Hi. I’m working on a new podcast. While it sounds good when I increase the loundess to -18 LUFS, I start to get distortion and considerable red-lining when I increase it to -16 LUFS. Does it need to be -16 LUFS for podcasts, and if so, do you have any advice for how to accomplish this, when going from -18 LUFS to -16 LUFS causes it to sound harsh? Thanks so very much. --Paul
Hi Paul, could you post a small sample of your audio at -18 LUFS for us to have a look and listen to?
IanT.
1 Like
Thanks, Ian. Here it is–everything is still in draft form. I’m still getting input from young men about the image for the podcast as well as everything else: Episode could not be found
Hi Paul, sorry for the delay in getting back to you. Thank you for the link and posting an example to ‘examine’. This is what I have found based on the fact that this is, I’m thinking, an encoded mp3 file and not the original recording…
On my initial listening (laptop speakers), I thought it was at a good volume for this kind of production. So, I then put your audio through some tests in my studio to check its compliance with the most requested specification that I have needed to work to in the past - Apples podcast audio spec. Last time I checked it was…
“ -16 dB LKFS, with a +/- 1 dB tolerance, and that the true-peak value doesn’t exceed -1 dB FS. The LKFS and true-peak values are calculated according to the ITU-R BS.1770-4 recommendation. “
There is more to the spec but we only need this bit for now to use as a target for measuring basic compliance.
I measured the audio sample over a two minute time period, and found it was already over the above spec in two areas. Bear in mind that when an AIFF or BWF (Wave) is converted into an mp3, certain values change in the process, so getting a completely accurate picture here isn’t possible.
Here is a screen shot of your audios waveform statistics using RX…
Notice the ‘True Peak Level’ at the top of the meter gives a value of -0.56 dB, which has exceeded or is ‘over’ Apples spec. Also look at the ‘Integrated Loudness’ (second from bottom window) with a value of -15.3 LUFS, this is also over.
Now because I found that both your ‘True Peak’ and ‘Integrated loudness’ are showing as overs, I just adjusted the overall gain of the audio sample downwards to get the integrated to the correct LUFS value. That in turn also dropped the true peak value into the spec as well.
Another pic with the new waveform stats of your corrected LUFS/True Peak values now falling within the compliance specification…
These results are based on a two minute sample, really I would/should measure the entire show to get an accurate picture of compliance and tweak any minor adjustments/fixes that need doing afterwards.
I would recommend that you try to get you LUFS and True Peak values on the target required by the hosting company you will be using. Also it means that your podcast will be at competitive sound level to any others casting on this or similar subject matter.
On to your harshness problem…
When increasing the gain or volume of an audio file, it is the peaks and true peaks which cause the harshness you are experiencing initially. Basically once the true peaks go above 0.00 LUFS the waveform will start to be squared off. This is a form of distortion and is referred to as 'Clipping’. In this case it should be avoided unless you like that crunchy sound on your voice. The more you push the audio signal beyond 0.00, the greater the amount of clipping/distortion.
To solve this a Compressor is required. You carefully set the compressor (using the various parameter controls) to in a sense, ‘squash’ or ‘compress’ the peaks/true peaks of your audio signal. Once set correctly, you will be able lift your LUFS level without clipping occurring or your channel meters showing red flags. You can find plenty of YouTube videos on how to set them up and what to listen for whilst searching for the ideal settings.
IanT.
1 Like
Ian,
I can’t believe your incredible thoughtfulness and kindness, and I can’t thank you enough. You are wonderful beyond words.
Best,
Paul