Resampling under OS X

[Masis]

I am wondering if there is a easy and friendly to use software for converting 24/176.4 WAV into 24/96 FLAC?

 

Thanks.

 

[Audio_ELF]

Assuming you mean non-on-the-fly resampling (which Mac OS X does reasonably well anyway)...

 

Okay so it's not free ... but how about Wave Editor. Around $80 and comes with the iZotope resampling software which does a good job.

 

The open source alternative is called Sox. Not tried it myself but I understand it's a command line (rather than GUI) software.

 

Eloise

 

P.S. I've just found a Sample Rate Conversion software comparison website which may be of interest to some.

 

 

[Masis]

Thanks Eloise,

I have read about SOX. I was hoping for something a little easier to use than a command line utility. I downloaded Wave Editor as a trial, but when I try to export the 24/176.4 file to FLAC, I am only given the option to change bit resolution, but not the sampling frequency. Do you know how?

 

ws

 

[Audio_ELF]

Sorry, not sure how Wave Editor works in practical terms ... I know others on here use it so maybe one of them will be able to help. You have selected the iZotope as the resampler haven't you?

 

Eloise

 

[Masis]

Hey thanks for the quick reply.

I figured it out. But now I see a lot of options for the Izotope filter for different slopes... very confusing...

 

[bdiament]

Hi wslam,

 

When you have Wave Editor open and set to use iZotope's 64-bit SRC, you will get excellent results if you simply move the "Quality" slider all the way to the right (best quality) and let the algorithm "choose" the rest of the settings.

 

Have fun.

 

Best regards,

Barry

www.soundkeeperrecordings.com

www.barrydiamentaudio.com

 

 

[The Computer Audiophile]

The original post mentions going from 24/176.4 to 24/96. Due to the math involved you may want to consider going to 24/88.2 as it's a multiple of 176.4. Of course your DAC and OS will need to support 88.2 which isn't always the case unfortunately.

 

[cfmsp]

 

 

Agreed on letting Wave Editor determine the settings, AND on sticking with even multiples (or, in this case, fractions) of the (presumably 44.1kHz) sampling rate when re-sampling.

 

clay

 

 

[bdiament]

Hi Chris,

 

In my experience, integer conversion is no longer necessary, provided one uses the better SRC algorithms. In my experience, iZotope's is by far the most transparent (meaning the results sound most like the unconverted original).

 

I would submit that iZotope's SRC can do 176.4 to 96k considerably better than most other algorithms can do 176.4 to 88.2.

 

Just my perspective of course. I've compared (and beta tested) many an SRC (and dither/noise shaping) algorithm. The criterion I always use is the unconverted original. Many algorithms add a "flavor"; with SRC, I find it in the spurious harmonics they generate in the resulting file. Most, to my ears, lend a certain brightening (and hardening) to the sound, that some folks will undoubtedly like. They might even create a bit less of this when used for integer conversion. The iZotope SRC seems to be devoid of this brightening and hardening. Its internal math seems sophisticated enough that it does not "care" whether the conversion is integer or not.

 

As always, I suggest folks try it themselves and draw their own conclusions. If the DAC can perform equally at 88.2 and 96, the test will be fair.

 

Best regards,

Barry

www.soundkeeperrecordings.com

www.barrydiamentaudio.com

 

 

[The Computer Audiophile]

Hi Barry - Thanks for the first hand experience. It's always great to bypass the speculation by armchair engineers like myself :~)

 

I do know some engineers who prefer to steer clear of any digital SRC and use one DAC and one ADC to do the conversion.

 

If it sounds good than it's good.

 

[Masis]

Thanks! I have just converted 10 HRx to 24/96 using Wave Editor. I am still new to the program and wonder if there is away to batch the action! Otherwise, this is gonna be quite a nightmare...

 

I will try 176.2/6 to 88.2 vs 96 and see if I can hear a difference.

 

For fun, I also converted from 24/176.4 to 16/44.1.

 

The difference is not subtle!

 

[bdiament]

Hi wslam,

 

Audiofile Engineering, makes of Wave Editor, have another app called Sample Manager that will do this.

 

Sample Manager also uses iZotope's SRC (and dither), though there are fewer options. Just set "Quality" to best and the results will be quite wonderful.

 

Best regards,

Barry

www.soundkeeperrecordings.com

www.barrydiamentaudio.com

 

 

[Masis]

Thank you so much Barry... downloading.

 

[bdiament]

Hi Chris,

 

Some engineers preferring the additional conversion steps only supports one of my oft repeated sayings:

 

Ask five engineers a question and you'll get at least six answers.

And seven of them might be wrong. ;-}

 

Best regards,

Barry

www.soundkeeperrecordings.com

www.barrydiamentaudio.com

 

 

[Masis]

Hmm the Sample Manager cannot open up FLAC files. Only WAV. I have some high res FLAC files too which I also want to downsample...

 

[Andrew S.]

Great thread - thanks.

A little OT: do you recommend upsampling all 16/44 files to 24/174 or similiar?: I ask because I have a friend with a dCS stack and the difference between using their upsampler to 24/74-192 (a Purcell I think) and not (16/44.1) is significant going into an Elgar.

I realise it will all depend on the dac but for most pro derived converters eg dCS, Weiss, Prism, Lavry etc what are your experiences?

Thanks to the OP for raising this.

Chers

Andrew

 

[Andrew S.]

