3 Replies Latest reply: Nov 6, 2012 5:25 AM by voiceoverworkshop RSS

    Evening out volume of two parties in phone recording; and NR

    JayNewWeb Community Member

      I sometimes record phone conversations for business purposes. I may interview a user of our product, for instance, or a technical expert, for a promotional audio or video (we'll add graphics to the audio).

       

      I'm using a DynaMetric recording device. It does pretty decently. However, it's not perfect and it is, after all, a phone conversation.

       

      I want to do two or three things here. I mainly want to even out the volume of both parties and am looking for help in finding the best way to do that. I'm not sure if I should just manually do it, or if a compression process is just as good (it's obviously much faster), or what's the best approach.

       

      Second, I want to increase the volume overall, as will be evident when you look at the sample. In that sense, I would guess that Normalizing is advisable, in which case maybe the evening out of the volume of the two parties is best done my by selecting segments of speech from one person at a time and Normalizing that selection to, say, 95% or whatever you suggest and so on (tedious, but maybe the best approach for all I know).

       

      And, since the increase in volume will dramatically increase the volume of the "noise" or "hiss" that's there throughout, I assume a Noise Reduction process is advisable, too. I know how to do that by taking a "profile" from a sample selection and applying it to the whole file. Assuming that's the best way to proceed here, I don't know if it should be done before I smooth out and increase the volume or after, or if there are any other pointers I should know about.

       

      Here's a sample from a file I'm working with:

       

      www.silver100.com/x/sample-for-vol-and-NR.wav

       

      Thanks!

        • 1. Re: Evening out volume of two parties in phone recording; and NR
          voiceoverworkshop

          JNW, I have no experience with DynaMetric, but have decades of experience recording telephone conversations in a variety of broadcast and studio settings.  If you do (or intend to do) a lot of this, and especially if the output is used for video or product reviews (where poor audio quality can defeat the purpose of the product review), I suggest getting a decent podcast style microphone, a small 2 or 4 channel outboard mixer and an outboard piece of telephone interface gear (Gentner comes to mind, but there are others, maybe even DynaMetric).  There's some detail in how to capture the call through the interface which is easily learned, but too lengthy for here. The gist of it is, run the phone line (with caller) through the telephone interface and into a channel on the mixer, where you can control the callers volume, and tweak the eq (if the mixer supports that). Run your voice through the microphone, into another channel on the mixer, where you can control your volume and tweak your eq if the mixer supports it. Run the output of the mixer to Audition.  With good use of eq and the ability to adjust volumes independently (combined with Auditions restoration, eq and mastering features) you'll be surprised just how good your phone productions can be

          • 2. Re: Evening out volume of two parties in phone recording; and NR
            SteveG(AudioMasters) Community Member

            voiceoverworkshop wrote:

             

            The gist of it is, run the phone line (with caller) through the telephone interface and into a channel on the mixer, where you can control the callers volume, and tweak the eq (if the mixer supports that). Run your voice through the microphone, into another channel on the mixer, where you can control your volume and tweak your eq if the mixer supports it. Run the output of the mixer to Audition. 

            Nothing wrong with this at all, but I think that one particular point isn't perhaps quite as clear as it could be - and that is that ideally, you should pan the channels (assuming it's a stereo output) so that for the initial recording, the caller is on one channel, and that you are on the other. That way, you don't have to worry so much during the interview that you're keeping all the levels exactly as they should be, and can concentrate on the interview itself. Then, after the event you can optimise just the caller channel so that it sounds best, without worrying about having to select bits of caller from two channels mixed.

             

            The other thing I should mention is that the best results I've had from doing something similar involve a small mixer, with outputs panned pretty much as described, and using Skype. With a little care, you can almost get it sounding like the caller is in the studio with you.

            • 3. Re: Evening out volume of two parties in phone recording; and NR
              voiceoverworkshop Community Member

              That's a good point Steve. In addition, a pro soundcard would raise audio quality and provide the ability to record the interviewer and interviewee on separate tracks and process each track independently as well. Recording in mono is a must, and at 44.1, 32 bit wouldn't hurt either. Other factors contribute to audio quality, such as room treatment, but I'm getting off topic from the original question.  I suspect the DynaMetric unit JNW refers to, records the call on a stick or some sort of internal memory card, as an mp3 or compressed format, which has to be brought into Audition, rather than be recorded there in the firat place. Thanks.