I found this description of part of the half speed mastering process by Miles Showell of Abbey Road studios (from What HiFi). Interesting that it goes from original tape via digital, which aids the correction process, before back to analogue for cutting. Also some mention of the low frequency limitation when done originally the analogue way.
And when we travelled to Abbey Road Studios to listen to Showell’s recent work, remastering four of Brian Eno’s solo records at half speed, we understood it was all worth his efforts. Here, he explains the whole process and how he helped it start again
The Eno job is a good example, because I was given access to the original tapes – a couple of originals are missing, but mostly I had the masters. It’s getting increasingly difficult to get hold of original tapes now, because people are aware that continually playing old analogue tape is a bad idea because it starts to wear away.
I have a fantastic Ampex tape machine, which just sounds gorgeous – I’ve got custom heads for it, which make it sound even better – so I was able to do a really nice transfer. I captured it at [hi-res] digital, 192-kHz/24-bit and, from there, did any repairs that needed doing. With clicks and extraneous noises, I had to do some de-essing to soften the ‘S’ sounds of the vocals to avoid sibilance on the pressings.
When I did the transfer, I was only thinking of records: there’s no excessive limiting or extreme compression, which you might do for a CD release to make it sound loud. No need for that on records – full level digital is too loud to cut as is, anyway, so it’s completely pointless to add further compression and further limiting to only bring it down even more.
The beauty of doing a high-resolution digital transfer is you can go in and micromanage the audio. One of these tapes had been played on a faulty machine that actually damaged it and left clicks all over it, so I went in and took out all the problems that had been created, which you couldn’t do cutting live from the tape. The only way you can fix that is digitally, and it’s next to invisible mending.
Once I was happy with the files, I started with the cuts. I have a recently restored and customised Neumann lathe in the room, which is absolutely beautiful – the best I’ve used – and I have a customised RIAA amplifier and filter.
If you’re cutting everything at half speed, you’re playing the source at half speed and running the lathe at half speed, then the frequencies and the filter are in the wrong place: you need a special one to do half-speed cutting.
When you’re doing a half-speed cut, you can’t hear what’s going down because it’s all slow and just sounds awful. So we insist that every client has an acetate. I’ll sit down and listen to it end-to-end on a domestic deck to make sure it sounds okay – there’s no point playing it on high-end [turntable], or even on the lathe pickup, because it doesn’t represent the real world.
From there it goes off to the client, and if they’re happy I’ll cut the masters. In this case, they got shipped off to the Optimal plant in Germany, probably one of the two very best plants in the world, so I was delighted when it went there. Six-to-eight weeks later I get test pressings back.
You do all the prep work first. I’ll go through and spend hours just checking things; I have to de-ess before I cut at half speed because the limiters don’t work. They’re looking for two things: listening for a specific frequency band, which obviously is not there any more; and also any acceleration, sudden fast transients it thinks are a bit sharp.
Because you’re going so slow it shouldn’t have any sharp edges. So any vocal you think might cause distortion with your sharp, bright ‘S’ sound, I’ll have to pre-treat that. But in a way it’s good, because I can use a specific digital tool which will allow me to just find the ‘S’.
Anything around it – any bright guitar or snare-drum hit or tambourine – is totally untouched, I’m just working on the vocal that’s going to cause a problem.
Then it’s just a purely mechanical job to cut it at half speed, which is pretty dreadful to listen to when you’re in the room. I’ve had 12-hour days where it’s just non-stop and you walk out of here going stir crazy – but when you get the records back and they just sound so open and so fresh, it really was worth doing.
All the hard work is done prior to the cut. Then I’m just checking that the lathe is working okay and the swarf, which is the waste product that’s being cut out of the disc, is going up the vacuum pipe alright and nothing looks untoward. Once I’ve finished, I can look under the microscope and check the groove is all fine.
The problem we had with some of the earlier ones, in the 70s, was people would be doing it live from tape, which sounds lovely but they have a bit of a roll-off on the bass end, from 30Hz downwards, which is just a factor of the machine.
Obviously at half speed that suddenly becomes 60Hz, so you ended up with these really clean, bright records with not much bass. The advantage of my method is that we have all of the bass and all of the top end, and everything just sounds nicer. Because nothing’s getting pushed to its limits everything sounds so much more real
(To me a question is are these remastered digital versions sold as HiRes, avoiding the vinyl step - and if so is the tweaking such as dealing with the ‘S’ sounds that cause problems with vinyl also beneficial In digital replay, or a negative effect compared to tge original recorded sound?)
