Why is it when testing the difference between 2 DAC/AMPs it only becomes apparent after you have listed to One for a while?

[stereo coffee]

So you can hear differences under controlled blind listening tests between the mentioned opamps in an audio application with just a few times gain ?

Yes, but not using op amps as devices to achieve potential difference between inputs , that's just too boring. They are capable of so much more,

 

[solderdude]

Op-amps are the most universal non programmable devices/components around.


They are only boring to people who just see a piece of plastic with pins sticking out.

 

[stereo coffee]

Op-amps are the most universal non programmable devices/components around.


They are only boring to people who just see a piece of plastic with pins sticking out.

It's sad to see the usual nasty comments arising again, but have a nice day anyway..

 

[solderdude]

What's specifically nasty about my remark ?

Opamps are the most universal non programmable devices/components around and about 99.99% of the earth population will find talk about opamps is boring.

 

[ernestcarl]

Yet if you listen to one source ... and then suddenly switch the difference is so huge then when you switch back and forth quickly the difference becomes 0.

For simple ABX comparisons, what is your methodology of switching and playing music between different DAC-AMPS at the same time?

Do you use something like this: https://www.bhphotovideo.com/c/product/748807-REG/Sescom_SES_IPOD_AB_iPod_Stereo_Audio_MP3.html

In JRiver, one can play the same music sync'd and outputed to more than one DAC -- basically DACs playing in parallel. Properly level-matched and combined with the mentioned linked physical switcher, you can do quick and reliable ABX comparisons. It can be done blind *(well, maybe not exactly as you will notice when the switch is engaged) if you have a second person doing the switching and scoring for you. It's a cheap way to do it anyway.

 

[majingotan]

What's specifically nasty about my remark ?

Opamps are the most universal non programmable devices/components around and about 99.99% of the earth population will find talk about opamps is boring.

Even Schiit themselves proudly say opamps are boring (from their magni heresy blog) so they make discrete amps but cannot even reach the level of opamp transparency. Funnily enough, their Bifrost 2 DAC that I have has opamp output stage but Schiit couldn’t engineer it enough to go past 10 bits of resolution.

 

[ernestcarl]

For simple ABX comparisons, what is your methodology of switching and playing music between different DAC-AMPS at the same time?

Do you use something like this: https://www.bhphotovideo.com/c/product/748807-REG/Sescom_SES_IPOD_AB_iPod_Stereo_Audio_MP3.html

In JRiver, one can play the same music sync'd and outputed to more than one DAC -- basically DACs playing in parallel. Properly level-matched and combined with the mentioned linked physical switcher, you can do quick and reliable ABX comparisons. It can be done blind *(well, maybe not exactly as you will notice when the switch is engaged) if you have a second person doing the switching and scoring for you. It's a cheap way to do it anyway.

Here's what I have which is better: https://www.bhphotovideo.com/c/product/1023552-REG/sescom_ses_ab_switch_3_5mm_stereo_audio_a_b.html

It has a "middle-way" or in-between state, which allows you simulate the illusion of switching to from A to B whilst not -- say, switch from A to A to A... this probably is blind enough with said other person doing the operation.

 

[Ratatoskr]

Even Schiit themselves proudly say opamps are boring (from their magni heresy blog) so they make discrete amps but cannot even reach the level of opamp transparency. Funnily enough, their Bifrost 2 DAC that I have has opamp output stage but Schiit couldn’t engineer it enough to go past 10 bits of resolution.

I am curious how are you arriving at that 10 bits of resolution from the Bifrost 2 APx555 measurements Schiit published? Thanks.

 

[majingotan]

I am curious how are you arriving at that 10 bits of resolution from the Bifrost 2 APx555 measurements Schiit published? Thanks.

There’s a measured one here and got 10 bits only. You can’t use the axis that Schiit use since it’s not in dBFS but dBV instead. Then again, referring to the main topic, I can’t discern a sound difference between the sound of a 10 bit DAC (Bifrost 2) against a ~17 bit one (Chord Mojo) on sighted volume matched AB using my Schiit Saga preamp in passive mode (120 dB Sinad) and Yamaha HS7+HS8S speaker

 

[ernestcarl]

There’s a measured one here and got 10 bits only. You can’t use the axis that Schiit use since it’s not in dBFS but dBV instead. Then again, referring to the main topic, I can’t discern a sound difference between the sound of a 10 bit DAC (Bifrost 2) against a ~17 bit one (Chord Mojo) on sighted volume matched AB using my Schiit Saga preamp in passive mode (120 dB Sinad) and Yamaha HS7+HS8S speaker

View attachment 41295

Did you try testing with more resolving headphones?

 

[stereo coffee]

Did you try testing with more resolving headphones?

Or attenuation that does not use contacts for audio paths ?

