|'); } // End -->|
In Reply to: Vinyl compression posted by Christine Tham on January 2, 2007 at 15:14:18:
The most straightforward example of relationship between dynamics and how the grooves look is any Tchaikovsky 1812 Overture, but any dynamically wide-ranging LP will do. (Pop, Rock, or elevator music will not do, but classical chamber music should also work for my example).
As I'm writing this I have before me Telarc's 1812 Overture on vinyl. By looking at the grooves (with the naked eye) you can see where the canons hit. With even a magnifying glass of even weak strength can see where the canons hit to the nano second, and you will say to yourself, wow! It is no wonder that tone arms have difficulty staying in the grooves at that point. The grooves have shear rises and drops. The grooves are very wide, almost like the unused space between the end of the music and the label.
I also have before me Ormandy's 1812 on RCA Red Seal. This version is very much compressed compared to the Telarc. So, the grooves, compared to the Telarc, are close together. Nevertheless, the loud passages are, too, very visible compared to the quieter passages on the same disc.
This is why you can fit so much more chamber music on an LP than you can music with wide dynamics.
(Also, with an LP with both quiet and loud passages if you hold it up to the light the more uniform quiet passages will look dark black, while the more uneven louder passages will look "gray/black")
Interestingly, Jared Sacks of Channel Classics (see threads below) explanation on why the entire Mahler 2nd could not fit on a single disc, cited the wide-ranging dynamics of the piece which would not allow for it all to fit. That was brand new to me. I did not realize that digital had those types of "analog" limitations.
Robert C. Lang
Thanks for your reply. I think I know what you are getting at.
*** As I'm writing this I have before me Telarc's 1812 Overture on vinyl. By looking at the grooves (with the naked eye) you can see where the canons hit. With even a magnifying glass of even weak strength can see where the canons hit to the nano second, and you will say to yourself, wow! It is no wonder that tone arms have difficulty staying in the grooves at that point. The grooves have shear rises and drops. The grooves are very wide, almost like the unused space between the end of the music and the label. ***
The widening of the groove spacing is a common technique to allow loud passages to track better. But it does not mean that a well designed and properly setup turntable is not able to track loud passages in narrow grooves.
As I've said before, I have transferred quite a few LPs to digital (probably a quarter of my collection). I can tell you there is very little correlation between the actual dynamics present on the LP and how wide the groove spacing (apparently) is. I have seen LPs with very narrow groove spacing that has dynamics far in excess of LPs with wide groove spacing.
I think you are probably underestimating the ability for a well designed stylus to track very sudden and loud transients. If you think about it, that's what a scratch (which creates a loud "pop") looks like to a needle. Most good LP decks should have no problems even tracking very deep scratches (of course, whether the amp and/or speaker can reproduce the resultant spike is a different matter - I tend to agree with you that most systems struggle with very large and sudden transients - I have measured "pops" from scratches in excess of +10dBFS - which is why I always allow 12dB of headroom when transferring from LP).
I generally find that even deep scratches produce very clean transients, in fact far cleaner than I would have thought possible given issues with arm and/or platter resonance, stylus tracking etc. On most of my transfers, scratches/pops come out as an almost perfect square impulse lasting 1-5 samples, which means a "pop" generally lasts for less than a millisecond, with no signs of ringing or other artefacts associated with mistracking. They are so clean I can easily digitally edit them out (and I do, for the really intrusive ones).
So, if a stylus can track pops so well, one can infer it should be able to handle "normal" music (yes, even the cannons in 1812) with no problems, even if the groove spacing is narrow.
As for SACD, the amount of compression you can achieve depends on the material. I wouldn't be surprised if a very dynamic recording of an orchestra compresses less than a quiet recording. Hence issues in terms of the maximum amount of playing time per disc - it varies (slightly). In the extreme case, pure noise does not compress at all, hence a recording of pure noise would take up twice as much space as a "normal" piece of music (DST, the lossless compression scheme for SACD, is designed for a target compression ratio of about 2:1).
