zondag 17 juni 2012

Temporary stop blog 33 toeren klassiek


Gisteren kreeg ik deze reactie:

I do appreciate your uploads; but one thing MUST be acknowledged. Many, if not most, have a background buzz signature that consists of a series of discrete peaks at harmonically related frequencies, running from the lower midrange all the way to the cutoff of your digitizer lowpass filter. The buzz will register at a varying amplitude, sometimes as high as, cumulatively, about -35 dB on my digital meter. I have devised a very complex filter to remove it, but of course this cuts little narrow "holes" into the spectrum. It seems to me that possibly something in your computer is leaking RF interference into your audio system (do you use a tube preamp?) and that you are not monitoring carefully on an extremely wideband system or perhaps have a personal slight hearing loss at the high end. I find that it takes me about an hour to two hours to clean up a typical LP for saving, quite a burden but helpful if, in many instances, the item is completely out of print and has never been issued on CD. So I hate to "look a gift horse in the mouth" but would respectfully suggest that you experiment and see if wiring routing, equipment placement, common power socket connections, superfluous grounds, or other issues can be checked to see if this can be MINIMALISED in future. For example: if you run a digitizing program, leave the stylus OFF the record but examine its meter WHILE 'recording' a blank file. Play it back and greatly increase the gain: you can immediately see the buzz, which is louder than turntable rumble and typical record surface noise and intrinsic tape hiss. Yours respectfully, a retired recording engineer

Ik dank de anonieme geluidstechnicus voor zijn reactie.
Kennelijk is er dus een serieus probleem dat ik niet zo gemakkelijk kan oplossen. Dus dat betekent dat ik de blog stop tot het probleem van de buzz opgelost is. Aangezien ik bepaald geen geluidstechnicus ben neem ik aan dat het wel enige tijd kan duren voordat ik de buzz eruit heb.
Dus graag geduld, en tot later.

Satyr

15 opmerkingen:

  1. I haven't noticed any hum or buzz on your recordings, but I do know that my computer will introduce hum into my transfers if I use AC power to the PC while transferring - even if the PC has a common ground with the turntable. The solution there is to run the PC on battery power. I also know that flourescent lights cause a buzz on my system, so I avoid them. You might also try reseating or replacing your interconnects, and making sure your cabling does not cross paths with AC cords.

    BeantwoordenVerwijderen
  2. Thanks a lot for your reaction! My PC cannot run on battery power, but I'll try the other solutions you recommend!

    BeantwoordenVerwijderen
  3. Satyr, you could try several solutions; but since you 78 issues runs there must be easy to solve that problem.
    a)check your wiring system, use 1st class wire
    b)avoid your signal wire pass around AC power sources
    c)check your ground connections
    d)check if you hear this distortion on the loudspeakers or headphones.
    e)if a variable gain output is in use take care of use the less sensibility in the input line, in other words, set the line in volume less than 25% and increase the output line.
    f)set of the PC playback during the recording operation, it will help you avoiding realimentation.
    g)always use line/recording output from your amplifier for made a recording, never use a headphone output.

    If you want you can mail me telling about the setup connection you usually use and how you connect it.

    BeantwoordenVerwijderen
  4. Dear Satyr, you have my compassion, because I know how easy it can happen, that you have a buzz in your transfers and how hard and sometimes apparently impossible it is to find the reason for that. As I am no maniac with high sound quality and mostly don't listen very loud to the music I hear, I would say: Better a rare recording with a little buzz than no recording at all! I know you will keep on trying to avoid this, but don't panic if you are not really successful with it. Just keep on: your recordings are valueable and fine (to my ears, which maybe are not really the best anymore...)!!

    BeantwoordenVerwijderen
  5. Beste Satyr, I have never been a "sound fanatic". For me it is the music and the interpretation that count. Also a sign of your generosity and preparedness to share your knowledge and collection. Thank You.

    BeantwoordenVerwijderen
  6. I appreciate all points of view on this situation, from "I don't care at all; the recording is rare" to mine, at the opposite end of the spectrum, which might be summed as "don't post unless you have added nothing spurious to the transfer."

    This is why I have taken the route, for the past months, to FIX IT MYSELF employing the complex filter that I have divised.

