Oldvinylplayer dropped in the other day on his way through town to have a peek at my newest speakers. He was aghast when I fired up iTunes to play some music.
This surprised me. I've made extensive use of iTunes since I got my first Mac in 2011. All our music is ripped in ALAC format to our Mac mini media centre computer, and shared to the various Macs, iPads and iPhones that we own. I have speakers in the bedroom and garage hanging off airport express boxes, plus an Apple TV in the living room, that allows me to airplay throughout the whole house.
I really like iTunes. I have a big investment in the whole apple ecosystem and Perry reckons I have a shrine to Jobs.
Oldvinylplayers comment was that the quality of reproduction of iTunes was lacking. He uses JRiver. I had a bit of a look at JRiver and it looks much like iTunes, but maybe a bit harder to use.
Anyway, this quality of reproduction thing had me wondering. Previously I've used Audacity to play 1KHz tones for testing amps using my MacBook Air, which was pretty good. I wondered what sort of THD performance I could get playing a prerecorded tone on my iMac via an airport express over our home wifi network.
I started by creating a 30 minute long 85% volume 1KHz tone at 44.1Ksps using Audacity. I exported this as a 16 bit AIFF (apple lossless) file, and simply double clicked on the file in finder to get iTunes to play it. I turned the actual speakers off and played via an airport express at full volume plugged directly into the 1MOhm input on my HP3585B spectrum analyser.
The plots attached show the result.
First is DC to 100KHz, showing the 5dBm 1KHz sine wave and not much else. SINAD here is around 90dB up to 30KHz, limited more by the resolution bandwidth I was using for the plot.
Next I wound the span down to DC to 10KHz, with 10KHz resolution bandwidth. SNR is ~97dB, as one would expect from 16 bit samples. There's some distortion at H2 and H3, but this is typical for my spec-an.
Last plot is with a 1KHz passive notch in circuit. This attenuates the fundamental by ~50dB, H2 by 9.6dB, H3 by 5.8dB, and H4 by 4.1dB. Our fundamental now measures -45dBm. H2 measures -106dBm, and H3 and onwards say -110dBm or better.
So I make H2 in reality -96dBc, H3 onwards ~-105dBm. I'd say that was reasonably good.
This surprised me. I've made extensive use of iTunes since I got my first Mac in 2011. All our music is ripped in ALAC format to our Mac mini media centre computer, and shared to the various Macs, iPads and iPhones that we own. I have speakers in the bedroom and garage hanging off airport express boxes, plus an Apple TV in the living room, that allows me to airplay throughout the whole house.
I really like iTunes. I have a big investment in the whole apple ecosystem and Perry reckons I have a shrine to Jobs.
Oldvinylplayers comment was that the quality of reproduction of iTunes was lacking. He uses JRiver. I had a bit of a look at JRiver and it looks much like iTunes, but maybe a bit harder to use.
Anyway, this quality of reproduction thing had me wondering. Previously I've used Audacity to play 1KHz tones for testing amps using my MacBook Air, which was pretty good. I wondered what sort of THD performance I could get playing a prerecorded tone on my iMac via an airport express over our home wifi network.
I started by creating a 30 minute long 85% volume 1KHz tone at 44.1Ksps using Audacity. I exported this as a 16 bit AIFF (apple lossless) file, and simply double clicked on the file in finder to get iTunes to play it. I turned the actual speakers off and played via an airport express at full volume plugged directly into the 1MOhm input on my HP3585B spectrum analyser.
The plots attached show the result.
First is DC to 100KHz, showing the 5dBm 1KHz sine wave and not much else. SINAD here is around 90dB up to 30KHz, limited more by the resolution bandwidth I was using for the plot.
Next I wound the span down to DC to 10KHz, with 10KHz resolution bandwidth. SNR is ~97dB, as one would expect from 16 bit samples. There's some distortion at H2 and H3, but this is typical for my spec-an.
Last plot is with a 1KHz passive notch in circuit. This attenuates the fundamental by ~50dB, H2 by 9.6dB, H3 by 5.8dB, and H4 by 4.1dB. Our fundamental now measures -45dBm. H2 measures -106dBm, and H3 and onwards say -110dBm or better.
So I make H2 in reality -96dBc, H3 onwards ~-105dBm. I'd say that was reasonably good.
