Hi,
I thought people might be interested in my evaluation of various formats of Love Over Gold, completely ignoring any subjective audiophile language and concentrating instead on evaluating the various recordings objectively.
I surprised the hell out of myself. The LP (aside from the damned pops and clicks) is as-good or better than any other release.
See Suzy's Blog: Telegraph Roads - an objective comparison of Dire Straits - Love Over Gold in various formats.
I thought people might be interested in my evaluation of various formats of Love Over Gold, completely ignoring any subjective audiophile language and concentrating instead on evaluating the various recordings objectively.
I surprised the hell out of myself. The LP (aside from the damned pops and clicks) is as-good or better than any other release.
See Suzy's Blog: Telegraph Roads - an objective comparison of Dire Straits - Love Over Gold in various formats.
Interesting findings Suzyj, thank you for sharing.
p.s. And great choice of track and album.
p.s. And great choice of track and album.
Last edited:
It's interesting that the 1996 CD that performs so poorly has super bit mapping. If I recall correctly, besides some trick to switch between different types of reconstruction filters, super bit mapping involves rounding the highest peaks with some third-order nonlinearity and correcting for it again inside the super bit mapped CD player. Without super bit mapping decoder, the signal is simply soft clipped.
Super Bit Mapping was a vacuous marketing term. Sub-band Nyquist dithering wasn't friendly.
The analog master would have been copied for the various territories, for LP production, changing the absolute noise floor. CD/SACD could have been mastered from a copy as well.
The LP lathe compressors are sought after today.
The analog master would have been copied for the various territories, for LP production, changing the absolute noise floor. CD/SACD could have been mastered from a copy as well.
The LP lathe compressors are sought after today.
One of the few albums I overplayed in my teenage years to the point I got a second copy. Not suprised by your findings. CDs are rarely mastered to make the best of the medium and often have less DR than the original vinyl release.
Also, since a quite a detailed investigation has been made, it would be interesting to see if you could re-master any of the digital versions to be better to your liking. In the computer domain 16 bit CD is no longer a limitation.
Please ignore post #3, as I've mixed up super bit mapping with HDCD. HDCD has the soft clipper, High Definition Compatible Digital - Wikipedia
No worries MarcelvdG. As a separate exercise we could list all the marketing terms that actually don't mean anything, A lounge topic maybe.
Love Over Gold is also one of my favorites. I have a particular attraction to Industrial Disease! It would be interesting to see some spectral analysis of the musical sections to see if anything has been done.
This is actually my plan. I’ve got an interim “cleaned up” version where I’ve removed the pops and crackles from the track using the Audacity repair tool. This is straightforward as the pops are very short transients. Using the CD as a reference I’m able to discriminate a pop from a deliberate transient, like a snare drum.
However it’s extremely tedious, and while it most definitely sounds great, there’s an element of guessing what the audio is during the pops that I’m not terribly comfortable with.
So, I’ve ordered another two copies of the 1982 Australian LP. My plan is to digitise the three identical copies with identical equipment, so the speed is the same and the levels match, then to time align using a correlation function. Then I average the three time-aligned tracks sample-by-sample, which should yield a reduction of 2dB noise (assuming the LP noise dominates). To remove pops I can simply set a threshold where for each sample if one recording disagrees with the other two, it’s not used.
I’m writing software (very bad C code) to do this, as I can’t find anything to do it.
This is all predicated on the pops being randomly distributed on the LPs, and thus not lining up. I imagine there will be some that come from the master, which won’t be removed using this technique. I can keep buying LPs, and thus get closer to the original LP master.
Last edited:
This has bugged me for years, the technology for recording music is getting better but the new album releases are becoming more and more compressed. I've spent a long time at the DR Database trying to find the best versions of the music I love. Love over Gold was one of those. I've found the earliest generation CD is usually the pick of the digital copies and I stay away from remasters like the plague. Some of those are hopelessly compressed. I have also noticed that many LPs have a wider dynamic range than the CD version.
Clickrepair authored by Australian mathematics professor Brian Davies is said to work well.
The waveforms show compression applied at various levels. Tools in Audacity can reverse the process if you know to use them.
The waveforms show compression applied at various levels. Tools in Audacity can reverse the process if you know to use them.
Last edited:
I record by running the phono pre straight into a Schiit Jil ADC and use Vinyl Studio. I've found it to be better than anything else I've tried. It has the virtue of being designed specifically for the job. Noise reduction is good. I'd say click reduction is on a par with Click Repair and low level noise removal is better. Cost wise there's not much difference than with Click Repair.
