1. Welcome to the new MacRumors forums. See our announcement and read our FAQ

Best practices for digital audio recording for voice overs and podcasts

Discussion in 'Digital Audio' started by VirtualRain, Jun 15, 2009.

  1. macrumors 603


    I'm a veteran videographer but a noob at voice over audio and podcasting.

    I would like to know if there are any good web references on the internet that discuss best-practices for vocal recordings.

    In terms of gear, I've got a couple of mics I'm evaluating... the Audio Technica AT2020 and the Rode NT1-A. I'm using a Lexicon U22 USB audio interface to connect the mic's to the Mac. I'll likely be using Final Cut Studio (Soundtrack Pro) for my audio work.

    Some questions:
    - Any comments on my microphone choices? At this point they sound the same to me and the Rode is twice the price of the AT.
    - I assume it's best to record at the highest gain possible without clipping? Or do you ensure you are several DB below clipping and then use a compression effect or something to maximize the level?
    - What's are considered essential effects to add to a voice over mix if any?
    - As requested above, are there any great resources for getting up to speed on best-practices in audio recording in a home DAW studio?

    Thanks for any thoughts, advice, or resources you can provide!!! :)
  2. macrumors 601



    i shoot for lower levels, between -36 and -18. but, that's for music tracks when there will be many mixed together. for a voiceover, i'd probably go for -8 or -12. even then, you'll get some interstitial peaks going over 0.

    once recorded, you don't need to compress or maximize it to get more volume. you can simply turn it up. what you want to do is get it right in the performance, not try to fix it after. so strive to keep your tone, tempo and volume consistent.

    unless you're going for a specific effect, i'd say as little as possible. the best thing to do is track in a treated area and try to emulate what you'd hear on, say, NPR. if you don't feel like dropping any extra money on something like a realtraps portable vocal booth, then maybe set up your mic so you're facing a rack of clothes. that should add a little deadness / "NPR" quality.
  3. macrumors G4

    I think if you are recording with 24-bits per sample you can afford to use lower levels to make shure you don't get any clipping at all.

    But I think what has more effect on the sound are little things about the environment. Are the walls reflective and "hard" or softer? The distance from the the person speaking to the mic. Use or not of a "pop filter". If you listen to pro recording engineers they will talk about little details like the specific brand of mic or preamp. The is because they have already taken card of the basics. By getting the basics down you are 95% of the way there, The engineers are talking mostly about the last 5%

    The other thing is the voice actor. I think people who do this well know how to speak into a mic. I can't. But the trick is not is make annoying plosive pops and hissing "S" sounds. Kind of like with musicians, some are just better than others. If you have someone like me at the mic you could try a large diaphram pop filter and EQinf out the "S" sounds but someone good who does not need that and can control his dynamics will sound better.

    I was just talking with someone a few weeks ago who for the first time hired a pro voice actor. They were amazed at the difference. The guy came in and had already studied the script at home. He read the lines in one or two takes and there was not much they needed to do to the recording. Before, they had been working very had with many re-takes, EQing and filtering, spicing together the "good" parts and so on. Taking 4X longer to get a product that did not sound as good. They did not pay a lot either.
  4. macrumors 603

    Let me start by saying "Thanks!" for the detailed, thoughtful posts!

    Can you elaborate? I like the value but wondering what your experience is.

    So just to be sure I understand, you guys are suggesting I record at levels that will ensure I don't clip and then... leave it? or increase the master mix level when finalizing the project?

    A lot of what I read indicates that the most common effect is to use a compressor to reduce the dynamic range so you can then boost the overall level to make the whole track/mix sound louder and clearer.

    I don't think anyone wants to have to max their volume control and then strain to hear my mix.

    Also, common sense (?) would tell me that it's best to record at the highest level possible (that won't cause clipping) to maximize SNR. If I record at a soft level, then when I later boost the levels, the noise floor is going to come up too.

    Can you guys either elaborate on what you were saying or correct my misunderstandings?

  5. macrumors 601


    i have a pair of nt1a's and rarely use them -- they're sizzly. the 2020 can be like that too, but i've just always liked the result better. i don't have one but have used them on a couple of occasions.

    digital has no self-noise, and the self-noise of the mics and preamps, properly gainstaged, will be very low. so you don't really need to worry about SNR.

    what you do need to worry about, besides digital clipping (very very bad), is running the a/d converter out of its sweet spot. they tend to sound better if you don't push for "full bits". you also need to worry about interstitial peaks -- basically those on-screen meters don't always tell you the truth. figure that your signal is easily 15 dB higher than what it's showing you, if not more.

    if i shoot for -18 dB, i still get actual peaks above zero. smassey (plug in maker) has a protools meter plug that shows these peaks. after i started using it, i turned my stuff WAY down.

    regarding compressors, i *like* some dynamic range. as long as your performance is even, i don't think you'll need one. and before reaching for a compressor, i always automate volume on the track. doing so means any compressor i do use won't have to work as hard.
  6. macrumors G4

    I don't disagree with your technique. But "digital" does in fact have sources of noise. One of them is quantization noise. If you remember the days when all we had was 8-bit samples it was horrible. "Telephone quality" at best. The other source has to do with numerical techniques used in the software. But today these are all small

    I think what you are saying is that with 24 bit A/Ds we have so many bits that we can reserve four of them for peaks you can't see on a meter and stuff like that.
  7. macrumors 601


    agreed. but as a noisefloor, it's so much lower than analog that people don't need to worry about "using all the bits" to avoid it.

    to the OP: if your levels are barely registering, then you're too low! keep it around the middle and you'll be fine.
  8. macrumors 603

    Thanks guys! I'll do some recording this weekend and play around with it some more.

Share This Page