Is it just me, or does Sox deemph overdo the de-emphasis?

[CastleJohn]

Thanks for all your help. I won't worry any more about it.

 

[Mario_JR]

So, SOX is good a good solution to de-emphasis?
I have a Jap Michael Jackson CD here that uses pre-emphasis and I would like to rip it =)

 

[Wombat]

Here's a thought: deemph is a low-pass shelving filter, exactly equivalent to 'treble -9.477 5283 0.4845s' (you can use the --plot option to check that the filter coefs are identical), so you could use 'treble -5 5283 0.4845s' for a 'halfway house' solution (or whatever peak attenuation value sounds best in this case).

Nice and simple idea to use the sox treble setting. I once tried several pre-emphsis samples and ended with this reasoning: YMMV
CD red book de-emphasis? - Page 4
So maybe someone wants to try these values with sox treble -9.31 5135 0.48
Would be even easier as to use the VST plugin and you should not have to worry about the VST Host dither and other probs.

 

[CastleJohn]

What is the effect of Sox de-emphasis (or treble adjustment, or for that matter, level normalisation) on pre-existing dither? As Sox will re-dither when it truncates the data to 16-bit, is the low-level information that was encoded in the original dither fully retained?

 

[phofman]

What is the effect of Sox de-emphasis (or treble adjustment, or for that matter, level normalisation) on pre-existing dither? As Sox will re-dither when it truncates the data to 16-bit, is the low-level information that was encoded in the original dither fully retained?

The low-level information is encoded by the higher-frequency components introduced by the noise-shaping dither. IMO it will be lost when these ultrasonic components are eliminated by subsequent processing.

 

[Wombat]

If someone thinks to dither the output does degrade the resulting file in any way you can of course tell sox to put it out as 24bit file.
Since nearly all recording won´t use the full 16bit even after de-emphasis i don´t do but people that are more concerned about it may try this.

 

[phofman]

If someone thinks to dither the output does degrade the resulting file in any way you can of course tell sox to put it out as 24bit file.

Or use the -D (--no-dither) parameter to tell sox not to add the dither effect at the end of the chain automatically.

 

[Mario_JR]

Or use the -D (--no-dither) parameter to tell sox not to add the dither effect at the end of the chain automatically.

Gonna test this option.
Ive made some ABX tests in foobar using Waveemph and SOX (only using deemph param) and to my ears the Waveemph file sounds much better, sox file sounds kinda congested.

 

[highpass]

SoX deemphasis is matched to the spec. and is slightly more accurate than wavemph, which was matched to a Denon test CD, which was slightly off-spec. However the difference is very small so I'd be surprised if it is audible.

 

[Wombat]

WaveEmph doesn´t dither at all and uses linear filters afaik. The biquad sox uses for this HighShelf filter has the same phase behaviour as the analog de-emphasis unit. We can discuss if the phase does anything here but anything that WaveEmph does it does most likely worse as sox or the EQ i linked. If you hear it to be "better" i´d be surprised.