Jump to content
IGNORED

Audio MIDI setup settings on Mac OS X


Recommended Posts

I've just got a DacMagic, and have been reading some of the threads here about the device, using iTunes for playback, etc. I've got a question regarding the Audio MIDI Setup settings. When I look at the available settings for the digital out (I'm connected to the DacMagic via Toslink), it shows 44.1 KHz as the sample rate, but allows me to change that to 48K or even 96K, though the music I'm playing right now is 44.1. Is there any reason why I would want to make such a change? As I understand, the DacMagic does upsampling to 192 KHz anyway. I'm mostly surprised that such an option exists in the Audio MIDI Setup settings...

 

Thanks.

 

I write about Macs, music, and more at Kirkville.

Author of Take Control of macOS Media Apps

Co-host of The Next Track podcast.

Link to comment

 

 

"Is there any reason why I would want to make such a change?"

 

My recommendation is to maintain the same sample rate as your source material unless you have very specific reasons for changing it, e.g. experimenting with upsampling being performed by CoreAudio.

 

Some have argued that it might not hurt much to set Audio Midi at the highest sample rate your DAC can support and relax - but frankly, I believe this is akin to the lazy man's guide to enlightenment. Doing so would avoid the 'fatal flaw' of iTunes - having to restart iTunes every time one varies the sample rate during playback - but at the cost of upsampling in real-time in each instance where the highest sample rate is NOT native resolution of the piece being played.

 

clay

 

Link to comment

I fully agree with Clay. Don't let your computer do the real-time sample rate conversion. You'll lose low level detail. It's really best to match the Audio MIDI Setting to the native sample rate of the file. If you wanted to be really lazy and choose one setting then choose the one that corresponds to the bulk of your music. Better though imo to use itunes to organize your library and the latest version of 'play' to play your library; that way the issue goes away.

 

(Btw, I got the tip about the latest version of 'play' from someone on this forum but can't find the post now. Anyway, whoever it was; can't thank you enough. It's the perfect solution for me.)

 

- John.

 

Link to comment

 

"Btw, I got the tip about the latest version of 'play' from someone on this forum but can't find the post now. Anyway, whoever it was; can't thank you enough. It's the perfect solution for me."

 

That would be CG, who apparently has left the building!

 

clay

 

 

 

 

Link to comment

Hmm, so if I want to listen to files at higher sample rates, I can't do so without making a manual change to settings...

 

As for Play, it doesn't play higher sample rates correctly for me - or at least I would still need to change the sample rate in the Audio MIDI Setup. I've got some Linn files in high sample rates, and when I play them (converted from FLAC to Apple Lossless) in iTunes, my DacMagic recognizes the higher sample rates. When I play them through Play, it shows everything as being 44.1 kHz, unless I change the settings in Audio MIDI. So Play isn't doing anything for you with files like that.

 

I write about Macs, music, and more at Kirkville.

Author of Take Control of macOS Media Apps

Co-host of The Next Track podcast.

Link to comment

I know that its generally accepted here to play the sample rate the same as the source material. Makes sense and this is what I do.

Having said that, I came across this yesterday:

http://www.usbdacs.com/Macintosh/Macintosh.html

 

Gordon Rankin recommends “With all the 24 bit DACS you can set the Format to 44.1, 48, 88.2 or 96K. If you are using only Red Book CD’s that are ripped to your hard drive I would suggest setting the output to 88.2K. This seems to be the best sounding setup for Red Book (i.e. 16/44.1K). This will tell iTunes to upsample from 16/44.1 to 24/88.2K. The idea here is that even multiple upsampling is much better sounding than odd (i.e. 44.1*2 = 88.2 will sound better than 44.1*2.177 = 96) If you would rather not upsample then set this to 44.1K and iTunes will merely pad the 16 bit data with zero to make the output 24/44.1.”

 

I this is the case would this not apply to other DAC’s?

