Head-Fi.org › Forums › Equipment Forums › Sound Science › iMac onboard DAC vs. external?
New Posts  All Forums:Forum Nav:

iMac onboard DAC vs. external? - Page 3

post #31 of 82

I'm not sure how they derive the S/N ratio of a digital output, as that is totally different to the S/N ratio of an analog output. 

 

Anyhow, here is a quick measurement of the headphone output of my MacBook Pro versus the the analogue loopback of the ULN-2 which I used for the measurement. These shouldn't be taken as professional measurements, just as a comparison. The measurements were done at around -6dB. I think a 20-40 dB difference in the amount of distortion at the levels shown is significant enough. 

 

 

 

 

 

To answer the OPs question though: You can do better than the analogue output of a Mac by a considerable margin. I haven't compared DACs with the TU-882R, but I'd say a basically decent DAC would benefit it.

post #32 of 82
Quote:
Originally Posted by OldHarbourGreen View Post

Joining the thread late.  iMac specs from Apple below.  S/PDIF has higher SNR and lower distortion.  

 

I/O ports

1. Headphone port

Line/headphone output 
The headphone output is automatically selected for audio output if no external S/PDIF optical digital output device is detected. The headphone output supports a stereo data stream at bit depths of 16, 20, or 24 bits per sample and at sample rates of 44.1 kHz, 48 kHz, or 96 kHz. The headphone output volume can be adjusted from 0.0 dB to -43.0 dB.

During playback of a 1 kHz sine wave at -3 dBFS voltage level, 24-bit sample depth, 44.1 kHz output sample rate, 100 k load (unless otherwise specified), the audio output has the following nominal specifications:

  • Jack type: 3.5 mm (1/8-inch) stereo combo
  • Maximum output voltage: 1.4 VRMS (+5.15 dBu)
  • Output impedance: <24 ohms
  • Frequency response: 20 Hz to 20 kHz, +0.5 dB/-3 dB
  • Signal-to-noise ratio (SNR): >90 dB
  • Total harmonic distortion + noise (THD+N): <-80 dB (0.007%)
  • Channel separation: >85 dB

 

 

 

These are frankly pretty poor, certainly worse than red book requires and done with a 24 bit signal no less and at -3db which makes me think it may clip at full whack - compare that to 

 

 

http://www.stereophile.com/content/marantz-cd5004-cd-player-marantz-cd5004-cd-player-measurements - a budget CD player

 

http://www.ebay.com/itm/Professional-HiFiMAN-EF5-Headphone-Amplifier/170696111860 - a cheapo HK made DAC/headphone amp or basically any number of low cost eBay Chinese DAc/amps

 

Of course this does not mean it sounds bad in any way but spec wise it is unimpressive !

post #33 of 82
Quote:
Originally Posted by nick_charles View Post

 

 

These are frankly pretty poor, certainly worse than red book requires and done with a 24 bit signal no less and at -3db which makes me think it may clip at full whack - compare that to 

 

 

http://www.stereophile.com/content/marantz-cd5004-cd-player-marantz-cd5004-cd-player-measurements - a budget CD player

 

http://www.ebay.com/itm/Professional-HiFiMAN-EF5-Headphone-Amplifier/170696111860 - a cheapo HK made DAC/headphone amp or basically any number of low cost eBay Chinese DAc/amps

 

Of course this does not mean it sounds bad in any way but spec wise it is unimpressive !

Red book is a set of CD specs that include nothing about audio performance.  It's formatting, data structure, etc.  There are no red book audio requirements.  

 

All D/As clip at "full whack".  None would clip at -3dB.  

 

Careful when comparing a measured noise figure to a noise spectrum plot.  The plot will always look much better because the measured noise is a sum of the entire bandwidth, whereas the spectrum plot is a measurement based on the maximum resolution bandwidth, which is very small.  There's also nothing about weighting, noise shaping, etc.  

 

The eBay DAC specs are unrealistic, highly doubtful, and unprovable by almost any buyer.  Not to say it's not "good", just likely exaggerated.  

 

The difference between the Mac and the other devices would likely be too small to be a audible with any actual audio recording, though it may not measure as well.  

post #34 of 82
Quote:
Originally Posted by jaddie View Post

Red book is a set of CD specs that include nothing about audio performance.  It's formatting, data structure, etc.  There are no red book audio requirements.

