DAC burn-in/break-in? My Gustard X16 measurements

[Pdxwayne]

Doesn't really matter since I don't trust non double blind tests.

Hmm? This is online blind tests. As blind as it gets.

I passed 0.5 dB with $50 headphones on my first pass. How about you? You can't hear things doesn't mean no one else could.

 

[solderdude]

I turned down level of REW a little to typical listening db. Test again after 2 hours of powered on. E30 (filter 1) and X16(filter 2).

The dB levels of REW are NOT calibrated. You need to find out at which level the input clips and then dial the source 3 dB down.

The Gustard uses 2 DAC chips, the E30 one.
Besides, after the chips there are opamps with resistors. The latter have tolerances.
These tolerances and component differences can easily become 0.2dB (about 2%) which I think is on the high side for electronics.
I haven't seen many headphones yet with a channel balance within 0.2dB over the entire audible band.

Yes, on AB you can hear 0.5dB differences. (I can detect them too) but not 0.2% using music. 0.1% I cannot detect, nor with tones nor with music.
About 30 years ago I built this device to check my (and those of others) hearing limits.

 

[Pdxwayne]

The dB levels of REW are NOT calibrated. You need to find out at which level the input clips and then dial the source 3 dB down.

The Gustard uses 2 DAC chips, the E30 one.
Besides, after the chips there are opamps with resistors. The latter have tolerances.
These tolerances and component differences can easily become 0.2dB (about 2%) which I think is on the high side for electronics.
I haven't seen many headphones yet with a channel balance within 0.2dB over the entire audible band.

Yes, on AB you can hear 0.5dB differences. (I can detect them too) but not 0.2% using music. 0.1% I cannot detect, nor with tones nor with music.
About 30 years ago I built this device to check my (and those of others) hearing limits.

You use 0.2%, same as 0.2db? So you can hear as low as 0.2db with tones? That is cool!

Anyway, how about 0.3db with simple repeated tones music? Thanks!

 

[solderdude]

0.2dB overall level difference with tones (is close to 0.2%) on direct switching ... NOT when there is a pause between switching.
Not so with music. That will be between 0.2dB and 0.5dB as I could detect 0.5dB but not 0.2dB

http://www.sengpielaudio.com/calculator-thd.htm
This calculator lets you calculate dB to % and % to dB

 

[Frio]

Hmm? This is online blind tests. As blind as it gets.

If you read here something you don't understand you can Google it. https://www.verywellmind.com/what-is-a-double-blind-study-2795103

 

[Pdxwayne]

0.2dB overall level difference with tones (is close to 0.2%)
Not so with music. That will be between 0.2% and 0.5% as I could detect 0.5dB but not 0.2dB

http://www.sengpielaudio.com/calculator-thd.htm
This calculator lets you calculate dB to % and % to dB

Thanks!

Considering the combining effects of difference in filters and the difference in channel mismatch, certain freq difference can be around 0.3db between e30 and x16.

Possible at all to hear that 0.3db diff if music is consists of simple tones?

Thanks!

 

[Jimbob54]

Now, on to the test that matters. Can you tell whether its the e30 or x16 playing in a proper test. Does your ability to pick it change if either dac is low temp /just booted.

You can't start seeing quirks in measurements then saying cumulatively they might add up to the lack of sparkle you heard in uncontrolled listening.

 

[Pdxwayne]

Now, on to the test that matters. Can you tell whether its the e30 or x16 playing in a proper test. Does your ability to pick it change if either dac is low temp /just booted.

You can't start seeing quirks in measurements then saying cumulatively they might add up to the lack of sparkle you heard in uncontrolled listening.

I thought I would find nothing....But the difference are at the edge of the limit set by Gustard, which is 0.2db. They said they would replace one chip if the difference is bigger than 0.2db. Great for them to be willing to do that.

I am curious. What is the smallest dB change you can sense?

 

[Pdxwayne]

If you read here something you don't understand you can Google it. https://www.verywellmind.com/what-is-a-double-blind-study-2795103

So what exactly is the lowest dB difference you believe anyone can sense?

 

[Frio]

I am curious. What is the smallest dB change you can sense?

Why because you are trigger happy to hear that somebody hears those 0.2db. ok, I can hear that.

 

[Jimbob54]

I thought I would find nothing....But the difference are at the edge of the limit set by Gustard, which is 0.2db. They said they would replace one chip if the difference is bigger than 0.2db. Great for them to be willing to do that.

