Academy Dashboard Forum Production Mastering Correlation Meter/Stereo width

  • This topic has 6 replies, 3 voices, and was last updated 5 years ago by Erik Duijs.
Viewing 7 posts - 1 through 7 (of 7 total)
  • Author
    Posts
  • #48462
    Daniel Ashcom
    Participant

      Hi everyone,

      I'm taking my first attempt at mastering a track I mixed and using Voxengo's Span plugin and comparison with reference tracks, I see that my correlation meter is at or very close to +1 through most of the song, whereas two of my reference tracks, while touching +1 regularly, fluctuate much more between zero (or just below) and +1. I looked this topic up and know very little as of yet, but doesn't this mean that while the song will sound fine in mono, it has a very narrow stereo image?

      What sorts of things would I do in recording and mixing (or mastering, but I'm under the impression it's the first two) to ensure as wide a stereo image as possible? Or shouldn't this be a concern?

      I recorded bass DI, drums with 1 mic and double-tracked guitar with one mic and I'm not sure whether this approach has affected the stereo width, hence sharing the details.

      Thanks,

      Dan

      #49231
      James Gorman
      Participant

        Hi Dan,

        > I looked this topic up and know very little as of yet, but doesn't this mean that while the song will sound fine in mono, it has a very narrow stereo image?

        Yep, this is pretty much it. In and for its self there is nothing good or bad about being mono (some very good records are mono!). The issue is only if you are a trying to fit into a genre that tends towards stereo.

        That said, the short version is, does it sounds good? If it does you're set, if not then you'll want to look at things.

        Thing you can look at for stereo image (given what you've got as input):
        - are the guitars hard panned left and right?
        - If they are, try setting them to centre and flip the phase of one to see how much cancels out. If a lot is cancelled out then they are too similar and you might need to rerecord one. You could use different guitar, cab settings, a different mic placement, or playing a part with small variations.
        - Is there reverb on things? With mostly mono mics, only the guitars can give you much space if you have a typical pop/rock mix of drums and bass down the middle. A short, diffuse room-y reverb might add a sense of space to things (of hall/chamber depending on the track)
        - Is the bass reamped (or virtual amped)? If you try Warren's high pass trick, you can feed this to the reverb. Some v-amps let you play with a stereo room mic too.

        Are you confortable sharing you track (and reference)? I'd be happy to have a listen and see if I can help more.

        James

        #49241
        Daniel Ashcom
        Participant

          Hi James,

          Thanks for taking the time to help with this. I'm happy to share the track but please know that I'm still finishing up the mixing stage as it turns out with help from a couple people so I'm not quite to mastering as I'd thought. You can see the disaster that my experimentation gone wrong turned it into as well as its current (much improved) state here:

          https://soundcloud.com/candidategalaxy/orange-shades-5

          I actually re-recorded the hi-hats and toms today but haven't uploaded the latest version yet. I only have one amp and one electric guitar and my audio interface has 1 mic input so I can't help but think that's contributing to the mono tendency. The guitars are double-tracked hard left and right.

          For my references I have The Black Keys' 'Everlasting Light' (guitar, drums, stereo field),the Walkmen's 'Wake Up' and Beck's 'Gamma Ray' (bass, drums). The bass is going directly into the audio interface and Guitar Rig. I'm following Warren's bass mixing tips from Youtube with splitting low- and hi-pass so the hi-pass signal may have reverb already -- I will check that as well. It's late here so I will do as you suggest to check the phase tomorrow morning as well as upload the latest version.

