Encode to MP3 at 320kbps CBR -> Attached file. Encode to MP3 at 320kbps CBR -> Silence.Ģ) 1kHz sine generated at -115dBr. on the audibility of the quantization error generated.ġ) 1kHz sine generated at -115dBr. I don't think much quantization noise is accounted for there IMO.'Closest match'?ĪFAIK MP3 divides the frequency spectrum into 512 bands (which are then converted to the frequency domain.) It allocates bits to each band based on e.g. To think that even 320k encodes will preserve information below 70dB.maybe in a very quiet passage luke a recerb trail but joint stereo artifacts and frame switching should create more artifacts in reverb trails than dither would help "avoid" to begin with IMO.Īnd for louder passages, again mp3 doesn't give you "your audio" but rather frames of the "closest match" of spectral data from mp3 frames that already exist. Can I live with that tho? Of course, cuz I don't want the bottom 8 bits of my 24 bit song literally hacked off which is what happens if u don't dither (on a final master).ĭither for mp3 output is pointless, mp3 "encodes" to frame data where analysis of each audio chunk to be turned into an mp3 frame assigns the "closest match" based on the dominant spectral & uncorrelated (noise) frequency data the encoder "sees" during analysis. I will say that when clicking between ON & BYPASS on a dither plug-in always makes the song feel a lil collapsed when it's ON. The common ones are: Sonnox Limiter's dither, Waves L2's Type I dither, Izotope Ozone 4's MBIT+ dither, Waves IDR dither, & POW-r dither. My mind "thinks" it can hear or FEEL a difference. That being said, I still audition & compare different dithers when printing a final master & choose the one that makes the music feel the best. I have tried it every way, and no dither going from 24bit wav to mp3 sounds the best to me.Reading this again by Steven Massey makes me feel a hell of a lot better about obsessing about the dither process during mastering or even when creating a reference CD for a client. You want a technical reason why it doesn't really matter? Read this: Massey Plugins Inc. At the time, we could pick out which songs were 16bit & which were 24. It was very slight but the same song sounded smaller & more closed in w/ the 16bit.
I tried this about 7 years ago back when I was a Pro Tools LE user (now on HD) & myself as well as a very experienced engineer at the time could hear a difference when we mixed 16 & 24bit sessions for a 7-song EP. The plug-ins will be processing in 24bit & sound more open. Mainly because in a 24bit session the plug-ins will sound better working in a 24bit environment. Kind of thinking it's irrelevant if I can't hear it and I could definitely use the extra processing power back.If you're talking about working in 16bit sessions as opposed to 24bit, I wouldn't do that. No one noticed, I can't tell a difference. My last record I did one song with 16bit and the rest in 24. In fact, I am considering going back to 16bit.
If it's a final mix, I'd do it, but I don't hear the lack of dither when bouncing to mp3 from 24bit. I never had anyone complain to me nor compliment me on whether I had bounced and dithered to 16 bit or just did a straight bounce to mp3 from 24bit.Īt some point in the past, I inadvertently stopped putting dither on my tracks when bouncing down and I never noticed anything wrong. If you think about that last part, it sounds silly. Because of that I can't recommend you to listen to it." However the real question is will the listener know any better as to what you did? I don't think you will hear any reviews that say, "The song was good, but it sounds like they just did straight 24bit to mp3 conversion without dithering to 16bit first. If you can answer yes to any of those I applaud you. Last can you hear the difference between and mp3 that has been dithered to 16bit and then converted to mp3 vs. Can you tell the difference between different brands of dither (mbit+ vs uv22hr vs etc). We all know the theory of dither but can you really hear it? Can you hear the difference in 16 bit truncation vs. I think here's a few questions you have to consider in this argument. If you bounce it out, dither and then convert to mp3, you can then say that you did the most optimized audio conversion for the "cleanest sound." I do think perhaps if anything it's more of a "piece of mind" thing. Honestly I really don't have an answer other than my simple logic. Seems logical thoughĭo I need to dither when converting to mp3? | Ame oneI seem a little late to this but I will stand by what I wrote. Found this but can't say if it's right or wrong.