I am curious. What is the smallest dB change you can sense?

Irrelevant- and I have an inkling my ears are broken anyway- I have started adding a balance on headphones of over 0.5dB to the right to stop everything being too biased to the left. And thats not a DAC, HP or amp issue as its on multiple units of each.

Focus on the thing you are seeking to prove, dont measure things willy nilly and embark on a voyage of discovery every time you hook something up. The Gustard didnt sound right compared to the E30- so design a test that will prove that is a repeatable phenomenon. Then use the experts here (not me, obviously) to help understand what measurements might explain that.

 

[Frio]

So what exactly is the lowest dB difference you believe anyone can sense?

It is not about 'believe' we are not in church here and you totally ignore that every step in an audio chain can have a channel imbalance nothing will lead to your dac or something you heard two weeks ago.

 

[Jimbob54]

Irrelevant- and I have an inkling my ears are broken anyway- I have started adding a balance on headphones of over 0.5dB to the right to stop everything being too biased to the left. And thats not a DAC, HP or amp issue as its on multiple units of each.

Focus on the thing you are seeking to prove, dont measure things willy nilly and embark on a voyage of discovery every time you hook something up. The Gustard didnt sound right compared to the E30- so design a test that will prove that is a repeatable phenomenon. Then use the experts here (not me, obviously) to help understand what measurements might explain that.

So, moving things on. Had the X16 not had the imbalance you measured, what would you conclude from the tests you have run so far?

 

[solderdude]

0.2dB overall level difference is not enough to hear obvious differences and make treble sound 'not sparkly' and 'echo trails' to be longer.
You would need many dB's for that. There aren't any 0.2dB level difference between L and R in your room on speakers is already much smaller than those and even your headphones will have smaller differences.

When you want to know just use your favorite recording and make the L or R channel 0.1dB less loud, make one with 0.2dB less loud, 0.3dB, 0.4dB and 0.5dB and use ABX tester to determine WHAT you can hear using music. Then you know for yourself. use the E30.

You need to determine YOUR hearing thresholds.


Another test would be to put those 6 files (incl the original) in a folder. Use a player that can randomly play from that folder and just listen and hit 'next' now and then without looking at the screen. Then guess what dB difference there is and then check to see if you got it right.

Enjoy.

 

[chris719]

As above, whatever your perceived problem is, it's not a 0.2dB difference in level between channels. Maybe your Gustard is broken somehow, but this is not the right test to determine that either.

 

[mansr]

These tolerances and component differences can easily become 0.2dB (about 2%) which I think is on the high side for electronics.

1% seems to be the standard choice for general-purpose resistors these days. For parts where matching matters, I'd probably use 0.1%. They're cheap enough that I see no reason not to. 0.01% starts getting expensive.

 

[lizhuoyin]

Unfortunately. Same issue. Right is lower by ~0.2db.

Top two lines are X16 XLR outputs. Using Filter 2. It was powered on from cold 3 hours.

View attachment 115387

It looks like the dB diff between SE and XLR is less than 6.02dB. Does that mean the Vrms change is less than doubled? Since X16 scales well and SE is fixed at 2Vrms at 0dB, can we conclude that output from XLR at 0dB is less than 4Vrms in general or at least in this sample?
I would consider it a more serious issue here.

 

[solderdude]

1% seems to be the standard choice for general-purpose resistors these days. For parts where matching matters, I'd probably use 0.1%. They're cheap enough that I see no reason not to. 0.01% starts getting expensive.

Yes, I meant all tolerances (chip and various resistors) can easily add up (become) 0.2% (in either direction)

 

[Pdxwayne]

Why because you are trigger happy to hear that somebody hears those 0.2db. ok, I can hear that.

If you can't even show you can hear 0.5db difference (giving excuses why you don't trust an online test), and then simply resort to childish respond, would you please just ignore this thread and be happy somewhere else??

Have a great rest of day!

 

[Pdxwayne]

So, moving things on. Had the X16 not had the imbalance you measured, what would you conclude from the tests you have run so far?

I already said in https://www.audiosciencereview.com/...my-gustard-x16-measurements.20679/post-688756

Quote:
"If indeed I still find nothing significant with x16, I will let it go, thanks all the experts here, and enjoy musics."