 

CD theoretically anyway has a SNR of about 96db the Apple onboard device at 90db is around 15 bits, probably not noticeable but certainly not state of the art, in truth you are unlikely to notice the difference between 14 bits and 16 bits except at absurd volume levels like the Meyer and Moran study

 

All D/As clip at "full whack".  None would clip at -3dB.

 

Interesting, then why do so many engineers seem to insist on hitting the end-stops if it just leads to distortion frown.gif

 

 

Careful when comparing a measured noise figure to a noise spectrum plot.  The plot will always look much better because the measured noise is a sum of the entire bandwidth, whereas the spectrum plot is a measurement based on the maximum resolution bandwidth, which is very small.  There's also nothing about weighting, noise shaping, etc.  

 

The eBay DAC specs are unrealistic, highly doubtful, and unprovable by almost any buyer.  Not to say it's not "good", just likely exaggerated.  

 

Hmmm, I've seen far more "optimistic" specs on some mainstream gear (add another couple of zeros) and some like the Benchmark stuff seems to have measurements to back it up, the one positive thing about Stereophile being that they do actually measure stuff, not turntables of course.... wonder why wink.gif

 

But yeah the general lack of actual measurements is highly annoying !

 

 

The difference between the Mac and the other devices would likely be too small to be a audible with any actual audio recording, though it may not measure as well.

 

Agreed , but it is always fun to bash Apple just a little bit...

 

 

post #35 of 82
Quote:
Originally Posted by jaddie View Post

The difference between the Mac and the other devices would likely be too small to be a audible with any actual audio recording, though it may not measure as well.

That's really all that needs to be said. When the specs on a player exceed the specs on our ears by a fair margin, there's no need to throw out numbers any more.
post #36 of 82
Quote:
Originally Posted by nick_charles View Post

CD theoretically anyway has a SNR of about 96db the Apple onboard device at 90db is around 15 bits, probably not noticeable but certainly not state of the art, in truth you are unlikely to notice the difference between 14 bits and 16 bits except at absurd volume levels like the Meyer and Moran study

 

Yes, but no 16 bit device ever does the theoretical 96dB without quite a bit of trickery like oversampling DACs and noise shaping. But it doesn't matter because if you hit it with 16 bit audio data you'll never get 96dB anyway, and you wouldn't want to because you do want that data dithered, and that alone will kick you down to 93 or less.  If the audio was digitized at 16 bits there'll be 3dB at least of least significant bit jitter.  And if digitized at 24 bits, the conversion to 16 should never just be truncation, sort of defeats the purpose.  So 90dB, real world, is about right for anything.  The stuff that measures better isn't measured with a digitized analog signal, and yes the measurements are real, but you have to consider the noise floor of real audio material too, and noise-shaping, over-sampling DACs can't help that.  There are better measuring DACS, but that doesn't mean they'll sound better.  And, 96dB or even 90dB is pretty much more that we need.

 

quote: "Interesting, then why do so many engineers seem to insist on hitting the end-stops if it just leads to distortion frown.gif"

 

It's the loudness war, and the end stops they are actually hitting aren't in the DAC, they're in the data processed by some sort of DSP that may involve clipping or hard limiting.  Yes, it's nasty, but until people stop buying, and start returning music that's so loud its distorted for their money back, engineers (technically, producers) will keep making it.

 

 

quote: "Hmmm, I've seen far more "optimistic" specs on some mainstream gear (add another couple of zeros) and some like the Benchmark stuff seems to have measurements to back it up, the one positive thing about Stereophile being that they do actually measure stuff, not turntables of course.... wonder why wink.gif

 

But yeah the general lack of actual measurements is highly annoying !"

 

Yup, sure is.  And yes, there are lots of things that measure better.  Like I said, it's not all real once you start playing audio though.  It might be me being cranky and old, but I tend to ignore DACs that measure better than theoretical because once you shove real digital audio into it instead of computer generated test data you're back to 16 bits again.  Sort of pointless, but since so-called 24 bit ADCs behave like 18-20 bit ADCs, and when audio gets shoved up to the top end and squashed there, the discussion is really academic.  Yes, it's fun to bash Apple, but frankly, they put pretty decent audio into their devices, on average way better than the integrated PC sound junk. And there are some VERY smart people in the audio department there.  It's likely to get even better.

