copy of an interesting review. I'll see if I can dig it up and give you a cite. Bottom line: DA vs T depends on the type of dielectric, especially when it's a polymer getting near the Tg. And either there's a whole lot of literature on the Johnson noise from DA or there's none, depending on which post you read here.
No I don't think the difference between CD and SACD or Bluray is a small difference. It's all about what the difference means to person as to how small it is.
I'm not concerned about comparisons of SACD to CD on mid-fi equipment. I don't listen on a mid-fi system when I do serious listening. I have my MP3 player when I do casual listening.
Anyway, this is just another subjectivist vs. objectivist argument. The LP sounds better to me than CD. Sure I hear faults with the LP like distortion, but it's not something that bothers me. The LP has a higher dynamic range than the CD in the high frequencies, due to the RIAA curve you mention.
The ears are very good at filtering out noise as well, especially pleasant noise, and there can also be significant recorded signal below the noise in analog recordings.
Many find that the LP sounds more like what they hear live, than with the CD.
Apparently the errors the LP makes bother you much more than me. That's fine. There's no right or wrong about it, just a matter of what you prefer.
John
Well, to save my time during Post synchronization sessions, i had made measurements of envelopes of the same sentence said by an very "constant ?" actor at different distances, outside and inside in a very dump studio. And designed an equalizer with a unique button to adjust this distance (and a switch inside/outside). And i can tell you the response curve is very affected.
...
" I suspect that your anechoic chamber just wasn't built with the same high budget.."
LNE (France) > 500M²
I suspect, on my side, that Bell was not the inventor of the telephone ;-)
I was not arguing one sound better than an other. This kind of argues are subjective. I was telling that, comparing a master to his vinyl copy reveals a lot of differences, and CD or blue ray can be very close and bit copies absolutely identical to the original. When the differences are less than the ones i can produce between two mixing studio sessions, i do not complain too much about.
Maybe you have a bad DA converter on your CD player ?
There is a great misunderstanding and confusions bettween "good sounding", and transparent reproducing.
And, yes, as far i'm concerned, i prefer to listen to a good tune on MP3 than to listen to impressive little samples on high end equipments. I'm not an audiophile.
Last edited:
That is not true. The CD medium has higher dynamic range than LP at any frequency. Typical pop CDs simply do not take advantage of the available dynamic range, but that's not the fault of the medium.
As Chris Hornbeck pointed out: It is fairly easy to record vinyl to digital and burn a CD-R that sounds identical to the record. Thus, CD certainly has at least as much dynamic range as LP (and actually more). This is true not only in recording, but also in live mixing and sound reinforcement situations where there is digital processing between the LP and the speakers. On a proper system, nobody can tell that it's not unadulterated vinyl.
Also, the RIAA curve ideally restores the sound to its original response, because the same curve is applied in reverse ahead of the cutting lathe. RIAA is not a single-ended process. The end result is to filter the surface noise of the record without filtering the musical content. But that doesn't really expand the dynamic range of any frequency range of the original content, just the surface noise. You might say that the raw, undecoded audio on the record has more amplitude in the higher frequencies, but I hope you're not listening to vinyl without the RIAA decoder - that would sound worse than the worst CD.
Note that CD has a rough equivalent of RIAA, called pre-emphasis. This allows the noise floor to be reduced by an output filter in a similar way to vinyl. It's largely unused because it typically is not necessary.
This is quite true. It seems to me that when it comes to amplifier design, it makes sense to pay a little more attention to removing distortion rather than noise, because pleasant noise can become subconscious where certain distortions cannot. But it does all come down to the types of noise and the types of distortion.
Analog magnetic tape around 60db. LP, well, it is different as noise is not linear, but little dynamic at high frequencies due to distortions (that's why peak compressions needed during mastering), and a surface rubbing noise, to compare with > 96db thd+noise of a CD ? But i agree, Vinyl sound is quite "natural"... Not transparent at all, but natural.
