Printable View
Hope the experts like Mike Phillips and Kirk Harnack can give me some advice on a problem I am having. All my recordings are a little low. You need to turn your laptop speakers up all the way to hear me and its still a little low. If i go any higher I get distorted. I am using a simple Mackie mixer. I have tried everything I can think of but its still a problem.
If you are willing to post a 1 minute segment, I would be happy to take a look at it. It's difficult to make suggestions without a clear indication of how it sounds at the source level ...Quote:
Originally Posted by TommySulivan
-paul.
@produceNewMedia
What Paul said.
I think this question warrants an answer, even with TommySullvan's lack of response.
The single most important aspect of establishing a distortion free workflow is Headroom within a good quality recording. Remember that there is a difference between the input (recording) level and the distribution (perceived) Loudness.
First of all - proper source management is imperative. I'm referring to the quality of the hardware, mainly the source of gain (preamp) and the mic - as well as implementation. For example if you are using a low output mic, cranking up your preamp in order to drive the mic is not recommended. This will result in problematic noise that will eventually need to be dealt with. Without the proper tools in post - aggressive noise reduction never turns out well.
In terms of gain on the input, I'll share a tip - basically how I handle dealing with a low output mic. I have my mic plugged into a voice processor with a decent preamp. I'm routing the output of the voice processor to a line input on a Mackie mixer with a gain trim. The mixer then feeds my DAW. I insert a gain trim plugin on the channel that's receiving the mic input.
Basically I've set up what can be referred to as gain staging. I use a combination of three stages to sufficiently drive the mic with low noise. The preamp on the voice processor is the initial stage of gain, and it is set no where near it's maximum. I then use the gain trim on the mixer's input channel to pick up additional gain while still maintaining a minimum amount of noise. Lastly, the software gain trim plugin is fine tuned until the input signal is hot enough to record with ample Headroom.
Regarding Headroom ... we'll classify this as the difference between the Peak Ceiling (peak amplitude) of the audio coming in relative to 0 dBFS. For example if your maximum peak on the input is -12 dB, that's 12 dB of Headroom to work with in post. How much Headroom should you shoot for? Well that's subjective. There are many variables including the skill level of the producer, the gear, as well as the availability of proper tools.
If you are wondering what my subjective recommendation would be ... for now I'll say that anything close to -6 dB or so at the input is cutting it close. I would be more comfortable suggesting that you leave yourself 8-12 dB of Headroom to work with.
Note that at this point if there is distortion present in the signal, you're in trouble. It's one thing if your post production processing results in distortion. This can probably be corrected. However if there are problems at the source level, well then you need to revisit your recording techniques and figure out how to alleviate the problems BEFORE you take the recorded audio into post.
Moving forward - let's assume you were able to record a fairly clean piece of audio with sufficient Headroom. Again - this is a source level recording and 99% of the time it will be unsuitable for distribution, referring specifically to it's Integrated (Program/Perceived) Loudness.
First thing you need to do is assess the dynamics and decide whether Dynamic Range Compression is necessary. For spoken word you want to maintain a much tighter dynamic range than you would for music. I'm not suggesting overly aggressive compression. The point is if the recorded source is sort of "choppy" with levels all over the place, it's going to be difficult to properly bump things up in post.
And you guessed it - you use a Compressor to get this done. Start off with a 3:1 or 4:1 Ratio, fairly fast Attack time and moderate Release. The Threshold will vary based on how hot the signal is in the first place. You will need to experiment with the combination of all settings in order to achieve the proper results, where the transients (peaks) are tamed without overly compressing.
After this initial pass of Dynamics Processing the audio will be much more suitable for bumping up to the recommended distribution targets. And this is where things can get very complex. It will depend on (as I said) skills and availability of tools.
Note that I didn't make reference to additional aspects of post production (equalization, noise reduction, etc.) that will also play a major role in the end result. In terms of targeting distribution Loudness, problems that are perceptually audible in the source will be much more noticeable after bumping things up post.
Bumping Things Up
One consideration would be to use bundled processing options in applications. For example the Match Volume options in Adobe Audition. Auphonic (the web service or stand-alone application) is also worth considering. For advanced producers who are comfortable using various plugins for "manual" processing, I have plenty of recommendations.
Here's an exercise. You'll need a Compressor plugin and this free, cross platform plugin: (consider donating to the developer :) I have no affiliation):
LoudMax
I'm going to use a few screen shots from one of my own blog posts. Note that the source example was actually a distributed Podcast. In essence it is not a source level recording. However I think it will serve the purpose of displaying various stages of processing that will result in a distortion free piece of audio suitable for distribution.
This is the recorded source. In order to make it suitable for Loudness processing, we need to tame the Dynamics.
http://f-video.s3.amazonaws.com/headroom/source.jpg
This is the result after the initial stage of Compression:
http://f-video.s3.amazonaws.com/headroom/initial.jpg
At this point additional processing has been applied in preparation for final Loudness processing. Let's refer to this as the Processed Source.
http://f-video.s3.amazonaws.com/head...termediate.jpg
Lastly, the Processed Distribution Master:
http://f-video.s3.amazonaws.com/headroom/final.jpg
Note that I did not refer to shooting for any specific targets. I just want you to see the process as it evolves.
Try to replicate the various stages using one of your (clean) source recordings. As noted you will need to manage Dynamics and what ever else you deem necessary. When you reach the Processed Source stage - using the LoudMax plugin, set the bottom Output slider to -1.5 dB. This is the Peak Ceiling. Play the audio clip through the plugin and pull down the Threshold slider. You will notice an increase in Perceived Loudness. Also be aware of the amount of Gain Reduction (vertical meter).
http://f-video.s3.amazonaws.com/headroom/lm.jpg
Since we are not really shooting for any specific target, use your ears and get things to where you feel comfortable in terms of Loudness. Finally, bounce your audio clip through the plugin.
Remember this is sort of an arbitrary exercise . If your recording was in good shape, and you managed things well throughout the workflow - the resulting audio should be significantly louder with no hint of distortion.
Lot's of subjective stuff here guys. But that's my take. Again - there are tons of variables.
Hopefully you'll find these general guidelines helpful.
-paul.
@produceNewMedia
Very helpful post ptfigg!
Spencer,Quote:
Originally Posted by Spencer Kobren
I'm glad you found it helpful ;)
-paul.
@produceNewMedia