    I cannot quite fathom the graph that Satyr has posted, as it is not quite comparable to ones that I have made. What I perceive is a series of (apparently harmonically related) narrow "buzz" peaks that start around the lower part of the midrange, extending all the way up to the digital filter cutoff, just below the top extreme response of 44/16 audio. The worst (highest amplitude) of the narrow peaks will tend to be dispersed from about 2k to 7k, with another peak at about 11.9k. There may be as many as 16 or 17 of these very discrete "buzz spikes" which are only a few Hz in width. The loudest one -- of course it varies from transfer to transfer -- will be in the 2-4k range, about 35 dB below maximum digital modulation. However, this is where the ear is most sensitive so the buzz is ALWAYS constantly audible (to me; I use good speakers.) It is easily discernible in the fade up at the start and end of files, and stays in the unmodulated lead grooves.

    I have noticed this buzz in virtually all of your uploads; it does slightly vary in "tone color" and intensity from one to another.

    Now: here is a strange thing. There is one other blog that has this exact same problem; in fact, I can use the identical 'correction filter' I've devised for files from BOTH of these blogs. It's the blog that has the Mercury Living Presence records, here:
    http://rareclassicalvinyl.blogspot.com/

    That is where I first noticed the buzz, and undertook to analyze it with a Fast Fourier Transform and devise a complex low-level dynamic filter to remove it. When I later heard this identical buzzing in the "Satyr" files, I wondered if the noise signature was similar; overlapped files of unmodulated grooves; and discoveret IT WAS VIRTUALLY IDENTICAL: the same filter worked with files from BOTH blogs.

    I won't say what occurred to me at this point. You might guess!

    At any rate, the other blog brags (or at least did at one point) about using a tubed preamp for the digitizing. I have a great deal of experience in the design of older audio technology and, indeed, once designed a series of tube amps and processing devices back in the sixties and early seventies. Some wideband tube designs have much more bandwidth than certain modern solid state devices. The danger is that RF signals can leak in, intermodulate, and turn into AUDIO that can be detected by ear. RF hash from computer display oscillators, the main PC switching power supply, and any number of other sources might leak into the phono cartridge pickup wires or other vulnerable parts of a high end audio system; it has happened to me during the many years I've been responsible for the maintenance of high quality recording studio gear.

    I will say that I have the point of view (which others won't necessarily appreciate) that positive criticism that is fair, honest, and diplomatic, encouraging the IMPROVEMENT of a product, can't be bad!

    With respect to your graph, it seems to me that you might try to HEAR the buzz so that you don't rely on the slow process of a computer trying to analyze and graph it. Use enormous gain and listen to the output of your files with the stylus OFF the record. As you make this 'blank' file, move your arm and hand AROUND the area of the pickup head and tone arm, and AROUND the wiring leading to the preamp phono cartridge input.

    I will have a bit more to say in my next posted comment.

    S.-retired audio engineer

    BeantwoordenVerwijderen
  7. You might also try to make two files, one WITH a ground isolator, and one WITHOUT one. An isolator that eliminates the contact of the center pin (US Standard--I don't know how the Euro AC adaptors work though this surely is readily available in a net search) can eliminate ground loops. Laws mandate the use of a three-wire connection now; but that third chassis ground wire OFTEN causes a ground loop that can transmit noise and hum.

    This is indeed why battery operated preamps are sometimes preferred!

    I have to put ground pin isolators on virtually EVERY single component in my stereo systems, and in the main computer and the power amp connected to it. In addition, I "reverse" the non-polarized two-pin AC plugs until I get the lowest buzz.

    But, this buzzing signal that I am describing does not resemble "ground loop induced hum" to me. It is much more like a demodulated RF carrier that has caused a nonlinear distortion in a very wideband audio circuit. This distortion product turns what would be perhaps one big electrical signal peak into a harmonic series.

    I must say that persons who have not had hands-on design experience, using spectrum analyzers, in checking "out of band" performance of audio gear have NO IDEA what happens when you operate such devices in the proximity of a strong RF signal. Typically, nonlinear 'detection' somewhere in such a device will create either a constant wide-band "hash" noise (a soft sshhh sound) or, as I perceive here, a discrete series of isolated 'tones'. This will rise and fall sometimes when one even moves the interconnects around in such a field!

    S-retired audio engineer

    BeantwoordenVerwijderen
  8. Earlier I explained that ONE of the digital "buzz peaks" in the harmonic series might measure at about -35 dB below maximum modulation; but that can be misinterpreted, and seems -- on the face of it -- to be inconsequential. I've done one further check, obtaining the average of all peaks I've measure in your upload of the Bizet recording by Fournet (1953 Philips Lp) during my run of the filtering process to clean out the buzzing; if I add together the series from 2 to 12k, I get an approximate average signal amplitude of about -18, which means that the buzz registers louder than the pp to ppp passages of the music; it's about equal to mp and covered up by mf passages; but it is then instantly audible again the moment the music dynamics fall below loudest. As I said, the human ear is very sensitive to the region where I measure the loudest peaks of this harmonic series. My hearing is not at all abnormal; at my age I have a distinct loss about 15 k and can BARELY detect a steady 14 k tone, requiring a bit higher amplitude of that, compared to 1 k, to hear it. However, in the middle frequency range my hearing is not affected by age and, since I have NOT been a rock'n'roll recording engineer but one who has focused primarily on classical music, I have not wrecked my hearing with constant high amplitude din!