Anyway, as says rsdio, you can copy your LP with a major difference: yourCD copy will be better than the LP after 10 runs of it on your turntable. ;-)
In fact, there too is a misunderstanding. Analog tapes as well as vinyl shows an increasing distortion with the level, specially on transients. (when CDs stays very low distortion till peaking). We have to notice that, recording at 0 Db on an analog tape (looking at a vu meter) will bring the peaks to ~+8db, and at this level treble are reduced by the saturation of the tape at high frequency. The response curve is no more linear.
That why Lps gives the impression to sound louder, warmer, some can feel more dynamic, but it is distortion, illusion.
I would like to add that the main reason people thought that Cds where too much brilliant, at the beginning of the Digital story, is that studios (magic rooms) had their monitors per-compensated for the losses of treble in LPs (so sound engineers have to add treble in their mixes for good listening) . Working in the same studios, with the linearity of the digital and there where too much treble on the master, for sure.
Last edited:
I really find much just said as rumor, and half accurate. As a designer of analog master recorders, I have found analog recording superior, subjectively, to digital recording, from the very beginning, until today. Much said in the last few messages is self-serving 'myth'. Perhaps we should talk about analog recording, its true limits, and its extra qualities.
"Subjectively", i agree. The question is, omho, what you consider a recorder for. Helping to create good sound or beeing transparent and reliable ? In other word, if you use your tape recorder as a effect, i agree.
I used to over modulate kick drums on analog tapes. and get a very nice and dynamic sound. It recreate in an artificial way a realistic hit of the "hammer ?"on the skin's drum. I'am unable to reproduce that on digital equipment, for sure.
Same remark about mixing desks. I can fight with huge corrections on a digital desk without the changes i want on the sound 'personality ' (phases ? ). So i prefer for that analog correctors. But when i just want to filter, to clean a sound, i prefer DSPs, because they don't change the nature of the instrument. Same remarks with my good old UREI compressors, old tube Neumann, they really have a personality ;-) It is about Painting VS photography.
Did-you agree ?
Producers have often, in Europa, very bad musical practices: re-recordings. Bass+drums at first, then all the other instruments one after the other. Session after session you can hear your drums dynamics slowly decrease on analog tapes. Drums stays fresh on Hard disks. Last, listeniing recently to old 30 years master's copy at home, i was desperate. Looking in the bottom of the box if the sound was fallen there. (Yes it is worse on old DATs, you can read the user manual with the light of the error diode.).
"Rumor" was not nice. That was my life, and, like you, i can design and build all the studio equipments, including analog recorders, and i know what's happens at 10 000 hz because that is the way i tune the polarization level.
Last edited:
I have worked with Direct Disc recording, multichannel 15 ips analog recording 30 ips, 2 channel recording and designed and built the electronics for each. I have also made studio boards for Direct Disc and analog recording.
The real limit for analog recording is 80 dB (CCIR weighted), the effective bandwith will exceed 20KHz, and can be as high as 45KHz, without too much trouble, the distortion continually drops with level, for real, in analog tape recording, and it is almost impossible to get the same subtle information that analog tape gives from digital.
The real limit for analog recording is 80 dB (CCIR weighted), the effective bandwith will exceed 20KHz, and can be as high as 45KHz, without too much trouble, the distortion continually drops with level, for real, in analog tape recording, and it is almost impossible to get the same subtle information that analog tape gives from digital.
My understanding is that the best tape recorders have as much as 30 kHz bandwidth, so roughly "equivalent" to a 60 kHz sampling rate.
Supposedly, the best tape achieved almost 90 dB S/N, and it should be understood that signals below -90 dB are still audible because they are not completely masked by noise. Without dither, A/D simply cannot preserve signals below the quantization floor, so -96 dB is the smallest for 16-bit. However, with proper dithering, digital can actually preserve signals well below the quantization noise floor just like analog tape.
With analog tape, you need good quality blank tape, proper biasing, heads that are free from dirt or magnetization, and careful attention to levels if you want to stay out of saturation (or even if you want saturation, you still have to pay careful attention to levels). With digital recording, you need proper clocking and well-designed circuits. Levels adjustment is as simple as avoiding red. It's not easy to obtain a good digital recording system, but once you do they're easier to maintain than analog. The problem is, if you're not an expert in analog or digital recording, then you probably will not get the ultimate results from the medium that you choose.
