Recording Mono Vinyl: Summing L+R Channels

Here is a summary of an archived thread called When Stereo Is Mono... (No wonder it took me forever to find it!)

By our esteemed leader, Steve Hoffman:

Am I glad I stumbled on that archived thread! I didn't know about this before. I used to transfer mono vinyl in stereo: L channel left, R channel right. After reading the above comments, I tried various methods of summing the channels. Wow!!! What a difference! As explained above, summing cancels the constant "swoosh" of the vinyl (which, when heard through headphones, sounds like it's about to push my eardrums in). Also as explained above, summing can reduce distortion products as well. When I tried it, it dramatically reduced sibilance and high-end distortion. The music just seemed to "pop" into focus out of a nice quiet background. Remarkable!

Now for my comments on the various methods of summing the channels. You may have noticed a difference of opinion above about whether to do this in analog (using a Y-adapter) or digital (using your mastering software). Here's what I found out.

First I should tell what equipment I'm using. I have the same crappy Dual 505-2 table and Stanton L725E cartridge I had back in high school. (The Stanton cart got a new elliptical stylus in about 1992.) This is connected to a Musical Fidelity X-LPS v3 phono stage. The signal then goes to an M-Audio Audiophile 2496 sound card, whence it is recorded at 24-bit/88.2kHz using Peak 3 recording software. The software runs under Mac OS 10.3.

I compared the following methods of summing the channels: (1) recording in stereo, then folding the channels to mono using Peak's DSP; (2) recording in mono "on-the-fly" using Peak's recording settings; and (3) using a double Y-adapter between the phono stage and the sound card, and recording the resulting signal in stereo (creating a dual mono sound file).

The winner? Method (3). For some reason, method (2) didn't work at all. I got a mono recording, but with the stereo noise left in. All I can think is that Peak only recorded one channel. Methods (1) and (3) both worked to remove the stereo noise. Actually, the results were startling, with surface "swoosh", sibilance, distortion all disappearing as if by magic. (I played the stereo and summed-mono files back for my wife, who was amazed at the difference.)

However, method (1), the digital fold-down, resulted in a "digital" sound, IMO. This may be the fault of the software I use, but I find that DSP often leaves a hard, edgy signature. (Sample rate conversion and normalization both do this in my experience, and I wish I didn't have to use them.) Method (3), summing in analog, retained more of an "analog" sound.

I made my own double Y-adapter by buying 2 4" male-to-female Y-adapters and 2 4" female-to-male Y-adapters from Radio Shack. Then I crossed them so as to split and recombine the channels, and inserted them between the phono stage and the interconnects that carry the signal to the sound card. If the Rat Shack adapters degraded the signal, I couldn't tell; the improvement caused by summing was so remarkable that the effect of the adapters, if any, was undetectable.

By the way, I also tried folding the dual mono channels that resulted from method (3) into a single mono file. This did not bring any improvement. I heard a slight degradation, if anything, plus the addition of that unwelcome "digital" sound.

I've transferred a few mono vinyl recordings in stereo over the last couple of years. Time to do those over! Dual mono, here I come!

 

If I'm going to do any manipulation, I prefer to keep it in stereo. Why? Sometimes one groove is cleaner than the other due to uneven wear, or sometimes a scratch plays in one channel and not the other. In the latter case, I can just copy/paste a split second of one channel into the other, and it's clean.

If I were just doing a straight needle drop with no manipulation, I'd definitely go the analog mono route. It's amazing how much of the noise is cancelled out that way!

 

So you do the manipulation, if any, in stereo, and then sum to mono using DSP?

 

Summing the channels does indeed give a great improvement on most mono disks; if the disk is in really great unplayed condition, there's not a lot in it but it does reduce all the noise/hash/distortion on worn and dirty (not that your disks should be dirty) ones.

Note that by summing the signals from a stereo cartridge you are cancelling with two *different* signals; although gross errors are rare, the left and right channels of a stereo cartridge are not identical in response.

There are a few genuine mono pickups about now, along with vintage ones for the well-monied, from Lyra, Grado, Denon and Ortofon. These are very worthwhile on older mono disks that aren't cut on a stereo lathe (like they all are now). Generally they have optimised mono styli too.

There's the old archivists trick of using one channel only on really damaged disks - these days they record in stereo and flip the channel to the quietest as often as they like for the 'best' result, in this context.

 

For most of my mono vinyl, I simply hit the MONO button on my receiver and away I go. However, some records get that swishy out of phase sound and/or really harsh sibulence. In those cases, I just burn it in stereo, pick whichever channel sounds better, and copy/paste the better channel to the other channel in EAC (after deleting the inferior channel first of course).

 

Hmm. I have a few of the older mono discs you mention -- cut in "true mono", so to speak. Why is it better to use a mono pickup, as opposed to summing the L+R channels? I'm not doubting your word, you understand -- just wondering what the difference is.

 

Don't forget! If you sum to mono in analog using a double-Y cable: Before you burn the audio to a CD pick one channel - doesn't matter which - and make that one both left and right. No processing is involved when you do this so you won't get any digital artifacts. There's no need to burn the noise or any channel imbalances of the sound card to a CD, right?

 

I didn't think of that, but it makes perfect sense. I'll try it!

 

Sure, give it a shot. It works great! It's similar to what the pros do when they play back a mono tape using a stereo head. They pick the best channel (hopefully, sometimes they screw it up) and make it both during the A/D conversion. That's why if you take a properly mastered mono CD, extract the audio, and then invert and overlap the right channel over the left...you're left with nothing! They cancel each other out perfectly.

 

There seems to be a difference between 'electrically' losing the unwanted (vertical) noise and distortion by summing the channels and 'mechanically' losing it by having a cartridge that only responds to lateral movement.

Users I know report much less noise and better sound from most mono records.

A stereo cartridge works by having two generating systems mounted orthoganally, so that they only respond to the separate channels. A mono cartridge only responds to lateral movement, completely different geometry to a stereo cartridge. Most early mono records have a different and larger groove profile, and some of the new mono cartridges have a larger stylus to fit this better.

This difference is appreciated enough for Lyra to be making a special $5500 mono version of their Titan cartridge.


Hope this helps.

 

OK, I get it. A dedicated mono cartridge -- one more item for the wish list.

Another advantage: only having to remove the clicks from one channel! That's what I'm doing now. Then comparing one channel to the other to hear the difference.

 

I used to do method (1), and I had used method (2) a few times as well. But these days, I use pretty much method (3) - but I currently use only one male-to-female Y-adapter and one female-to-male Y-adapter on my setup. This setup sums the channels into a single channel before spiltting that single channel into two.

This is a bit different from your method, which splits each of the two channels into two more channels before combining each split from the original left channel with each split from the original right channel to form new left and right channels.

In either case, the channels should be summed in the analog domain - not the digital domain - for the best results.

 

I don't know about the date range.. most of my friends apply 'try it and see'. It's a second cartridge (and turntable setup) for most of them, or maybe just an arm tube. This isn't interchangeable headshell territory. Sometimes the stereo cartridge playing is preferred.

One I know is used a lot for mono 60's singles.

I'm not in that price range now (though I was in better times) but I still enjoy thinking about it, expecting those times to return. The wonders of 'a career in IT'.

 

And if you are going to sum in the digital domain make sure you record in 24-bit and do your transforms in 32-bit. You might want to do it this way if, for example, you want to use only the best portions of each channel. Also, mono songs with a lot of crackle tend to be easier to declick in stereo because you don't have to deal with the sum total of the clicks and pops overlapped onto one track.

 

I went to Radio Shack and bought a pair of Y-adapters so I could hear the real mono on some Beatles vinyl. Sure enough, the sound reproduction is awesome.

My problem is I picked up some new hum despite cleaning the connections and ensuring that the connections were secure.

Any recommendations?

 

IMO, if the record is in good shape (clean, good pressing) sum the channels in analog and record to best A/D convertor you have (DAT, soundcard, etc.) at 16/44.1 (avoid dithering/processing color). De-click by hand (if you choose) and avoid any processing like normalizing, etc. Configure your tracks, burn to CD and enjoy. Less is more.

 

I combine the L/R channels with a Y-adapter when I record. This works the best for me on mono vinyl. Once in a while, if one groove is really damaged but the other is ok, I'll record in stereo and copy the good channel over the bad one.

Not only has my background noise been reduced, but the records seem to have more punch as well.

 

My approach has been different. I have a spare V15 Type V and got hold of a couple of little "straps" (small metal strips) that connect L to R channels and L to R grounds together, making the cartridge output mono (essentially, dual mono). Seems to work fine. No further manipulation and no extra cables necessary to get a nice, quiet mono signal.

 

I record in mono on my MD deck. Not audiophile quality, obviously, but it works for listening on the road and in situations I just want to listen to mono music, if I want pure mono quality or if I want to play mono LPs, I can just turn off one of my phono stereo channels on my DJ mixer.

 

I cannot sum the channels in analogue. If I attempt to, using two y connector chords, a hum appears in the chain. Local audiophile store told me the hum is a result of the Rega turntable not having separate ground wire. The ground is in the left/right RCA cables. As a result of this all of my mono needle drops over the past year were summed during the recording phase in Cool Edit Pro.

Is there a way to get rid of the hum?

Todd, it was my understanding that I should be recording in 32 bit float (88.2 not 96), "cleaning up" the sound and then converting to 16 bit/44.1. Someone mentioned to me that if you record in 16 bit and then do anything to the files, you will end up with less than 16 bits when the file is saved.

 

No.

 

It depends on what you do to the file. If you record at a low level in 16-bit (say, it peaks near -12db) and then amplify in software you'll end up with less than 16-bit audio. If you have a way to set your levels in the analog domain so that you're up near 0dB then you'll have no problems recording in 16-bit mode. Manually declicking, as Todd suggested, can be done without any harm to a 16-bit file because you're working on pieces of audio here and there that are just a few milliseconds in length. Now if you wanted to do some heavy duty decrackling, or EQ'ing, or noise reduction, that would be a different story.

 

For what it's worth, I wouldn't worry about it too much. (And I worry about everything! ) I've never found it to make that much of a difference especially when working with 24-bit audio in the form of 32-bit files. All the software is really doing is finding the average of the left and right sample values. It's not perfect but there's not much that can go wrong. Although some people feel that anything "wrong" is not good enough. I suppose it comes down to how much you want to stress over it.

 

I actually do from time to time a bit of EQ'ing and a wee bit of decracking. Never use noise reduction. I try to manually declick where I can. However it's a bit difficult when you're dealing with a quiet LP passage that has faint crackling fireplace type sounds in the background.

Is there some problem dithering down from 32 bit to 16 bit that I'm not aware of?

I'm not worried about any recent needle drops, however I want my needle drops to be as best as they can be. If summing the left and right channels through the analogue stage is the best way to go, then that's would I should be trying to do.

Again, is there a way to get rid of the hum from the Rega turntable?

Thanks.