    I discussed this buzz with a friend who is an audiophile, record collector and amateur transferer, and a professional musician & trombonist (he has sat in brass sections all his life!) -- and HE can hear the buzz, constantly, on your transfers. In fact, he appreciated ones that I'd cleaned up since yours were obtrusively noisy in his opinion.

    So, two "old guys"--with different gear, on opposite sides of the US continent--and we both agreed exactly on your buzz, and heard it with the same discomfort.

    S-retired audio engineer

    BeantwoordenVerwijderen
  9. Oh--I might also add that I have a private alternative transfer, done by my east-coast USA friend, of the Eva Bernáthova, Janácek SQ: Dvorak (DGG, 1957) that you have posted (I have not DLed yours) and it has NO TRACE whatsoever of any buzzing noise.

    Nor does another blog's transfer of the Jochum Eroica; again, I've not checked yours.

    S-retired engineer

    BeantwoordenVerwijderen
  10. To confirm: yes, your transfer of the Dvorak Op. 81 (DG, 1957 Lp) has a LOT of buzz. I checked the second movement: it is an easily perceived "alien thing" above the music, and to my ears sounds like two or maybe even three discrete tones; on my spectrum analyzer I see exactly the same multiple harmonic series from lower mid up to inaudibility. It REALLY stands out in the slow movement as being *particularly* irritating. My friends old alternative private transfer, unfiltered, has NO TRACE of this same buzz so, I'm sad to say, I'll not be keeping your transfer.

    To those who say: "It doesn't matter..."

    What is your cutoff? HOW BAD does a hum/buzz or other intrusion have to be before YOU say, "um, that bothers me"?

    We are all very different.

    I can say this, as a person who has professionally produced classical LP releases and did major broadcast syndication recording for one of America's top six orchestras: this would NOT be acceptable for professional broadcast or commercial release.

    Furthermore, the noise level is at least 35 to 40 db HIGHER than would be permitted by American FM radio station broadcast standards.

    It's NOT inconsequential. It's NOT inaudible.

    S-retired audio engineer

    BeantwoordenVerwijderen
  11. Thanks for all your reactions and thorough research on my uploads, dear audio engineer! You don't need to convince ME that the buzz is there and that it is important to do something about it - I agree with you, and again thank you that you let me know the buzz is there! That's why I temporary stop this blog, to take time to do something about it.
    Greetings,
    Satyr

    BeantwoordenVerwijderen
  12. Hello Satyr.
    Sorry to read about your problems - alerted to you by Mr.Waldee (I wasn't aware of them as only downloaded 78 dubs!).

    If you had given some hints about the equipment then maybe some 'educated' suggestions could be made - but your 1kHz spaced 'spikes' are a normal feature of 'EMI': there must be some reason why the hash seems mainly located @ 12kHz (easily seen via Audacity 'analysis' on the silent-end grooves of the Dvorak) and appears to spread outwards from that frequency.

    I checked some early transfers made through a Leak valve (tube) pre-amp and can't see anything similar: the only difference would be overall phono S/N ratio (~-60dB as opposed to ~-85dB for SS).

    One solution may be to get a CD-Recorder; which I continue to use 'for convenence'...having bought 3 from eBay (£100 all-in)..
    Pioneer PDR-509 (20bit circuit: there's a later 24bit PDR-609) + TEAC RW-800 (24bit - very fast CD-RW erase, etc - virtually identical to their 'pro' TASCAM CDRW-700).

    CDRW discs are cheap (though not everlasting) - but should (unless this noise is a 'feature' of your HiFi..) avoid computer EMI hash - at least from my experience with various equipment.

    BeantwoordenVerwijderen
    Reacties
    1. Thanks for your suggestions, Tin Ear! I hope the problem will be solved soon.

      Verwijderen
  13. I do miss your posts.
    Wish you a very nice summer.......

    BeantwoordenVerwijderen
  14. Hello Satyr,

    Likewise I miss your uploads and the interesting information you post with them.

    I look forward to you recommencing your blogs again soon.

    Thanking you and have a good summer.

    Cheers,

    Douglas (UK)

    BeantwoordenVerwijderen