First to keep things in *some* perspective, this is germane to the discussion on the Fischer Mahler 2nd and numerous other SACDs (and DVD-A I'm sure) is that now there are *routinely* many wide-ranging recordings out there that are not found and will probably not ever be found on vinyl (not without dialing back the dynamics so that all (not a select few) turntables can track them *perfectly*. There are now more discs that have "1812" type dynamics than ever before as the quest to approach the dynamics of a live scenario (for better or worse) continues. The beauty of SACD is that any player can track the most dynamic recordings out there (or at least it doesn't take an expensive set). This is simply not true for vinyl as far as I know. Remember tracking and tracking perfectly are very two different things.
****I have seen LPs with very narrow groove spacing that has dynamics far in excess of LPs with wide groove spacing.****
Absolutely! That's why it is important you compare a specific disc(s)that has *both* very quiet passages and very loud passages in the same performance. In that case (unless the engineer was riding gain, that is, compressing the loud passages) the loud passages will always consume more vinyl space (wider passages) than the quiet passages. (Or unless they are trying to squeeze more on the LP side than can be comfortably accomodated). The loud passages need more real estate everything else being equal.
This is *definitely* the case with the two specific 1812 Overture examples I gave to you.
So, agreed, you may (will) see LPs with very narrow groove spacing that has dynamics exceed LPs with wide groove spacing but probably not from the same label (or recording team). There are many examples of 6 cylinder engines that are more powerful than 8 cylinder engines, but generally not within same product line.
So before you compare labels to each other (which is what you have seem to have done) first compare Telarc to Telarc or RCA to RCA, preferably with the same recording team. You will find in both cases that the loud passages will have wider grooves that the quiet passages. True, you may find that the loud passages on the RCA are actually less wide than the more quiet passages of the Telarc. You will also find that the Telarc with the wide spacing for the loudest passages will (generally) *sound* less compressed and much more dynamic.
"So, if a stylus can track pops so well, one can infer it should be able to handle "normal" music (yes, even the cannons in 1812) with no problems, even if the groove spacing is narrow."
I don't see a "complete" correlation here, (especially since typically the narrow grooves, that contain the quiet passages are easier to track) so I won't address that point. But all 1812 recordings hardly give tone arms the same challenge. Some are no challenge at all! And they sound like it. What wide ranging 1812 Overture that you own that utilizes narrow grooves (especially in relation to the quiet passages on *that* disc)? Even my compressed Ormandy version has grooves that are *visibly* wider than the quiet passages on that disc.
Also, there were then and as now set ups that could track the Telarc 1812, but only with *great* difficulty and at great cost or compromise to sound quality. The very best sounding (that are often the heavist tracking) couldn't track it very well. Even today high-end dealers do not want to risk embarrassment by that disc. How well does your set up track it (Telarc)?
While I have thousands of vinyl I have not purchased any in 7 years. I'm sure there may be *some* but in the vinyl world are discs being routinely pressed that approach what we are seeing today with SACDs?
Do you have the Fischer Mahler 2nd?
Before we go much further, I must point out I'm not trying to pick a fight here, and as you well know I like SA-CD a lot and I have stated in the past vinyl is my least favourite format for many reasons.
*** The beauty of SACD is that any player can track the most dynamic recordings out there ***
Conversely, I could argue that all digital formats have a hard ceiling in terms of maximum amplitude that can be captured, whereas analog on the other hand has no artificial upper limit.
Think of it this way, you can never record a digital signal higher than 0dBFS, by definition. If you do, the signal clips.
Whereas there's no artificial limit in terms of how "hot" you can cut a signal on LP. Theoretically, if you master at a slow enough cutting speed, and your stylus is up to it, you can record a groove containing a signal hotter than a notional "0dBFS". This is not just theory, it actually happens in real life. I did mention that I record my LPs with about 12dB headroom. That headroom is necessary, because I regularly encounter signals hotter than the notional 2V maximum (for line levels coming out of my phono stage). It depends on the recording, and the pressing. But it's not uncommon.
