CHANGSTAR: Audiophile Headphone Reviews and Early 90s Style BBS

  • December 31, 2015, 10:10:39 AM
  • Welcome, Guest
Please login or register.

Login with username, password and session length
Advanced search  

Author Topic: Measurable Difference of NOS DACs on Headphones - An Experiment  (Read 1240 times)

0 Members and 1 Guest are viewing this topic.

Hands

  • Pizza the Hutt
  • Mate
  • Pirate
  • ****
  • Brownie Points: +331/-8
  • Offline Offline
  • Posts: 1591
  • Master of Revelatory Bird Calls and Fine Art
Measurable Difference of NOS DACs on Headphones - An Experiment
« on: August 09, 2014, 06:02:50 PM »
Most or all of us have seen some of the typical characteristics NOS DACs usually exhibit in measurements, and we all know they sound like absolute garbage (that's why I use them). But how much and how so, if at all, do they affect a headphone's performance on a measurable level? I decided to grab my HD600 (stock) and find out.

Normally for headphone measurements, I use my Creative SB1240 as the DAC, feed that to my Leckerton UHA-6S Mk.II's input, run my headphones from that amp, and mic them back to the SB1240's mic-in. This is from my desktop, which will provide more than adequate performance for headphone measurements (DACs are much more sensitive in this regard if you want to get the measurements just right). In Windows, I have the DAC set to a volume setting of 100, and I set the mic-in to 40/100. I have found that 40/100 provides the best looking and most consistent harmonic distortion measurements. The headphones are calibrated to roughly 90dB at 1KHz using a basic SPL meter and coupler.

For this test, I simply substituted the Metrum Quad, being fed by the JKSPDIF Mk3, as the DAC when necessary. Both output at 2V, so no level matching was necessary (or, close enough). While the SB1240 performs objectively excellent via it's USB input, better than the Metrum Quad in most/all regards, I want to make it known right away that the JKSPDIF is indeed one extra part of the experiment that could skew the results (i.e. some differences may be due to that and less so from Metrum Quad itself).

In this post, you'll see I've attached FR/THD and impulse response measurements for both. Do note that the impulse response graphs aren't scaled the exact same. The windows are set to the same amount of time, but don't necessarily line up exactly (Metrum exhibits that drop before the impulse, SB1240 exhibits pre-ringing that is more noticeable zoomed-in, so it's hard to match up the start of the window exactly). Next two posts will have more pics. These were done in 24/48.
« Last Edit: August 09, 2014, 06:14:18 PM by hans030390 »
Logged
The other master and I invite you to visit our digital museum of fine art and revelatory bird calls: https://www.facebook.com/SchrodsonkMuseum

Hands

  • Pizza the Hutt
  • Mate
  • Pirate
  • ****
  • Brownie Points: +331/-8
  • Offline Offline
  • Posts: 1591
  • Master of Revelatory Bird Calls and Fine Art
Metrum Specific Measurements
« Reply #1 on: August 09, 2014, 06:06:03 PM »
Metrum specific measurements. Attached to this post, you'll see a 24/96, 999Hz sine wave test at 0dB. I only did 999Hz because ARTA likes to switch my numbers to something slightly off if I go straight for 1KHz, so, whatever.

Next is 999Hz at 24/48.

Next are 15KHz measurements at 24/48, at both 0dB and -10dB.

Keep in mind that measurements will show this well below 0dB due to the mic-in volume setting. It's either that, or you start introducing more background noise into the picture or just get garbage.

In hindsight, I should have grabbed 16-bit measurements. Maybe later.

Update: Took out the waveform measurements due to how the ADC/software report the results.
« Last Edit: August 10, 2014, 06:09:50 PM by hans030390 »
Logged
The other master and I invite you to visit our digital museum of fine art and revelatory bird calls: https://www.facebook.com/SchrodsonkMuseum

Hands

  • Pizza the Hutt
  • Mate
  • Pirate
  • ****
  • Brownie Points: +331/-8
  • Offline Offline
  • Posts: 1591
  • Master of Revelatory Bird Calls and Fine Art
SB1240 Specific Measurements
« Reply #2 on: August 09, 2014, 06:07:34 PM »
Same as above, but for SB1240.

Update: Took out the waveform measurements due to how the ADC/software report the results.
« Last Edit: August 10, 2014, 06:07:17 PM by hans030390 »
Logged
The other master and I invite you to visit our digital museum of fine art and revelatory bird calls: https://www.facebook.com/SchrodsonkMuseum

firev1

  • Cynophobic Puss
  • Able Bodied Sailor
  • Pirate
  • ***
  • Brownie Points: +52/-0
  • Offline Offline
  • Posts: 490
Re: Measurable Difference of NOS DACs on Headphones - An Experiment
« Reply #3 on: August 10, 2014, 11:19:31 AM »
Slight shifts in the minimal distortion, can't say much since the WM61a is not really meant for these sorts of measurements. Are they very repeatable?
Logged
Time spent on enjoyment is not time wasted. - someone

Hands

  • Pizza the Hutt
  • Mate
  • Pirate
  • ****
  • Brownie Points: +331/-8
  • Offline Offline
  • Posts: 1591
  • Master of Revelatory Bird Calls and Fine Art
Re: Measurable Difference of NOS DACs on Headphones - An Experiment
« Reply #4 on: August 10, 2014, 02:50:11 PM »
I took about 3-4 measurements of FR and HD each to verify consistency before saving any results. Didn't shift the HP on my head at all between DACs, was still, in same sitting position, etc. to minimize any differences, though I'd be curious to see if they'd still hold if I were to purposely shift things around a bit with each measurement. Not to mention, I have so many other stock measurements of the HD600, and they always come out looking like the SB1240 shots here as well with almost zero effort (such a well behaved headphone for measurements!). So, all in all, results should be quite repeatable. Definitely can check that over again in the next few days. :)

I kinda posted this thread in a hurry (which I didn't make clear), but I'm sure there are explanations for some of slight differences that stem more from the entire chain and how the pieces work together rather than the individual DACs themselves. There are still some things I'd like to double check and new tests I'd like to run, but these preliminary results are interesting for what they are.

Hold up on those 15KHz measurements, though...something else is definitely going on there that I need to investigate further.

UPDATE: It would appear my SB1240 has issues with output and/or recording higher treble measurements. Just did loopback test on it with my tablet to rule out some ideas. Spectrum results can be great, but the waveform itself takes on that weird look. I'll need to investigate further. My current theory is that it's a limitation with on the ADC side.

UPDATE 2: Definitely the ADC and/or the software combination. I think you really need something closer to an oscilloscope to capture these sort of waveforms (sorry if this is the noobest of noob knowledge, but I'm still piecing some of this together). Verified with my oscilloscope that the SB1240 puts out nice looking 15KHz sine waves and the Metrum + NOS1704 I have on hand definitely do not. Though it is interesting watching it improve real-time as you increase the sampling rate.

Anyway, spectrum measurements should still be OK to get some information from. I am fairly confident in what my FR/HD measurements report.
« Last Edit: August 10, 2014, 06:06:54 PM by hans030390 »
Logged
The other master and I invite you to visit our digital museum of fine art and revelatory bird calls: https://www.facebook.com/SchrodsonkMuseum

Boner Stabone

  • Able Bodied Sailor
  • Powder Monkey
  • ***
  • Brownie Points: +14/-15
  • Offline Offline
  • Posts: 26
Re: Measurable Difference of NOS DACs on Headphones - An Experiment
« Reply #5 on: August 11, 2014, 03:32:26 AM »
i dun understand. can you explain? too many graphs.
Logged