Meditation Can Help You Escape Suffering, With Dr. Mark Epstein | Full Podcast Interview: Dan Harris

[@lindaelarde2692]

I'd love to hear you interview Lisa Feldman Barrett and Anil Seth...their work lends a fascinating dimension to the ideas you're discussing.

[@markfuller]

The audio is voltaile (low, then jarringly high). Look into loudness normalization, EBU R128, the "loudnorm" filter in ffmpeg.

[@TenPercentHappier]

Noted. Thanks for the feedback!

[@markfuller]

@@TenPercentHappier ffmpeg is command-line & can be painful. The way I use the "loudnorm" filter (which implements EBU R128): *Step 1 (analyze):* ffmpeg -i 'input-filename' \ -vn -sn -dn -af loudnorm=I=-16:TP=-1.5:LRA=11.0:print_format=json:linear=false \ -f null /dev/null 2>&1 The I, TP & LRA values are constant (I found something a long time ago saying those were the values recommended for online content. There's some other values for broadcast tv/radio). I always use what you see above. FWIW: I = "integrated loudness target" (range: -70 to -5.). TP = "true peak" (range -9 to +0). LRA = "loudness range" (1.0 to 50.0). Defaults are -24, -2, 7, respectively. That command will output to the screen measured values. The vn, sn, dn bypass processing video, subtitle and data streams (since it's just analyzing audio). Runs faster. *Step 2 (normalize):* ffmpeg -i 'input-filename' \ -af loudnorm=I=-16:TP=-1.5:LRA=11.0:print_format=json:linear=false\ :measured_I=-nn:measured_TP=-nn:measured_LRA=nn:measured_thresh=-nn:offset=nn \ -acodec aac -b:a 160k -ar 48k -vcodec copy -c:s copy \ *output-filename* Replace the "measured_" values (and "offset") with what was displayed at the end of step 1. Note: offset can be negative or positive. The ar parm is necessary because the loudnorm filter up-samples the rate to 192khz. You have to tell ffmpeg to resample the rate back to 48k. The "copy" parms cause video and subtitle streams to be copied without processing (you're only processing audio). Note that those command-lines work with Linux. Windows might be different. Ffmpeg can be frustrating. Basically the -af parm is passing everything to the loudnorm filter (until a space is encountered). It works very well if you can get it to work. The only thing that ever changes are the measured & offset values in step 2. You can tell it to output a different bitrate (instead of 160k. Probably best to use same rate as the input file's audio bitrate. There are tools to show what that is.). You could output to something other than aac , i.e., if the input is different and you want to keep that format.

[@mindfuleats4517]

Loved this. Thanks

[@TenPercentHappier]

Glad you enjoyed it!

[@charlesmay3759]

Very low volume

[@Meta4Monky]

Bros, you should mark yourelves as what people are looking for if they thought they liked sam harris 😂