 

Link to comment

I was initially going to side with those who recommend against letting iTunes up-sample, based on a couple of tests I'd tried with the Apple AU resampler in Wave Editor. However I thought I should try routing the output of iTunes via soundflower to Wave Editor. Unfortunately soundflower doesn't support 88.2khz so I had to use 96khz which should "intuitively" be worse than 88.2khz. The results were quite surprising

 

I've used RMAA's spectrum analyser with default settings to generate attached pngs.

 

Procedure was to generate a two tone .wav with one tone at 1khz -60dB and a second at -10dB 10khz. The file was generated at 32bit/44.1khz then dithered down to 16bit.

 

Using soundflower set to 2 channel 32bit/96khz as the default input and output I played the test file in iTunes and recorded the output in Wave editor 1.4.6.b4. This procedure should capture any alterations caused by iTunes resampling to 96khz.

 

To my eye upsampled output looks pretty clean when compared with the original file, but it appears the noise floor is increased by roughly 6dB.

 

cheers

Paul

 

Link to comment

Paul (pj), I think you erred in assuming that the SRC in Wave Editor using "Apple AU Converter" is the same as the SRC employed by iTunes.

 

The Core Audio SRC has several quality levels, and we don't know which quality level is used by either Wave Editor or iTunes.

 

HQPlayer (on 3.8 GHz 8-core i7 iMac 2020) > NAA (on 2012 Mac Mini i7) > RME ADI-2 v2 > Benchmark AHB-2 > Thiel 3.7

Link to comment

kirk mc: What do you mean by Core Audio quality?

 

What I refer to is the AudioConverter framework. It has two parameters that, in combination, determine SRC quality:

 

1. SampleRateConverterQuality (srcq) has 5 possible values ranging from Min to Max.

 

2. SampleRateConverterComplexity (srca) has 3 possible values: Linear interpolation (line), Normal quality (norm), and Mastering quality (bats). iTunes never uses Mastering quality (according to an Apple engineer on the Core Audio discussion list).

 

HQPlayer (on 3.8 GHz 8-core i7 iMac 2020) > NAA (on 2012 Mac Mini i7) > RME ADI-2 v2 > Benchmark AHB-2 > Thiel 3.7

Link to comment

 

"Unfortunately soundflower doesn't support 88.2khz so I had to use 96khz which should "intuitively" be worse than 88.2khz."

 

Barry diament has stated that after much testing of SRCs, he finds the iZotope 64bit SRC (used in Wave Editor) to be one of the absolute best, claiming that their algorithms are of such quality than the intuitive advantage integer math might have is not in play, IOW, any suspected detriment from the more complex math is not evident in the 'hearing'.

 

clay

 

 

Link to comment

I've taken another look at the resampling using a 32/44.1 20hz to 22050hz sine sweep, and looked at the spectrograph of the resampled output. iTunes 44.1 -> 96 SRC doesn't look so good lots of aliasing. I've also upsampled using WE's iZotope SRC "Ultra-steep, Linear phase" preset and the result appears to be far superior to iTunes SRC.

 

I guess I have to retract my previous comments praising the quality of iTunes resampled output.

 

cheers

Paul

 

 

 

Link to comment

People have differing tastes, but personally I find that for up-sampling a CD born file, linear phase filtering isn't my favorite. I much prefer using a minimum phase filter to go to 88.2 KHz, 24 bits. I don't even try to do this conversion in real time. The files are converted and stored on the disc as ready to play 88.2/24 aif files.

 

As they say, YMMV.

 

Link to comment

If you are going to upsample - converting files "off-line" with a high quality SRC does seem to be the best approach.

 

I was wanting to explore for my personal edification whether there were any negatives to allowing iTunes to upsample 16/44.1 source material as a default. Based on the spectrograms of the swept sine I'll be playing files via iTunes at their native sample rate in future.

 

Link to comment

Using iTunes at the native format seems like the best approach. If you try to do any processing in real time, you tax the processor more, which forces more use of virtual memory, which all causes more power supply activity. In the end that adds more noise to an already noisy system.

 

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...