Separate names with a comma.
No, No, and no.
Because the volume variability between discs is huge. And even within a single record there can be huge leaps in volume. It's more difficult to set levels when miking an acoustic machine than doing electrical transcription. I've tried to record my phonographs before, but I always seem to bump into a part that blasts and blows out into digital distortion. The dynamic range seems broader than for most LPs, but that may be because they didn't have limiters or any way to adjust levels back then. The sound that went into the horn was what was encoded in the grooves. When singers are very close to the horn, it sounds almost exactly like the range of a real human voice. When that voice is Caruso, the range is massive.
I don't know if you've ever had the chance to hear a record that was recorded and played back acoustically, It sounds amazing, but it isn't at all following the rules for modern hifi... the response curve has big wolf tones, the dynamics are not natural, the sound is all in the midrange, the horn is projecting the sound in front of the machine in a very directional manner, and the room expands the sound and adds a natural feel to it. The overall impression is so striking and present, it can make the hair stand up on the back of your neck. For some reason, it is VERY difficult to recreate this sound with electrical transcription.
I think the various sound labs at the manufacturers in the early part of the 20th century were experimenting with psycho-acoustic principles that we don't deal with any more. We are able to present sound more accurately. But they were able to optimize very inaccurate sound in a way that still maintained the spark of life and presence in it. I've been playing around with my phonographs testing them and trying to recreate their sound. It's very difficult to figure out. I wonder if anyone has done controlled tests on just what effect the needle, mica diaphragm and horn have on the sound. It would also be interesting to compare different brands and models, because each manufacturer had its own house sound and recorded their records to sound best on their own machines and vice versa.
1. If you're going to claim that 24 bit audio is necessary to record acoustic 78s played on an acoustic phonograph, we have to deal with numbers, not subjective terms like "huge" and "huge leaps". 40dB of chance, for example, may be described as "huge" by most people. Some think 20dB qualifies as "huge". We can't have this discussion with 20dB+ dissagreements in term definitions.
2. So one pass to establish maximum peak then adjust for that doesn't work for you? Start low, observe the highest peak-hold value, then adjust up if necessary?
3. The DR of a 78 might seem broader but it isn't. https://en.m.wikipedia.org/wiki/Dynamic_range
The 78 dr reference is from a cited AES paper. 40dB, with wear, 30dB. If we add a resonant peak of 20dB (is that huge?), we still max at 70dB. That's 12 bits.
4. The sound that went into the horn, got distorted by innumerable horn resonances, again by the nonlinear behavior of a ridged diaphragm, again by the shape of the cutting stylus, and again by the spring-back of the wax. Then most of that all over again during playback. It is amazing the result is at all pleasing, and recognizable.
5. I have, many times. I was impressed by the playback volume, very much so. I was also impressed by how completely unique to that system the sound is. Nothing sounds like that, not remotely, not the original event either.
I'm specifically talking about acoustic 78s played back acoustically. That is entirely different than electrically recorded 78s played back with an electric pickup. You can't generalize about acoustic phonographs and acoustic recordings. Every manufacturer used a different standard. Recording them isn't at all straightforward. Once you deal with overcoming room tone by close miking the horn, you're dealing with a tremendous volume level. It can reach peaks of well over 90dB with a loud tone needle. If I put my head where the mic sits relative to the horn, my ears distort on the peaks and they ring afterwards. It's beyond the threshold of pain. The surface noise mostly exists beyond the response of the sound box, so it is much lower than with electrical transcription. If you plan to do noise reduction in post, having that quiet bed of surface noise well defined would help a lot with pattern matching noise reduction.
Playing a record on an acoustic phonograph involves cranking up the machine and changing the needle with every record side played. A record runs 2 1/2 minutes. It just isn't practical to set levels by pre-auditioning every record. You set one level with a ballpark volume setting that doesn't distort with the loudest peak and you play them all in a row without stopping the recording. It's like a live performance. But a Magic Note Columbia record may be recorded at a drastically different volume level than a Victor Red Seal. There was no calibration for pre-electrical recordings. You have to set for the loudest and capture the rest wherever they fall.
I've done a lot of electrical transcriptions of acoustic records and it is a ton of work and processing to get them even close to the way they sound through an acoustic phonograph. It takes suppression of the noise floor, dynamic expansion, and selective emphasis of specific bands with EQ. I spent a week working on a single record once, listening to it on my suitcase Victrola and then trying to match it on my transcription. I sorta got something that sounded similar. Different machines sound different yet the same... it isn't just horn resonances. My Cortez has a 27 inch spruce horn and my suitcase Victrola has a little tiny paper mache horn. Yet they both present acoustic recordings with the same general sound. I don't fully understand everything that's going on.