Even though digital is supposed to be 'easy,' there are still many ways for one system to vary from another. The dithering process of certain Sigma-delta converters can add a lot of noise in those frequencies between 20 kHz and 30 kHz where analog tape is probably a lot cleaner, relative to performance at lower frequencies. Certain converter chips work better at lower sampling rates than higher rates. Preamps and signal conditioning are key to obtaining a good recording. Another common problem with digital circuits is that the DAC may share the same power rail with the analog reconstruction filter because the design engineer did not realize that full-scale waveforms will exceed the power rails post low pass. Good DAC designs have a larger power supply rail for the reconstruction filter compared to the DAC chip's analog supply, allowing headroom for peaks without clipping. And, my biggest pet peeve is that AES3 and S/PDIF are completely flawed as a digital transports due to pushing the clock from the media source to the DAC, so if you're evaluating digital with any kind of digital audio cable then you're not really hearing the full capabilities of digital. FireWire and some recent USB interfaces have this right, but those interfaces come with their own challenges.
I could go on, but I see both technologies as less than perfect. Each requires a specific amount of esoteric knowledge to get excellent results. I do think that digital slightly edges ahead of analog in terms of being easier to operate without continual fine-tuning, but that comes after years and years of working with it, and I suppose the best analog decks can be just as easy once you get into the habit of routine maintenance.
In my experience, most early digital sounds harsh due to poor clocking of the A/D. Unfortunately, this represents information that is lost forever to noise. There is no EQ, whether static or level dependent, that can restore information that has been lost. Such is the nature of random processes.
If you have an analog master that was poorly transferred, then just repeat the process with a better A/D. But if you're talking about a live performance that was only recorded with bad digital gear, then there's no form of post-processing that can fix it and a performance can never be duplicated.
Hi,
This is untrue. LP and CD have boradly similar usable dynamic range. It is important to remember that the 96dB figure for CD is pure advertising and "milkmaid math". Once we have made the number "analog compatible" we find that CD is equivalent to an analogue source with 87dB dynamic range.
Without dither (which reduces dynamic range) it is also not possible to use the full dynamic range.
LP in analogue terms and when done right has around -70dB background noise (excluding pop's and clicks which give an artificially high reading) and 14dB Headroom (it should be noted that old Analog Systems, from FM Radio to Tape and LP all had around 14dB build in headroo, in other words 0dB on these systems is comparable to -14dBfs for digital systems), so LP's possible dynamic range is about the same or slightly larger than CD with Dither.
LP may have a small advantage at high frequencies and is theoretically capable of rendering higher frequencies than CD, but the part that may be more crucial, the HF rolloff on LP is a by far lower order filter than that required for CD.
On the other hand LP has gross distortion at high levels, several percent 2nd HD (depending on the stylus shape, cutting process and so on - including tonearm alignment) at lower frequencies, though this is nothing one does not see from Speaker drivers at high power levels...
Correct, most recordings explore neither the limits of LP's or CD's dynamic range and that includes most recordings on SACD and DVD-Audio etc.
This is not our experience with "High End" equipment (example - Kuzma Stabi Turntable [the original suspended one], Triplanar Tonearm, Koetsu Onyx Cartridge, Phonostage of my design which incorporates a 114dB DNR 24/96KHz AD with low jitter clocks and low noise PSU's etc...), we find even 24/96KHz to be merely "quite close" to the original LP, but not identical.
It is possible however, that using lower grade analog playback equipment CDR recordings suffice.
In the "old days" CD Pre-empahsis was carried out in the Analog Domain. The improvement in sound quality was palpable. Just try the same recording on a Sony PCM F1 with or without pre/de-emphasis.
Of course, in those days we had real 16 Bit converters as well, not 9 - 12 Bit equivalent resolution "low-bit" converters that get advertised as "24 Bit" or even "32 Bit" and use the additions of >> 100% fuzzy distortion for most signal levels to pretend higher resolution in the audio band.
With De-empasis moved into the digital domain all it's benefits are lost and
given the current state of AD/DA converters this is especially regrettable, as these could benefit greatly from some pre-emphasis.