 

 

post #37 of 82

I own an iMac 27' core i7. I've always thought the sound quality is decent compared to other sound cards. 

 

When I got the JH16's the bottleneck was the sound card in my setup. I added a DAC/AMP ( Yulong D100 MKII) connected to S/PDIF and the difference is huge. I mean very huge.

 

1 - There is no audible white noise or hiss when nothing is playing no matter how loud the volume is

 

2- Bass is much more controlled with rumble, not bloated when volume is high

 

3 - Soundstage is definitely wider 

 

4 - Upsampling/downsampling and option to dynamically switch the sample rate and bit depth depending on the track playing.

 

5- No electrical interference ( HDD, ground or USB device noise )

post #38 of 82

If you were getting hiss and grounding interference at normal listening volumes, even very loud ones, there is something wrong with your iMac. If it's still under warranty, I would take it to the Apple store and get them to check it out.


Edited by bigshot - 5/27/13 at 5:15pm
post #39 of 82

You get a fair bit of beeping and buzzing from the headphone/line out of Macs. This is normal. It's much worse with very sensitive IEMs such as most customs. It doesn't show up in measurements because the input impedance of measuring equipment is 50 Ohms or greater and some sounds like buzzing comes randomly.

post #40 of 82

My Mac Mini and iMac 29 are stone silent. I can crank the volume way up with my cmoy and it's still clean. I have mirrordoor PowerMacs too and those are as clean as clean can be. Perhaps other models have higher defect rates. What model of Mac do you have?


Edited by bigshot - 5/27/13 at 6:47pm
post #41 of 82

I can also vouch that the headphone out of my iMac is very clean as well, it was even clean with the very sensitive Denon D2000 when I used to have that.

 

The only time I get a good amount of noise from the iMac's headphone out is if I hook it up to a my stereo receiver and turn the volume up quite a bit on the receiver for the speakers.  The sheer magnitude of the receiver's gain really amplifies the noise of the iMac's headphone out, but near maxing the iMac's sound out and using less juice on the speaker amp solves that.

post #42 of 82
Quote:
Originally Posted by bigshot View Post

My Mac Mini and iMac 29 are stone silent. I can crank the volume way up with my cmoy and it's still clean. I have mirrordoor PowerMacs too and those are as clean as clean can be. Perhaps other models have higher defect rates. What model of Mac do you have?

 

I'm listening now with the XBA-30s and don't hear anything. However, CIEMs tend to be much more sensitive. I don't have anything as that here at the moment to test with though. I do remember managing to get computer noise from it at one point, though I can't remember with what. Possibly it was with an amp.

post #43 of 82
Quote:
Originally Posted by bigshot View Post

If you were getting hiss and grounding interference at normal listening volumes, even very loud ones, there is something wrong with your iMac. If it's still under warranty, I would take it to the Apple store and get them to check it out.

 

I am afraid to say that there isn't anything wrong with my iMac, that's been the case with my two older iMacs as well. It's completely normal. With high sensitivity IEMs like Shure SE535 and JH16,

 

One important thing to remember is that this is mostly noticeable under BootCamp and not OSX because under the latter, the sound card goes into indle state after few seconds of silence whereas under Bootcamp, the soundcard driver is always active.

post #44 of 82
Are you talking about an impedance mismatch?
post #45 of 82

I did a little googling and determined that you are indeed talking about a mismatch with low impedance earphones. That isn't the fault of the DAC. It isn't even really the fault of the Mac any more than it is the earphones. Either select a headphone that matches well with the Mac, or use a headphone amp and the problems will be solved.

 

That's why I was completely confused by people saying that the Mac had bad sound. With my Sennheiser headphones, the output is pretty near perfect. When I amp it, I can crank it way up and it's extremely clean.

New Posts  All Forums:Forum Nav:
  Return Home
  Back to Forum: Sound Science
Head-Fi.org › Forums › Equipment Forums › Sound Science › iMac onboard DAC vs. external?