When mixing down to a stereo file, do you use the Pow-r 3 dithering?  I was told recently there are some harmonic errors created with this algorithm, and to consider using Pow-r 2.  So now when I don't like what I hear in my stereo file, I don't know if it's the quality of my work (likely) or the dithering (less likely).

I'm not able to hear differences between the algorithms. I use Pow-r 3. But I know I have a slight hearing deficiency.
I strongly doubt if the different ditherings can cause differences so big that they would make the resulting wav sound "not likable".

MorganT


When mixing down to a stereo file, do you use the Pow-r 3 dithering?  I was told recently there are some harmonic errors created with this algorithm, and to consider using Pow-r 2.  So now when I don't like what I hear in my stereo file, I don't know if it's the quality of my work (likely) or the dithering (less likely).

Hi,where did you hear about the algorithm errors?


If I hear weirdness,I always blame my vibrato,or my new obsession to get "everything louder,than everything else".
Bob

FWIW, I've done some (exaggerated) listening tests... and actually prefer Pow-r algorithm 1.
Algorithms 2 and 3 produce results that "irritate" my ears (for lack of a better description).

Take an isolated snare hit... and run that thru a reverb with a long decay.
Take that long reverb decay (into the noise-floor)... and use it to audition various dither algorithms.
Use headphones with the amp cranked up so you can really hear what's going on.
Obviously you want to be careful doing this... 

I like POW-r3 for Rock and Metal genres
I like POW-r2 for Rap, hip hop, Electronic, Trance, Reggae and Country genrea
I like POW r-1 for Classical and Easy listening music genres

Cj

Take that long reverb decay (into the noise-floor)... and use it to audition various dither algorithms.

Jim Roseberry


FWIW, I've done some (exaggerated) listening tests... and actually prefer Pow-r algorithm 1.
Algorithms 2 and 3 produce results that "irritate" my ears (for lack of a better description).

Take an isolated snare hit... and run that thru a reverb with a long decay.
Take that long reverb decay (into the noise-floor)... and use it to audition various dither algorithms.
Use headphones with the amp cranked up so you can really hear what's going on.
Obviously you want to be careful doing this... 

I like POW-r3 for Rock and Metal genres  I like POW-r2 for Rap, hip hop, Electronic, Trance, Reggae and Country genrea  I like POW r-1 for Classical and Easy listening music genres  Cj 




This is highly intriguing,as I have been trying to get as much info on the dithering process as possible,and although at times I think I'm kidding myself,I do not like the sound of recording everything at 16/44.1.
Bob

Not surprised - you should be recording @ 24 bit and only dithering down to 16 bit as the very last step.

If you're recording @ 16 bit and dithering down afterwards, you are seriously degrading your signal.

Dithering more than once does raise your noise floor unnecessarily, so you want to avoid that. But which algorithm you choose is pretty much irrelevant.

Bobguitkillerleft -- In Scott Garrigus' guide he states Pow-r 3 is the "most accurate" processing, but to use Pow-r 2 or 1 if it "overtaxes" your PC.  I felt I was hearing some harmonic distortion in a vocal when exported to stereo track that I was not hearing while monitoring the mix.  Am certain it was not an overload on processing power, so I tried triangular dithering to go to the other extreme, and it seemed to disappear.  I asked a friend about it, he said he recalled reading "somewhere" that this algorithm can induce some mild harmonic distortions at times, but wasn't sure.  I'm pretty sure now it was some of my effects / loudness...

Interesting that you all are noticing different impact of dithering on different styles of music.  I'll give that a whirl.

I generally record at 48k/24 bit.  Correct me if I'm wrong, but if you record at 44.1/16 bit you don't need dithering, as that is used to adapt bit size to 16 bit?

OK, so that last comment got me thinking.  This is an older project that I pulled out to "perfect."  (Bad habit, I know.)  I recorded it at 44.1/16 bit.  So if I mixed it down to stereo using dithering, wouldn't that be inducing further distortion?  (ie, shouldn't dithering be set to "none" in this case?)