*** the loud passages will always consume more vinyl space (wider passages) than the quiet passages. ***
As I've mentioned before, this is not necessarily the case. The loud passage can consume exactly the same space as the quiet passages.
It seems from your post you think the groove spacing must be increased for loud passages relative to the quiet ones, but this is not so. It is good practice to widen the groove spacing for difficult (i.e. loud and dynamic) passages, but it doesn't have to be so. In fact, you don't need to vary the groove spacing at all as long as it's wide enough to accommodate the hottest peak. As I mentioned before, based on my experience, there's little or no correlation between groove spacing and actual signal amplitude. In fact, the LP I own with the widest groove spacing happens to be a very quiet LP, it's because each side only contains about 8 minutes or so, so the cutting engineer decided to cut the record with very widely spaced grooves to fill out the side.
Mastering lathe cutting pitch does NOT effect the level of the recorded signal. It's the other way around. The level of the recorded signal controls the cutting pitch. The cutting system pre-delay simply tells the lathe what pitch must be set one revolution ahead for the upcoming signal. That's how more than 15 minutes of playing time can be achieved on a 12" LP. Prior to Vari-Pitch, all cutting was done at a constant pitch setting with no accomodation for levels.
I hate to burst a bubble here, but the vast majority of modern (mid-80s to present) disk cutting systems use a pcm digital delay to feed the cutter head, regardless of the recording source. In the case of analog tape, the direct tape machine output feeds the pre-delay control signal to control the cutting pitch. A pcm digital delay of the same signal feeds the cutter head. Exceptions would be where the system is set up with an analog pre-delay tape head prior to the main tape machine PB head. That's the way it was done prior to the advent of high-quality pcm digital delays. I'd like to think that ace mastering engineers such as Steve Hoffman would be using the all-analog method rather than a pcm delay in the chain. I've been out of the loop in LP mastering awhile so I don't really know.
At MikeL's house on his Rockport Sirius III with single sided half speed mastered 45... discs I thought (just a few times) I could hear a pre-echo, telegraphing or a shadow of what was coming a revolution later, since I don't believe it was a failing of the TT or the disc, I wounder if it was a change in the recording pitch when a change in dynamics was approaching. Does this sound possible or is it something more mundane?
What you're hearing is likely "pre-echo." That can occur as a mastering defect if the cutting pitch wasn't expanded enough. Normally, at the beginning of each track the pitch is supposed to be overly coarse (lots of land between the grooves) to minimize pre-echo. Pre-echo can also occur as a pressing defect, particularily if the vinyl press is set up with too much pressing pressure on the "biscuit" of vinyl.
You could also be experiencing "print-through" from the original analog tape if that was the source for the recording in question. You'll hear print-through a lot on older analog tapes with no noise reduction or that were recorded too "hot" to tape. Frequently, just being stored for a long time - and perhaps with too tight of a tape wind - the print-through occurs.
DSD has 6dB of soft clip at the top not hard like PCM, so in fact you can go to +6dB in DSD. Just like analog you want to avoid if possible since there is some loss of signal quality when you use it. I'm told it sounds like tape saturation.
In my experience, it doesn't sound as benign or 'euphonic' as tape saturation - I refer to it as 'scar tissue' as it's obvious (and difficult to remove) once it's there. Short peaks are fine, but any longer and "ouch!"
As usual, thanks for all of your helpful posts, particularly the recent SACD MCh discussion - I wish more people could/would hear it done well!
However, in practice, +3dB is not very useful, since on analog tape you can easily achieve +10-15dB without too much ill effects (provided the peaks are fairly irregular, otherwise the tape will over-saturate).