Great thread - thanks.

A little OT: do you recommend upsampling all 16/44 files to 24/174 or similiar?: I ask because I have a friend with a dCS stack and the difference between using their upsampler to 24/74-192 (a Purcell I think) and not (16/44.1) is significant going into an Elgar.

I realise it will all depend on the dac but for most pro derived converters eg dCS, Weiss, Prism, Lavry etc what are your experiences?

Thanks to the OP for raising this.

Chers

Andrew

 

[bdiament]

Hi Andrew,

 

I can think of only one reason to change the sampling rate of finished (i.e. ready to listen to) files to a higher one. That is, a good DAC is going to use gentler filtering at the higher rates and this may offer sonic benefits.

 

That said, depending on the DAC, there may also be a sonic price to pay in the form of less stable clocking at the higher rates. Some DACs in my experience, perform better at lower rates than they do at the highest rates. (This includes some "pro" converters.)

 

Also, when I speak of converting the sample rate, I'm referring to using an off-line (i.e. NOT real-time, while the music plays) algorithm of high quality, like iZotope's "64-bit SRC". With many other algorithms I've heard, even at integer rates, the brightening and hardening engendered by the lesser algorithms can easily offset any benefits of gentler filtering, regardless of the DAC.

 

So, for a few specific combinations of SRC algorithm and DAC, I think there may be sonic benefits. For most others, my view is the net result will be a sonic loss.

 

Just my perspective. As always, I suggest auditioning variations to see how you feel about it. You may agree or you may feel differently.

 

Best regards,

Barry

www.soundkeeperrecordings.com

www.barrydiamentaudio.com

 

 

[Andrew S.]

Thanks Barry - very helpful. I guess the dCS is "offline" in that it doesn't do it in the dac: I iimagine their math is pretty good, sounds great with their dacs.

 

Regards

Andrew

 

 

[bdiament]

Hi Andrew,

 

I'm not so sure it is off-line.

Doesn't it perform the upsampling while the disk is playing?

If so, I would call it "real-time", not off-line.

 

Best regards,

Barry

www.soundkeeperrecordings.com

www.barrydiamentaudio.com

 

 

[cfmsp]

 

 

"the brightening and hardening engendered by the lesser algorithms can easily offset any benefits of gentler filtering, regardless of the DAC."

 

Barry,

 

It's my considered opinion that 'brightening and hardening' is often mistaken (and therefore seen as a good thing) for 'transparency' or being 'more revealing' by audiophiles, and especially, ironically, when paired with gear that might NOT be exceptionally revealing.

 

Audiophile reviewers might even call a product such as this 'ruthlessly transparent', and 'utterly revealing', and even have the audacity to caution readers (this is my favorite, can you tell?) to be very careful with the use of 'lesser' equipment upstream in the playback chain (so as not to exacerbate the fault - aka brightening and hardening).

 

clay

 

 

 

[bdiament]

Hi clay,

 

I think this depends completely on the individual audiophile. Personally, I know many who would hear the brightening/hardening and not be fooled, especially those who have taken the time to assemble a system of relatively neutral components. Others will hear the difference even if they burn the files to CD-R and listen in their cars (with the windows open, on the highway).

 

An astute listener will not combine an brightened/hardened component (or recording) with others that roll that brightness and hear (or expect) a zero sum, simply because it doesn't work that way. That only works for less than astute listeners, which granted, will include many audio writers and very many record producers. =8-0

 

All that said, clearly some folks *like* a brightened/hardened sound. For them, iZotope's SRC would not be the best choice as (to my ears at least) it doesn't do any of that. Happily for them, they can use any of several other SRC algorithms and get their desired results. They'd probably be even happier with real-time SRC. ;-}

 

Just my perspective.

 

Best regards,

Barry

www.soundkeeperrecordings.com

www.barrydiamentaudio.com

 

 

[Bob Stern]

.

 

[Bob Stern]

To see detailed frequency response curves produced by different combinations of the Cutoff and Steepness parameters, download the free demo of Izotope RX Advanced. Although the graph only extends to -35 dB, you can move your cursor around the graph and the blue number at the top shows the attenuation at the frequency where your cursor is pointed.

 

The default settings all have Cutoff Shift = .96 or 1. To produce the "apodizing" filter used in the Meridian DAC, it would seem advisable to choose a much lower value for Cutoff Shift, perhaps even as low as .75.

 

The presets for Steepness jump from 4 to 32. Experimenting with intermediate values of Steepness seems a good idea. Lower Steepness produces less ringing, but 4 is too low for "apodizing", i.e., too low to attenuate possible ringing from the recording engineer's anti-aliasing filter.

 

One plausible combination (based on a few seconds of fiddling with the graph, not listening) is Cutoff = .78 and Steepness = 25, whose response is -1 dB at 16.5 KHz, -5.7 dB at 17.2 KHz, and -181 dB at 19.845 KHz. Another example with earlier treble rolloff is Cutoff = .75 and Steepness = 20, which is -1 dB at 15.6 KHz, -6.1 dB at 16.5 KHz, and -183 dB at 19.8 KHz.

 

[Alexey Lukin]

>> Although the graph only extends to -35 dB, you can move your cursor around the graph and the blue number at the top shows the attenuation at the frequency where your cursor is pointed.

 

You can also scroll the scales by dragging or mouse-wheeling them.