 

[Ratatoskr]

There’s a measured one here and got 10 bits only. You can’t use the axis that Schiit use since it’s not in dBFS but dBV instead. Then again, referring to the main topic, I can’t discern a sound difference between the sound of a 10 bit DAC (Bifrost 2) against a ~17 bit one (Chord Mojo) on sighted volume matched AB using my Schiit Saga preamp in passive mode (120 dB Sinad) and Yamaha HS7+HS8S speaker

View attachment 41295

Thanks for the input. Yes I read the original Bifrost Multibit review here but the Bifrost 2 uses upgraded 18 bit Analog Devices D/A chips. Perhaps the forum can review a Bifrost 2 in time.

 

[solderdude]

Or attenuation that does not use contacts for audio paths ?

Like the one that uses non-linear and voltage dependent light sensitive components that you said you would send in for measurements a year ago and which you are selling and still has input and output contacts in the audio path ?

 

[MRC01]

...See, we humans aren't very good at remembering what we hear. Especially when it comes to details in complex arrangements of sounds. ... The way to avoid these things is to eliminate bias and take measures to manage our difficulty with remembering details of what we hear. Thus, volume matched blind A/B/X testing...

...long term a large enough sound quality difference can be heard reliably, but short term is capable of better finer discrimination....

Short-term memory is indeed better than long-term, though still imperfect. Yet in an ABX test, we are always relying on memory. We can't listen to 2 different sounds simultaneously, so even with instantaneous switching we are comparing what we are hearing now, with our memory of what we heard a moment ago. When A/B/X testing, switching delays as short as a fraction of a second impair sensitivity, which tells us that even our short-term audio memory is imperfect.
From this one may imagine the conundrum that our hearing acuity, as good, bad, or inconsistent as it may be, is better than our memory of it.

 

[MRC01]

For simple ABX comparisons, what is your methodology of switching and playing music between different DAC-AMPS at the same time?
...

If it's comparing 2 audio files, I use an ABX software app. If it's 2 live devices playing (like a DAC or amp), I use a digital optical switchbox. Either provides virtually instantaneous switching of any 2 sources. It's up to you to synchronize and level-match them, but that's not too hard. And you can have a helper hit the switch for you, so you don't know which you're listening to, and neither do they.

 

[HammerSandwich]

wow didn't know our own ears are so unreliable

Youtube "mcgurk effect." You get only one chance to see it for the first time, so listen initially with your eyes closed. It will blow your mind.

 

[RayDunzl]

wow didn't know our own ears are so unreliable, I suppose it is just how we evolved after all.

Your eyes are unreliable too.

The squares marked A and B are the exact same shade of gray:

 

[ernestcarl]

Much harder when it's dac-amp combos

If it's comparing 2 audio files, I use an ABX software app. If it's 2 live devices playing (like a DAC or amp), I use a digital optical switchbox. Either provides virtually instantaneous switching of any 2 sources. It's up to you to synchronize and level-match them, but that's not too hard. And you can have a helper hit the switch for you, so you don't know which you're listening to, and neither do they.

Those are good... But with two dac+amp combos or two different HP amps, a physical switch still is better than plugging and replugging cables.

 

[stereo coffee]

Like the one that uses non-linear and voltage dependent light sensitive components that you said you would send in for measurements a year ago and which you are selling and still has input and output contacts in the audio path ?

As I said soon. In my implementation there are no contacts in the audio signal path neither for volume or signal input switching. The option of direct wiring rather than using a RCA socket is available, if we view the interconnect socket and plug also as a switch - so Yes free of any contacts.
The NSL32SR3 LDr is unique in providing a extremely high resistance Off figure which benefits ability for contact-less input switching, which I discovered over 10 years ago.

As for non linearity these are your words not mine. however I cover the topic here exposing what you allude to, namely the absolutely miserable ways some manufacturers have powered LDR's : https://www.stereo.net.au/forums/topic/293104-audio-purpose-ldr-discoveries/ and offer insight into what begin's to make a difference. You will gather from my findings, the LDR simply reflects in its audio capability, your good or poor efforts with powering its anode and cathode. My design effort has been toward improving the anode and cathode relationship... and is quite fascinating how far this can go.

 

[Sgt. Ear Ache]

Short-term memory is indeed better than long-term, though still imperfect. Yet in an ABX test, we are always relying on memory. We can't listen to 2 different sounds simultaneously, so even with instantaneous switching we are comparing what we are hearing now, with our memory of what we heard a moment ago. When A/B/X testing, switching delays as short as a fraction of a second impair sensitivity, which tells us that even our short-term audio memory is imperfect.
From this one may imagine the conundrum that our hearing acuity, as good, bad, or inconsistent as it may be, is better than our memory of it.

Sure, totally true. But it's much easier to pick out and compare details between short sample bursts than it is listening to one thing for an hour and then comparing it to another thing for an hour.