When you buy a CD of a Caruso record, they don't even try to make it sound like it does on a Victrola. They take the approach that they are going to present it as close to what the music originally sounded like as possible by removing everything that wasn't a part of the original sound in the studio. That subtractive process results in wimpy, flat, compressed sounding recordings, because they're operating solely with the shadow of the sound, not the full sound of the recording itself. The acoustic phonograph is a musical instrument and the sound it adds to the music is part of the effect. The resonant qualities of the phonograph aren't distortion any more than a violin's resonances are distortion. Totally different way of thinking of sound reproduction than we have today. But anyone who has heard a really good acoustic phonograph knows it's a totally valid one.
Interesting point, and one I've previously wondered about. Many instruments rely on resonance, and resonance has been increasingly engineered out of audio reproduction. What we're left with is a by the books method where the original performance is captured as accurately as possible by a mic, stored as accurately as possible in data, and played back later as accurately as possible through a transducer which neither adds or takes away form the energy captured in the data. The result, deductively, is that the outcome of that sound reproduction chain will sound just like the original performance did. Considering the outcome of this modern approach, it's quite practical, newer gear is more technically precise, and able to reproduce sound with less distortion. The downside is that the natural resonance characteristics of certain instruments, including their ability to oscillate, which I believe required a non-liner response from the speaker to truly sound life like, cannot be recreated without resonance (and by definition distortion). Examples are trumpets, clarinet, and violin. Their sound relies naturally on resonance. Perhaps this is the reason certain recordings sound more realistic to you on the phonograph, because rather than achieve a perfect recording, or a perfect transducer, the designers simply aimed to recreate the sensation of an instrument, or to build an instrument out of a transducer. And that requires introducing a little resonance. You'd never see THD specs on an acoustic instrument. It's a different realm of audio, and different approach. Perhaps they saw the phonograph more like that. Our attitude towards speakers nowadays is totally different. Speakers are not instruments, they are scientific reproducers of sound waves. They either achieve an ideal, or fall in various ways from that ideal. Interpretation or enmphasis is not a part of the engineering process anymore.
I play violin and have always been fascinated by the way sound vibrates through a the instrument so effortlessly, atleast a good example of one, and how much of a command you need to have over the resonance - being able to provoke it instantly, or make it bleed away softly - in order to make it sound sweet or forceful. Being too gentle with the bow, and shying from that resonance, will lead to a loss in tone, or the dreadful squeak. I've found that headphones with tame or polite treble don't do violins justice. Same with trumpet and clarinets. And the most lifelike reproduction of a trumpet I ever heard was through a loaded horn tweeter with lots of (probably resonating with) treble energy. I have never been a huge fan of excited treble delivery, but I came to appreciate it, including the loaded horn design, for its ability to recreate certain instruments and singers in a lifelike way. I wouldn't want that kind of a reproduction for all genres, so it depends on the content. I've always preferred a neutral system first and foremost, but I do think the struggle for neutral response has robbed some of the life from music. Speaker systems nowadays are capable of playing most situations with more accuracy, but the lifelike sensation of having certain instruments in the room is being sacrificed. I still regret not setting up a tiny 2 channel music system with those little Hsu horned bookshelves.
You guys should try the Final Audio 1601 or 1602 IEMs. These are designed to emulate the highly mid-centric and resonant horn speaker sound. Very strange things... but with the right music, admittedly pretty fascinating.
How about this set up? Maybe it would sound good.
1. How does recording with modern mics, digital audio and modern speakers stop an instrument from oscillating/resonating?
2. A system (or speaker) with a resonance which might be similar to one instrument or sound would necessarily not be similar to the resonance of other instruments/sounds. The resonance of each instrument (including the human voice) is different, are you suggesting a system/speaker with a resonance the same as a violin, another system with the resonance of a trumpet and so on ad infinitum?
3. If an instrument resonates (and I can't think of any acoustic instruments which don't) and if we can accurately record and reproduce that instrument, which must obviously include it's resonance, why would you want to add another resonance in addition? How would say a violin with double it's resonance "truly sound life like"?
Then surely the solution is more accurate headphones, not less accurate/linear ones with a particular resonance similar to a violin?
I've noticed that the horn channels the sound like a beam of light. It's very directional. Totally different than the dispersion pattern of speakers. If you sit a little bit off axis to the direction the horn is facing, there's an eerie effect that projects an aural image of the singer four or five feet into the room in front of the phonograph. The sound doesn't seem to be coming from the horn. It's as if the ghost of the singer is standing the same distance from the horn as when he was standing in front of it in the recording studio. This gives a particular sort of sound that modern recordings don't.
