@billshurv:
Fair point, I've never actually measured vinyl (it's quite a while since I had a vinyl setup), but I would have expected lower at low amplitude, I wonder how much of that is harmonic distortion and how much random surface noise (might make a difference in perception, don't know).
...Actually it must make a difference in perception, otherwise it would sound like sh*t.
Fair point, I've never actually measured vinyl (it's quite a while since I had a vinyl setup), but I would have expected lower at low amplitude, I wonder how much of that is harmonic distortion and how much random surface noise (might make a difference in perception, don't know).
...Actually it must make a difference in perception, otherwise it would sound like sh*t.
Last edited:
In my experience the use of special compressor/limiters to control the out-of-phase amplitudes at LF was rare. The bass was almost always panned close to centre and this pretty much gives you mono LF. Another advantage is that you are making the best use of two channels to reproduce large signals.
Anti-phase groove-noise signals are of course unaffected by panning.
Rather belatedly I have been googling to see if any data on LF groove noise levels could be found on t'Net.
Neither "Vinyl groove noise" nor "LP groove noise" have so far brought up anything quantitive or in any way useful.
Neither "Vinyl groove noise" nor "LP groove noise" have so far brought up anything quantitive or in any way useful.
The module had a functional name which I'm racking my memory without success to recall....
There are notable exceptions, and yes Beatles stereo stuff springs to mind.
Well I reckon it does. You can't get away from the fact that the groove velocity is much lower at the end of side.
A proper reference would help us do that.
I am not at all convinced. Can you summarise the evidence?
FWIW on CD mastering, I find that -18dB RMS is a common level for nicely dynamic CDs. Many of my classical are -22dB or a little lower. That the RMS level down from full scale. Metal and compressed Pop can be at -10dB or even higher. Kinda silly.
AFAIK the Beatles stuff is like that because it was recorded on 3 track aimed for mono release. When stereo came along mono was outdated and the tapes were remastered as best as was possible.
I've got a couple of other albums which started off in mono and were later remixed for stereo with the same 'problem'.
Last edited:
Well some might say (inc Mr Martin AFAIK(?) ) that much original production was done with mono reproduction in mind? Mono releases made first, for example, and far more time spent on them relative to stereo releases. Mono brain, nothing wrong with that at all in my book !
@SY:
You're right of course - I can't quote datasheet figures for THD alone, and yes I do know the difference. Unfortunately the N bit is heavily implementation dependent. I'll have to get the spectrum analyser and test tracks out again and try to come back with some supporting evidence.
I reacted to being labeled third person as a typical example of 80s audiophile thinking. Not a name I like to be associated with.
You're right of course - I can't quote datasheet figures for THD alone, and yes I do know the difference. Unfortunately the N bit is heavily implementation dependent. I'll have to get the spectrum analyser and test tracks out again and try to come back with some supporting evidence.
I reacted to being labeled third person as a typical example of 80s audiophile thinking. Not a name I like to be associated with.
Indeed - but what that means is that groove angle for any given signal amplitude increases as linear velocity decreases toward inner grooves. For practical purposes groove angle = signal amplitude, curvature = signal slew rate, both of which generally increase at inner radii.
The groove angle presented is mostly the limiting factor, and this limits what is trackable dependant upon the friction coefficient between stylus and groove; to visualise, steep groove angles tend to pull the stylus up the groove wall and the outcome as to mistracking mostly depends on friction vs VTF in its simplest form.
It's all about friction as to what angle can be tolerated.
The Baerwald 1941 paper. Baerwald considered geometry for minimum HD, not friction, though in practice they are one and the same in terms of what one has to achieve for best inner groove tracking performance. Except that friction in practice involves all three axes, not just lateral angle. Look at the meaning of his HD distortion coefficient, in particular the meaning of e vs e/2, and in particular the high accuracy required as to common calculators and protractors for lateral alignment.
Consider what happens to stylus upforce as steepness of groove angles increases in the context of stylus-groove friction. It's the elephant in the room !
Last edited:
There has been much interesting discussion, for which I thank you all, even if some of it is a bit off the original topic.
So to steer it back on track, I suggest that it would be straightforward to put the anti-phase ambience (APA) hypothesis to the test by increasing the interchannel crosstalk at LF. As noted upthread, this has been put forward many times as a better cure for subsonic disturbance rather than straight highpass filtering. However, the idea has never caught on, and as far as I know has never been implemented in a commercial product. (No doubt I will be very quickly proved wrong there) In fact it never seems to have got further than the letters columns. Here's an example from Chapter 8 of my SSAD2:
35] Lawson, J "Rumble Filter Preserves Bass" Letter to Electronics & Wireless World, April 1992, p317
There have been some similiar suggestions but I don't have references to hand. Can anyone help there?
Anyway, all we have to do on the hardware side is come up with an LF-crossfeed circuit that can be switched in and out without level or phase changes. The more difficult bit will be proper double-blind testing to see if attenuating the LF antiphase information removes the 'ambience' or 'anticipation' effect.