          Regards,

          Dan

          #49390
          James Gorman
          Participant

            Hey Dan, I had a listen to yours and your references. Listening to the latest it looks like found most of the stereo field issues. Two things that strike me for bass and drum in particular:

            - if you feed bass and drums to shared reverb with a HPF and LFP before the reverb you'll open them up a bit more; this is a key feature of the Walkmen and Beck tracks. Beck is also doing interesting things with the width of his reverb (see below)

            - it sounds like you've panned the drums really wide. Nothing wrong with doing this, but it is at odds with the live/garage-y vibe of the track (and references). It also means in the parts with the stereo guitars, they are on top of each other and there is a whole lot of empty space between the mid and sides. If you want to play with this you can strap a stereo widening plugin like Flux:: Stereo Tool or Hofa 4U Meter Fade and Pan and move the drums around.

            If you get Melda's MStereoScope, you can get a full stereo meter, and the ability to solo mid/side/left/right so you can look at what's happening other songs.

            #49398
            Daniel Ashcom
            Participant

              James,

              Thanks for taking the time to give me feedback on this. I haven't recorded or mixed drums before and honestly I hadn't considered panning and didn't realize it was panned much wider than my references. I tried panning things less wide and you were right -- it sounds a lot better for this track. The hi-passed and low-passed reverb shared between drums and bass took a lot of tweaking of the bass level not to overwhelm the drums but I really like the result though I think a lot more practice with other tracks is required to get closer to a bass sound I'm satisfied with overall. Would you typically also send drums to a separate, second, reverb track in this context? I figured it would be too much.

              I'm not totally sure how I feel about it so I'll let it settle for a bit and come back with fresh ears. Here's a version with the tweaks I made: https://soundcloud.com/candidategalaxy/orange-shades-2

              Regards,

              Dan

              • This reply was modified 5 years ago by Daniel Ashcom.
              #49437
              James Gorman
              Participant

                No worries Daniel, we've all got to start somewhere!

                Typically if I need a drum reverb I will either have it as part of my drum buss and send individual drum channels there, or send the drum buss to a dedicated drum reverb. On Crazy I used a single room reverb for everything (though vocals had a bunch of extra reverbs and delays), and a dedicated snare reverb to get the timing of the snare tail how I wanted it, but this was a spur of the moment thing because I really wanted to emphasise the all-in-one-room thing already going on. For your case, I think the single reverb works since it keeps that live feeling. If you were doing pop, or even just something more complicated, you might want to use different reverbs, even if they are just different versions of the same plugin with different settings.

                Wow, the space it much better filled up in this version. The whole thing sits a lot better in space now. Nice work! Listening back through your version can definitely hear the improvement. Taking a break is always a good idea, be good to hear where you end up with this

                #54605
                Erik Duijs
                Participant

                  My understanding of it is this:
                  * The correlation meter measures the linear correlation between left and right signals. So if the signal at one point is 100% on the left and 100% on the right, the correlation is 1. If 100% left and -100% right it's -1 correlation. Etc.
                  * So the correlation meter says something about the stereo separation. If correlation is always 1 then there's no stereo separation: It's mono. If the correlation is always -1 then there's also no stereo information, but the channels are in inverse polarity (meaning the track will be silent when playing back in mono).
                  * Remember most of the energy of audio is in the low-end of the spectrum, so the correlation meter correlates mostly to bass.
                  * So you could say the more 'stereo' it is, the more the correlation meter will go towards 0.
                  * Also remember that bass frequencies have very little directional information in it (because the waves are so long). There's a reason why it's rare to see more than one sub-woofer, and if there are they likely get the same signal.
                  * But depending on the music, you'll probably want to keep it on average well above 0 because chances are that bass frequencies might often be cancelling out between left and right otherwise. And again, most of the stereo image is not in the bass; it's in the higher frequency ranges.

                  Personally I start worrying if the correlation meter is all over the place and often gets even well below 0. Because that usually means to me that I have an issue in the bass. Maybe a panned bass-heavy track is out of phase or something. Maybe it means I have to look at stereo effects (such as reverb) on bass-heavy instruments (such as bass guitar, kick, synths, or even guitars).

                Viewing 7 posts - 1 through 7 (of 7 total)
                • You must be logged in to reply to this topic.