Ciao T
This is untrue. LP and CD have boradly similar usable dynamic range. It is important to remember that the 96dB figure for CD is pure advertising and "milkmaid math". Once we have made the number "analog compatible" we find that CD is equivalent to an analogue source with 87dB dynamic range.
Without dither (which reduces dynamic range) it is also not possible to use the full dynamic range.
LP in analogue terms and when done right has around -70dB background noise (excluding pop's and clicks which give an artificially high reading) and 14dB Headroom (it should be noted that old Analog Systems, from FM Radio to Tape and LP all had around 14dB build in headroo, in other words 0dB on these systems is comparable to -14dBfs for digital systems), so LP's possible dynamic range is about the same or slightly larger than CD with Dither.
LP may have a small advantage at high frequencies and is theoretically capable of rendering higher frequencies than CD, but the part that may be more crucial, the HF rolloff on LP is a by far lower order filter than that required for CD.
On the other hand LP has gross distortion at high levels, several percent 2nd HD (depending on the stylus shape, cutting process and so on - including tonearm alignment) at lower frequencies, though this is nothing one does not see from Speaker drivers at high power levels...
Correct, most recordings explore neither the limits of LP's or CD's dynamic range and that includes most recordings on SACD and DVD-Audio etc.
This is not our experience with "High End" equipment (example - Kuzma Stabi Turntable [the original suspended one], Triplanar Tonearm, Koetsu Onyx Cartridge, Phonostage of my design which incorporates a 114dB DNR 24/96KHz AD with low jitter clocks and low noise PSU's etc...), we find even 24/96KHz to be merely "quite close" to the original LP, but not identical.
It is possible however, that using lower grade analog playback equipment CDR recordings suffice.
In the "old days" CD Pre-empahsis was carried out in the Analog Domain. The improvement in sound quality was palpable. Just try the same recording on a Sony PCM F1 with or without pre/de-emphasis.
Of course, in those days we had real 16 Bit converters as well, not 9 - 12 Bit equivalent resolution "low-bit" converters that get advertised as "24 Bit" or even "32 Bit" and use the additions of >> 100% fuzzy distortion for most signal levels to pretend higher resolution in the audio band.
With De-empasis moved into the digital domain all it's benefits are lost and
given the current state of AD/DA converters this is especially regrettable, as these could benefit greatly from some pre-emphasis.
Ciao T
Lots of good information in your reply, Thorsten. Thanks.
How far does tape go? Can it compete with -115 dB at 16-bit? What about 24-bit where the levels would be even lower?
You are correct, of course, that dither reduces dynamic range. That's where the 91 dB figure comes after starting with 96 dB without dither.
Check AES papers from John Vanderkooy, et al, which show that DSD is not capable of proper dither due to there only being two possible values for the signal at any given time. His papers also show why lack of dither is a 'fail' for digital.
Maybe you just need 192 kHz.
I can agree that LP and CD are largely similar, but where do you get this 87 dB factor? My understanding is that a realistic number is 91 dB (ask Bob Katz). I would strongly disagree with the "pure advertising" criticism, because CD can preserve discrete tones as faint as -115 dBFS that are audible, and as low as -130 dB that are measurable.
How far does tape go? Can it compete with -115 dB at 16-bit? What about 24-bit where the levels would be even lower?
That comment is rather specious. Nobody, and I mean absolutely nobody, should be using digital without dither. If you're evaluating digital without dither, it's like listening to records cut without the RIAA curve - the media simply does not work properly without it.
You are correct, of course, that dither reduces dynamic range. That's where the 91 dB figure comes after starting with 96 dB without dither.
Check AES papers from John Vanderkooy, et al, which show that DSD is not capable of proper dither due to there only being two possible values for the signal at any given time. His papers also show why lack of dither is a 'fail' for digital.
No offense intended, but you're criticizing digital using your own digital designs as proof. I'm inclined to wonder whether there might have been a bit of self-fulfilling prophecy here. Or, are you saying that you designed the phono stage only, and not also the 24/96 A/D?
Maybe you just need 192 kHz.
- Status
- Not open for further replies.