SRC / Dithering

Posted:
Thu Mar 24, 2011 2:31 pm
by zenelectro
Hi Guys
Just a general question to all the mastering / recording gurus:
- For 24bit 88.2/96/176.4/192 k downconverting to red book, what SRC / Dithering are you guys using?
Thanks,
Terry
T & L Technical P/L
Electronics design / service / warranty repairs
Re: SRC / Dithering

Posted:
Sat Mar 26, 2011 11:48 am
by rick
I have a weiss src but i never use it i generally turn everything into analog at some point (with the weiss/maya super convertor)
but from the 24 files to red book
I still use apogee uv
although i have every option under the sun - i dont think the brand and style of dither is a big issue
Re: SRC / Dithering

Posted:
Sat Mar 26, 2011 2:24 pm
by Suburban Studios
Hello all,
You may want to check out this very informative link.
http://src.infinitewave.ca/It's a comprehensive database comparing different SRC algorithms.
Have a look at the FAQ and HELP sections for greater detail regarding the graphs/test results and how to interpret them.
There seems to be a great discrepancy between the different algorithms, including software based (built into DAW or plugin based) and hardware units.
Not sure how audible the differences are, but at least you have some objective measurements to base a decision on.
Cheers,
George
Re: SRC / Dithering

Posted:
Sat Jun 04, 2011 3:36 pm
by chris p
I was on the WWW yesterday and saw that CraneSong have released a "dithering" CD! Made me look more into dithering and what is is/does.
The short answer is that dithering is the application of quiet "random" (in fact, in most algorithms it is shaped to some degree) noise to a digital signal prior to word length reduction, to avoid quantisation artefacts. More sonically speaking, it reduces signal distortion at the cost of raising the noise floor, the underlying premise being that the human ear/mind is much more tolerant of noise than or distortion, so its a good trade. It words for ANY digitised signal - look for dithering under photography and you'll be able to see what dithering does very clearly - and in fact the principle was derived from WWII bomb sight mechanisms which apparently worked better in the air than they ever did on the ground - because the vibration of the airframe introduced "dither".
With the CraneSong CD, you apparently at the track at a volume equal to your wordlength reduction (ie 3db/bit), so 24 bit to 16 bit means you add this track at -24db compared to the source material and whammo - instant "analog niceness".
So far so good. Three points further for discussion -
(1) It was suggested in some forum discussions that it is better to do all ITB DAW mixing and bouncing to a final, high-wordlength saved track and then do wordlength truncation and dithering as a last, separate step - PT in particular was singled out as being less good at mixing, bouncing, dithering and WDR all in the one step. I've never noticed this as a problem but I've never tried the alternative. Does anyone have an experience with this, or is it just another forum myth.
(2) Someone said their used their own personal dither track based on a recording of wind through a nearby forest, with associated wildlife and bird noises. Claimed that there were good sublinals when using this to dither (hmm - but they are all supposed to be trruncated out, aren't they?) Anyone tried a personal dither track? Sounds like a great studio selling point (do you want wind in the forest, gentle rain or waves in a cave dithering?)
(3) When you think about it, AD conversion is also a wordlength reduction step, so dithering is as important at recording as it is when doing final mixdown. I record at 24 bit, and nothing in my recording arsenal comes close to a signal to noise ratio of more than 140db, so my understanding is that the mic and preamp noise floor provides its own dither in this case. Any digital recording experts care to comment?
And please avoid the obvious puns - no, I'm not all in a dither over this, nor am I on "shaky" ground. I will happily select POW-r in Logic Pro and truncate to my heart's content in the firm belief that my aged ears will never pick the difference (we are talking effects that take place in the -89 to -92db range). However, the human ear and mind never ceases to astound me with its capabilities, so I am open to being amazed.