Hello Folks:
I have quite a few high resolution flacs, in 24bit/96Khz, (ie HDTracks purchases), and occasionally have need to convert these to 16bit/44.1kHz for CD compilations or portable devices that can't handle 24bit:
(ie AirTunes/Airport/AppleTV2).
Being Mac based, I use xACT, XLD, Max, and Adobe Audition to capture/encode/manipulate my audio library.
At this time, I am using Adobe Audition to convert the 24bit/96kHz source files to 16bit/44.1 kHz, using the "Save As" function, and choosing the convert sample type choice to 44.1 kHz Stereo, 16-bit (Triangular Dither).
Is this a valid/correct choice using the tools I have available?
(PS: The output is WAV, and then I might do final conversion to ALAC via XLD for my iPhones/iPods).
I still treasure my original hi-rez tracks, but enjoy the process for making custom compilations, etc, and hope to improve my methodologies. )
Thank you for your feedback/consultation/corrections.
At this time, I am using Adobe Audition to convert the 24bit/96kHz source files to 16bit/44.1 kHz, using the "Save As" function, and choosing the convert sample type choice to 44.1 kHz Stereo, 16-bit (Triangular Dither).
Is this a valid/correct choice using the tools I have available?
I'd say yes. I recommend these settings:
Sample rate conversion slider set to 999,
Pre/Post filter check box not checked (doesn't make it sound better IMO, just increases the filter length), 1-bit triangular dither with
Noise Shaping B selected in the drop-down box (or if you have Audition CS5.5, some other shape).
Important: make sure in the Settings dialog (hotkey F4), the
Use Symmetric Dithering box is checked, even if you don't use dither.
Chris
Thank You. Quite helpful in confirming/modifying.
My research is giving me bits and pieces for Mac tools settings for this exact project,
as I want to create a perfect "preset" called : "Hi Rez to CD" that I could universally commit to...
I am using Adobe Audition from CS5.5 .
My settings:
Sample Rate: 44100 Hz
Advanced > Quality > 100%
[ ] Pre/Post Filter UNchecked (* thanks for that, never could figure out best.)
Channels: Same as Source (i love old mono reissues too)
Advanced > Left Mix 100% / Right Mix 100% (both greyed out)
Bit Depth: 16bits > Advanced:
Dithering > Enabled
Dither Type > Triangular
Noise Shaping > None **
** It's this last setting that has me a bit confused…
Should I use "Neutral (Light)" ? [there is no 'Noise Shaping B' type options]
Noise Shaping also had some additional questinable default parameters for Crossover: kHz (~16.74) and Strength: dB (~24), which sound wrong,
and adaptive mode submenu choices of: OFF ** or Dynamic or Constant.
I initially tested with Noise Shaping at None versus reasonable guesstimate settings, and could hear no difference, hence I set to None… but I could be wrong...
Thanks to all at this community, It has provided much valuable research on a myriad of issues, and I am trying to keep up!
This goes back to the early Syntrillium days. Cool Edit has always done very good sample rate conversions. Doctrine has been that no one can hear any difference by increasing the filter quality setting by the time it reaches 250. Setting it higher makes the run take longer, which may be on little consequence in light of the increased CPU power over the past 15 years, but it does not change what one will hear.
Of course, this being the place for ABX tests, everyone should know they can test this for themselves. I settled on
dither depth 0.5
p.d.f. shaped triangular
noise shaping C1
because objectively it adds measurably less noise than many other choices, a difference I could hear at very high amplification levels, but there probably are exceeding few circumstances, if any, where other reasonable combinations make an audible difference at real listening levels.
The 'best' approach is dependent on how the original material was dithered, to a degree. For tracks originally subjected only to unshaped or minimally-shaped TPDF, any kind of dither is sufficient. For content dithered with psychoacoustic dither, unshaped TPDF is probably the best method.
Generally it won't be clear to you what kind of dither has been applied to your source, so unshaped TPDF is what I'd recommend.
If the material was recorded at 24 bit, and kept at 24 bit, it is quite possible no dither was ever applied.
If you, or anyone else, can actually hear a difference, regardless of the type of dither used (after converting to 16 bit), that does not depend on some special circumstances, such as selecting a section between tracks and turning the volume up really high (much too high for listening to the music), it would be interesting to learn about it. Almost certainly, if you just convert to 16 bit with no dither, it will sound exactly the same as any dithering you can use -- except perhaps one that applies so much noise in the lower frequencies that you can hear the dither noise itself on low level passages.