Another thing that acoustic recording does differently than modern techniques is using the room for its natural characteristics. With speakers, you try to avoid reflections. With an acoustic phonograph, they tell you to put it in the corner of the room to get more reflections. The reason is that acoustic records have no reverb or room tone at all. The recording horn didn't pick up any sound further than ten feet away, so the sound is totally dry. If you play that in a live room, the ambience of the room wraps around the recorded sound and makes it as if the singer is in the room with you.
You could probably create a horn loaded speaker to simulate all of this, or just play an iPod into the tone arm of an acoustic phonograph. It's not that modern techniques are unable to do this, they just don't do it.
1. It doesn't. I was observing the modern reproduction chain, and how it lacks a sound behavior native to instruments.
2. No, obviously too costly. I've invested in a neutral system "first and foremost", but a loaded horn bookshelf for certain music would be nice to have for the right money.
3. You'll always capture the frequency of resonance with an accurate system, but dynamics, imaging, and attack of trumpet can be aided with the resonance of a loaded horn. I tend to lean toward treble emphasized headphones when listeing to violin, and think one that emphasize mids sounds great with cello. Variation from perfect accuracy can help some instruments sound more life like. The problem is picking a system for each instrument, so for value reasons neutrality is an ideal baseline.
4. Ideally, but "more accurate headphone" is kind of a loaded term. For instance, I think DT880s are neutral, but some people disagree, and think it has treble emphasis. Lots of people say the HD600 is neutral. I think it distorts the mid range, albeit pleasingly. Both headphones are pretty close to what most people would regard as 'accurate' though, both of them recommended often as monitoring headphones, but even in their case there's some major range in the interpretation. I own both, and find they excel at different things.
Accuracy and neutrality are important, but I think the point being made here is coming from a different perspective, and you're applying a traditional one about sound reproduction in order to try to understand it. To illustrate this, I know they sell pianos with robotic mechanisms that can play pre-programmed songs. So imagine that we managed to record not only the sound of Glen Gould's Variations in sound waves, but the exact inputs he made to the piano, down to thousandths of a second or mm resolution. We then took an example of Glen Gould's grand piano and hooked up a robotic mechanism that could recreate the timing and movement of the performance we recorded. We also played back his performance through the most accurate set of speakers we could find for the same cost of a his grand piano. Which would be the more "accurate" rendition? It's two completely different approaches, but in my opinion, there's no way the speakers could ever reproduce what the piano could. Yes, there would be deviation form the actual performance, and we would be missing the environmental acoustics, or Glenn Gould's humming, but we would gain a true representation of the resonance characteristics of that instrument. With millions of dollars in robotics and quality instrument purchases, you could have an amazing virtual orchestra that no speaker could ever beat. Perhaps an extreme way to illustrate the idea, but along the same vein nonetheless.
I still don't understand your logic. If you accurately capture the frequency of resonance of an instrument, how does adding more of that resonance "aid" the sound or make it more "life like". If it was captured inaccurately, with too little resonance and it wasn't compensated for during mixing, then adding more of that resonance in playback would make sense, providing of course it was the same resonance as the missing resonance and of course, it would be pure chance that the resonance of a horn speaker matches the missing resonance. The last thing I would want is a sound system which matches the sound behaviour of any one instrument, what happens if I want to listen to a different instrument or a group of different instruments? Unless I only ever listen to one unaccompanied instrument, which is always recorded inaccurately with the exact same resonance reduction, I want as linear a system as possible.
Don't forget the double basses, you'll need headphones with an emphasised bass for those. So what you ideally want is headphones with an emphasised treble (for violins), emphasised mids (for cello) and emphasised bass (for the basses). Explain to me how these ideal headphones would differ from linear/accurate headphones?
One of the problems, is how you define life-like or natural. A few inches away, say in the case of actually playing the violin, you will hear far more treble, details and dynamics than you will from say 20+ feet away in a concert hall. That's almost certainly why you prefer added treble, dynamics, etc., because it more closely matches the sound you are used to. BUT, what is more life-like or natural, what the musician hears a few inches away or what the audience hears many feet away? Music is typically mixed (though not typically recorded) from an idealised audience perspective, NOT from just a few inches away. If we recorded each instrument in the orchestra from just a few inches away and mixed it as if you were an audience member sitting just a few inches away from each of those instruments, it wouldn't sound much like any orchestra you've ever heard!
I don't get your analogy or what it has to do with the point at hand. If I were to record Glenn Gould and accurately capture all of the resonance etc., how does adding more resonance than existed at the time (by reproducing with a horn speaker) make it better or more life-like? It will make it less accurate and less life-life, the only way it could be described as "better" is if your personal preference is for inaccuracy in the form of too much resonance in a particular frequency range.
I know what I like, and I like what I know.
Me, I'm just a lawnmower - you can tell me by the way I walk.
When the sun beats down I lie on my bench
I can always hear them talk.