I've been doing a lot of field recording lately, and I certainly miss the headroom you get on tape! No matter how carefully I set the recording levels, I always get some clips. And I don't want to set the levels too low, otherwise the recording as a whole sounds rather murky.
With a supposed 70 Dbs, heck even a claim of 90 Dbs of possible range, why do you insist in over saturating the tape, thus causing other possible ill effects to the sound?
Deliberately over-saturating drums for example creates a very nice thud. You've probably heard that thud in countless 80s pop albums!
These days it's possible to digitally simulate the effect of tape saturation, but that thud sound is no longer "fashionable". The so called "New York" style popular in the late 90s and early 00s instead is created by hypercompressing the drums and mixing the result together with the original uncompressed version. Creates a different kind of "thud" - more edge to it.
Not that I hold the Wikipedia to be the best reference, it at least is widely available: "...Unlike CD, which sets the 0dB level right at the theoretical PCM signal limit, and doesn't take into account oversampling, SACD sets the 0 dB level at 6 dB below the theoretical full-scale DSD signal, and prohibits peaks above +3 dB..." http://en.wikipedia.org/wiki/Super_Audio_CD
Also here's a post from Graemme in on a conversation you (and Christine) were a part of: http://www.audioasylum.com/forums/hirez/messages/182566.html
Both say something I should have been more explicit about, tho there is some head room, it's really just 3dB not 6dB as I implied because using the last 3dB is illegal according to the spec.
Graemme weighed in other times: http://db.audioasylum.com/cgi/search.mpl?searchtext=db&author=Graemme&forum=hirez
Wait, lets not bring digital compression into the picture, OK that has not much to do with the the dynamics of the signal it has a lot more to do with the amplitude of the high frequency components in the frequency domain of the signal. A very dynamic low frequency signal will digitally compress a lot better than a loud but steady high frequency ring, and the loss less compression will affect the amplitude and the dynamic range of the signal exactly 0 db. Even lossy compression will not affect the dynamic range of the signal a great deal it will just limit how fast the signal can reach peak amplitude. Dynamic compression is a purely analog artifact and it affects only the amplitude differences of the signal, it only changes the frequency domain picture as much it reduces the rise time of the signal from volume A to volume B.
I hope this did not confuse anybody :-)
*** A very dynamic low frequency signal will digitally compress a lot better than a loud but steady high frequency ring, and the loss less compression will affect the amplitude and the dynamic range of the signal exactly 0 db. ***
This is not true, and you can check this out yourself by observing the compression ratios of various codecs depending on the source signal.
Steady state sine waves compress very well, regardless of amplitude or frequency. More so than a "very dynamic low frequency signal." If you don't believe me you can test it out yourself. Synthesize two signals: one a constant sine wave at 15kHz 0dB amplitude, and the other a low 50 Hz signal that varies in amplitude from -96dB to 0dB (you can modulate this using a sine or triangle wave).
The first signal will losslessly compress better than the second signal. At least on a codec that compresses based on a Fourier transform. Even WinZip will compress the first signal better than the second, and WinZip is notoriously poor at compressing audio.
I didn't bring digital compression into the picture, Robert did (in his reference to the Mahler recording). But you are right that lossless bitstream compression is not the same thing as dynamic compression. But I don't think that Robert was necessarily confusing the two either, he was simply citing two unrelated things that affect the maximum amount of recording time (groove spacing on LP, vs lossless compression ratio on SACD).
I was not thinking about pure tones, when i was referring to a ring i was thinking more like ringing of a bell or a triangle real loud, vs playing the lowest registers of an organ (better than a piano or standing base, because of the lower HF content. the bell or triangle will have an almost continuous Fourier transform in the HF region, the low freq pipe organ will have a pretty simple spectrum. Even though the pipe organ will have a larger volume, its digitized waveform will compress much better. If you start playing with synthesized tones, things will become very different.
True terror is to wake up one morning and discover that your high school class is running the country.
This post is made possible by the generous support of people like you and our sponsors: