Do you agree with this statement I found on the net and how frequently might it be significant when dealing with a 2 channel recording of a monaural source–typically a pre-1958 recorded performance:
“If the source is a stereo tape or CD, but the sound is mono, don’t mix L & R into a mono signal. Annoying phasing will occur depending on the tape transport. It’s better to record to a stereo WAV, open it with an audio editor and save the best channel in mono”.
How often is “annoying phasing” noticeable?
On 1 song in 4? 1 song in 100?
In the real world, how often can one detect a material difference between L and R channels by ear?
Etcetera.
Brian Davies makes this comment in the manual to ClickRepair:
“Merging will remove quite a lot of vertical low frequency noise”.
Over 80% of my song files are pre-stereo era, but nearly all are 2 channel. I have occasionally merged L and R, but am not certain there are audible advantages. The saved space is not an issue in my case.
So, should I merge to remove vertical low frequency noise? Or should I not merge so as to avoid “annoying phasing”?
I don’t see any reference in Davies’ manuals to simply selecting the better of 2 channels. Do any of you do that in actual practice?
You only get annoying phasing problems if your stereo performance had significant differences between “left” and “right” and you listened to the show on a mono speaker. Minor phasing problems show as muffled, woolly sound.
And yes, you can decrease noise significantly by adding up the two channels.
It’s hard.
We talked about making a phase meter where you could tell instantly if your show was in trouble or not, but doing that requires graphic management, not just throwing audio bits around here and there and is much more difficult.
This is typical of a stereo show with some portions or instruments “off center.” If it was a straight mono show, the green line would be straight up and down. This would tell you immediately which technique to use.
But alas, there are no good answers without extra tools. There are unconditionally safe choices. Always use Left. It may be slightly noisy, but the actual show will not be damaged.
Thanks for the response. Please clarify on the following points if you can:
Do you mean that if one listens to a mixed L/R mono performance on 2 speakers, one is immune from hearing “annoying phasing problems”–in which case there is no downside to mixing?
The obvious question is “why left”?
Most importantly:
For pre-stereo era (pre-1958 primarily) 2 channel recordings, what is the recommended best practice??
1: do nothing; stay with 2 channel as is
2: mix L and R
3: Always choose left channel and convert to a 2 channel recording, with each channel of the new version containing only the old left channel
4: Always choose left channel and convert to a 1 channel purely mono recording, with no right channel at all
5: other
I’m just looking for a distillation of the best practice given my source material. I certainly don’t want to mix and then wish I had not.
I invariably listen on 2 standard speakers in a home sound system or through standard headphones.
What do you in fact do personally and what is your reasoning for that decision?
The source you quote does not mention mono vinyl or shellac recordings - only tape or CD. I disagree with lumping CDs in with tape. Unless the original recording used for the CD had phase problems, the CD should not have phase problems. Tape is a different matter. Misalignment of the head azimuth will cause phase problems.
Also, when transferring tapes or CDs, you are not concerned with “vertical” groove noise.
You’ll have trouble detecting the difference between the L and R channels by ear, but if you mix them together and there are phase problems, you’ll hear it right away.
My advice would be to follow Mr Davies’ advice. But if you’re concerned about phase problems, listen to a portion of each resultant mono file before discarding your original captures.
No. The problem occurs when you mix the L and R channels of the capture. At that point it does not matter if you listen with one speaker or two, the damage has been done.
When a mono recording is reproduced using stereo playback equipment (such as a turntable or tape deck) it is possible (but unlikely) that the equipment may be mis-aligned, and this may result is frequency-dependent phase differences between the L and R channels. When you sum the L and R channels to mono, if there are no phase problems all you get is exactly the same show, only louder. If the are frequency-dependent phase anomalies, then when you sum the L and R channels some frequencies will get louder (normal) and some will get quieter since the L and R channel waves don’t “line up”. If the problem is severe enough that some frequencies are inverted, those frequencies will cancel out.
To hear this effect, proceed as follows.
Import any music track.
From the track drop down menu, choose “Split stereo to mono”
Delete the bottom track
Duplicate the first track
Click the Skip to Start button
Zoom in until you can see the individual dots the represent the digital samples.
Using the Move tool, drag the bottom track to the right by 5 samples.
Thanks for the comments. I’m not sure I grasp it all, but let me give you some more details.
You state you would likely follow Davies, which I take to mean that you would in fact combine L and R and then possibly listen for phase issues in the result. The advantage of Davies recommendation being that it can reduce vertical groove noise.
Here is the possible fly in the ointment:
My source material is digital mp3s that are typically 2 channel mono era recordings.
The mp3s are not my own captures from my own vinyl, shellac, tape, or CD.
The mp3s in question are either from a digital store such as Rhapsody or from someone else’s CD/vinyl/shellac collection. Most often from CD, but not entirely.
Reading between the lines of your posts—I assume that in my case, given my source material (digital mp3), there is no “vertical groove noise” to be reduced and that therefore combining L and R per Davies could not improve fidelity?
And that mixing L and R remains pointless on captures from my own tape or CD collection, even if the “original” recording was “mono”, recorded in say 1950? I do have a lot of this stuff, but haven’t yet begun transfers to PC.
And that mixing L and R would be recommended on any captures I might personally do from my own “monaural” vinyl and shellac discs per se?
I think that is what you mean–reading between the lines–but I could be misinterpreting?
Please advise on those specific points if possible.
Your MP3s might contain vertical groove noise depending on how the captures were processed before converting to MP3. If the MP3 is mono there is obviously no point in mixing L and R. If the MP3 is stereo, it might or might not exhibit vertical groove noise. If you import a mono MP3 into Audacity it will show up as one track. If you import a stereo MP3 it will show up as two tracks - but those two tracks might or might no be identical. An easy way to tell is to select Split Stereo to Mono from the track drop-down menu, invert one of the channels and listen to the result. If you hear nothing, the L and R channels were identical. If you hear a very faint version of the recording it could be that the L and R channels are not exactly at the same level, but are otherwise identical. If you hear mostly ticks, pops and crackle then the recording would benefit from mixing the L and R channels together.
As for your CD and tape collection, are these commercial releases, or captures you have done yourself? Whatever the case, you can apply the same test as with the stereo MP3s to see if there is any vertical groove noise present.
When you get around to capturing your own vinyl and shellac recordings, you can again apply the same test for vertical groove noise.
It is ultimately up to you how much effort you want to put into this process to get a result you are satisfied with.
– Bill
PS: I have not done any of this myself (all my captures have been from stereo vinyl from the 60s onwards), and have not read Brian Davies’ recommendations closely on dealing with old mono recordings. That said, I still think that he’s the expert, having spent much time capturing and examining many recordings and finding the optimum processing methods.
My mp3 are nearly all two channel and made from recordings of the pre-stereo era. They do display as 2 channels in Audacity, but I virtually never can see a difference in the 2 wave forms or detect any channel differences on headphones. In that sense, they are mono–but always 2 channel. None of them are taken from “reprocessed stereo” sources.
I will try your test using inversion on some random samples. I would guess I would hear little or nothing.
The CDs are commercial releases. I have maybe 500, but in reality I’d guess that fewer than 50 will ever be ripped as I already have the material on mp3. I realized that I nearly always only care about 1 or 2 tracks per CD or LP and thus have found it more time/cost effective to obtain those 1 or 2 tracks again in mp3 form, rather than rip from my own CDs, 45s, 78, or vinyl. All mp3s are just single songs–I drew the line in the sand and said “don’t obtain an mp3 for any song you don’t care about ever hearing again”—which eliminates probably 80% of my CD/vinyl/shellac collection.
I have only maybe a dozen cassettes left and they are all made from my personal LP/45/78 collection–which has continued to languish unexamined for a long time.
I also have maybe 50 minidiscs made from my own vinyl. They are all MONO because nearly all of the vinyl from which they were made was mono and because mono gave me double the running time on minidisc.
Yeah, the effort thing is a key point. Even if I got some minimal improvement on a few songs by combining L and R, I don’t know that it would be worth my time to wade through 20,000 mp3s to do the combining.
I’m asking as much for future mp3s as anything. In reality, most of my mp3s will never be opened in Audacity—but I’ve wondered if I happen to have one open for general editing, should I maybe combine channels while I have the opportunity.
I"m nearing the end of my 21 day trials of Brian Davies’ apps and have not yet decided which if any to buy. ClickRepair is the most likely candidate.
Ah, there’s the rub. Do you know if your minidisc recorder recorded one channel only (usually the left channel if so) or summed the L and R? In any case, there’s nothing you can do about it now.
I transferred 2 or 3 particularly prized minidiscs 2 or 3 months ago.
I have to assume they showed as 2 channels when captured in Audacity as if they had not I would have remembered it—I can’t recall ever seeing anything I had show up as single channel in Audacity other than the handful of tracks I have deliberately converted from 2 channel to single channel.
Well, let me actually check by opening an mp3 derived from one of those minidiscs in Audacity:
OK; they show as 2 channels in Audacity.
Somehow, that doesn’t make sense, but it is what I see. But as you say, I can’t do anything about it and the vinyl source material was “mono” anyway.
I’m not familiar enough with minidisc technology to know what’s going on. All I know is that they are vaguely similar to floppy discs and (at the time) offered 70 odd minutes “stereo” and 150 odd minutes “mono”. I chose mono. I think the running time is considerably above that now.
I guess they may be about extinct? I thought the ordinary entry level $200 unit I have did a super job on fidelity.
Yes, they would show up as stereo (2 channels) when transferred to Audacity if Audacity was set to record two channels. It doesn’t matter - the point is the L and R outputs of those mono minidiscs will be identical.