My main gripe with Click Repair is that the UI looks like it's AWT based and has not kept up with the JDK, and to use it properly you'll need JDK 8 or accept some novel behaviour of lists and radio buttons with an up-to-date JDK. Core functionality works well, and being able to listen to what's been removed is a really useful feature.
I record at 192kHz/24bit and an average click is anywhere between 40 and 120 samples or 0.0002083s to 0.000625s. Very often they're in the range 60-90. I know this because Vinyl studio lists clicks it automatically repaired.
My general approach is to use as little correction as I can and to accept some noise. VS provides settings to protect transients from click removal and this usually results in big clicks being ignored by the automatic system. They can be hand repaired in VS.
Honestly, I can't hear any difference between recordings and originals except for the absence of noise. In some cases it's glorious. Particularly with older mono recordings where the absence of noise actually reveals the music.
Playback Heybrook TT2, Audiomods S6 Arm, Ortofon 2m Black, Muffsy Phono Pre.
My main gripe with Click Repair is that the UI looks like it's AWT based and has not kept up with the JDK, and to use it properly you'll need JDK 8 or accept some novel behaviour of lists and radio buttons with an up-to-date JDK. Core functionality works well, and being able to listen to what's been removed is a really useful feature.
I record at 192kHz/24bit and an average click is anywhere between 40 and 120 samples or 0.0002083s to 0.000625s. Very often they're in the range 60-90. I know this because Vinyl studio lists clicks it automatically repaired.
My general approach is to use as little correction as I can and to accept some noise. VS provides settings to protect transients from click removal and this usually results in big clicks being ignored by the automatic system. They can be hand repaired in VS.
Honestly, I can't hear any difference between recordings and originals except for the absence of noise. In some cases it's glorious. Particularly with older mono recordings where the absence of noise actually reveals the music.
Playback Heybrook TT2, Audiomods S6 Arm, Ortofon 2m Black, Muffsy Phono Pre.
My experience is that more recent remasters are actually sympathetically done because they're at the behest of and under the control of the artists and they're really impressive. But I always read around before I buy a remaster...
Rory Gallagher remasters are a prime example of what can, and should be done. I've got others by Tangerine Dream that're triffic and without the ghastly mush from those crappy crappy crappy Virgin pressings.
I have Jimmy Page's Led Zep remasters. You can hear the squeak from JB's drum pedal...
For instant gratification and a lighter wallet, there is the Sweetvinyl Sugarcube. I can see these devices being a boon for large record collections where you have to balance the amount of manual effort against the cost. Either way, it leaves those who believe digitisation robs vinyl of being..vinyl, in a bit of a quandary.
Blimey. I'd never heard of these. They're attractive in a hideously expensive and rather pervy kind of way. A bit like square crisps.
I'd rather go the software route and get a transfer I really like, album by album, portable files and stream them. But, if you've a big wallet and the urge for convenience and a need to splurge.
Now I've got the whole Arduino bug, could imagine a small camera positioned above the TT that used image recognition to load an album side by album side profile for click and noise reduction as the vinyl is placed on the TT and before the puck goes down. Yeah, I'm sure that's doable.
I'd rather go the software route and get a transfer I really like, album by album, portable files and stream them. But, if you've a big wallet and the urge for convenience and a need to splurge.
Now I've got the whole Arduino bug, could imagine a small camera positioned above the TT that used image recognition to load an album side by album side profile for click and noise reduction as the vinyl is placed on the TT and before the puck goes down. Yeah, I'm sure that's doable.
It takes me hours of careful, painstaking work to remove pops from a typical track with Audacity, carefully identifying the pop using the spectrogram, zooming right in on it, and confirming that it’s a pop by looking at exactly the same spot in a CD, then using the repair tool over the minimum number of samples to remove it.
I really doubt that a real-time automation of that process is going to be anything beyond awful.
I really doubt that a real-time automation of that process is going to be anything beyond awful.
Try it. clickrepair is a free trial. ok you have to fight java, but that should be no challenge for someone of you skills 🙂
It would seem that way. I thought that I would never own noise cancelling headphones because they had to mess with sound. Now I own a pair and they are great. Algorithms have advanced significantly. Many things people complain about can be fixed. Github for example has got great stuff for headphones.
The concept of better-than-before will be hard for some to swallow, and messing with the sound isn't necessarily a bad thing.
You will likely find this a waste of time, as Heraclitus said "You can not walk in the same river twice". The speed stability of virtually all TT's is inadequate for this task, simply play the same LP twice and try to line them up. I have not listened to LOG in a long time but there might also be an issue with tape hiss which would not average out.
I've done the one at a time pop removal it is very tedious and by the time you are done it is hard to listen to the music for a while.