I would be glad to hear suggestions for the spec of the hardware. At what frequency do we start crossfeeding? With what slope do we introduce it- 6dB/octave, 12dB/octave??
So to steer it back on track, I suggest that it would be straightforward to put the anti-phase ambience (APA) hypothesis to the test by increasing the interchannel crosstalk at LF. As noted upthread, this has been put forward many times as a better cure for subsonic disturbance rather than straight highpass filtering. However, the idea has never caught on, and as far as I know has never been implemented in a commercial product. (No doubt I will be very quickly proved wrong there) In fact it never seems to have got further than the letters columns. Here's an example from Chapter 8 of my SSAD2:
35] Lawson, J "Rumble Filter Preserves Bass" Letter to Electronics & Wireless World, April 1992, p317
There have been some similiar suggestions but I don't have references to hand. Can anyone help there?
Anyway, all we have to do on the hardware side is come up with an LF-crossfeed circuit that can be switched in and out without level or phase changes. The more difficult bit will be proper double-blind testing to see if attenuating the LF antiphase information removes the 'ambience' or 'anticipation' effect.
I would be glad to hear suggestions for the spec of the hardware. At what frequency do we start crossfeeding? With what slope do we introduce it- 6dB/octave, 12dB/octave??
Stereophile may have saved you the trouble. They commonly run spectra on dithered low level tones- with nearly all modern DACs that were engineered (rather than designed), you can't see anything but the fundamental. The other clue is the linearity vs level plots which are pretty flat all the way to -100dBFS. These levels are ridiculously lower than the distortion and noise floor of even the best analog devices (e.g., cartridge Johnson noise).
Sorry for staying OT, but that matches the average I have seen. A number of boutique recordings i bought in 2015 were mid 20s RMS, but with the peaks close to 0dBFS. This may not be quite as dynamic as the live recording, but pretty good.
Check some CD's from the '80s, early '90s. They observed that rule for a while.
Unlike the broadcast industry, no one has any authority on CD levels, so the fraudsters used any conceivable trick to make them sound "better" to the unsuspecting public or the musicians looking for their record to "stand out" when played in a mix. With the "free market" excuse ("this is a free world, we'll do what we please, and let the market decide").
Stretching memory again, but IIRC I read of a product under the name 'Rumbleizer' or similar spelling, which attempted to exploit the premise that, generally, vertical stylus motion below a few hundred Hz was unintentional and safe to cancel by subtraction. 300Hz rings a bell. A quick search online didn't show it up, but variations of spelling might?
But IMO bass end authenticity is but one aspect to the OP. A whole array of mastering art and engineering in applying studio tapes to vinyl ultimately underpins the bona fide nature of vinyl's original sound. Rather as two artists painting the same subject in different media are pretty much bound to achieve separate results, so as to mastering 60's/70's vinyl vs current digital media. It's inevitable: the separation in decades of sonic taste and different constraints of the media make it so !
.
I did, and the lowest peak levels I could find was dire straights brothers in arms. So finding no evidence to support your original premise on reduced resolution.
DR Peak RMS Filename
----------------------------------------------------------------------------------------------
DR20 -2.00 dB -23.87 dB 01 So Far Away.wav
DR18 -2.00 dB -22.62 dB 02 Money For Nothing.wav
DR13 -6.98 dB -22.62 dB 03 Walk Of Life.wav
DR14 -9.61 dB -27.12 dB 04 Your Latest Track.wav
DR13 -14.86 dB -30.84 dB 05 Why Worry.wav
DR17 -4.78 dB -25.27 dB 06 Ride Across The River.wav
DR14 -5.21 dB -24.75 dB 07 The Man's Too Strong.wav
DR18 -2.95 dB -23.49 dB 08 One World.wav
DR15 -6.95 dB -26.21 dB 09 Brothers In Arms.wav ----------------------------------------------------------------------------------------------
Feel free to look here Album list - Dynamic Range Database to see if you can find even one album that does not have peak levels close to FS.
Last edited:
Back to low freq. noise. Easiest "proof" I can think of is to filter out the LF of a record with music that dosnt have much LF and see if the music sounds worse. I kinda doubt it.
Can you remind me how those "DR marks" are relating to alignment level, permissible maximum level and headroom expressed in dBFS?
P.S. And something to puzzle you. Did you know that professional analog recording used to have 24dBu of headroom? It was just that the "0dB" on the meters was 24dB lower than the clipping point. Which created much confusion when digital equipment came out with the 0dB on the meters indicating exactly the clipping point.
Last edited:
Yes. Whereas there is no strict FS in vinyl mastering, no little clip light, no rule as to how hot and what unnatural compression happens above nominal 0dB and high slews. This is another key difference relevant to the OP, IMO, and was (still is !) an artistic way round the dynamic range 'limitations' of vinyl for mastering engineers.
Though shalt not exceed 5cm/S. Imagine ! 😱
- Status
- Not open for further replies.