Sigh ... O.K.
While reading here on most of the forums i see that scientists prefer to spend time to intimidate people that hear differences instead of making more serious measurements.
The most used argument is that we are all audiophiles fooled by what we read on internet.
I like to think that 'scientists' as well as seasoned engineers etc. are trying to educate people that are convinced they hear differences into finding/using ways to eliminate the 'knowing' part in their test methods.
Why ?
Because they found this out already and understand that the mind is tricky.
It would be a better idea that asr would spend time to find new measurements to improve scientific knowledge.
Why is this ASR's responsibility ?
What new measurements are needed to improve scientific knowledge ?
I hope you are not suggesting to look for 'the yet unknown signal' that still needs to be discovered and explains the discrepancy between measurements and what is heard sighted.
If so... I wish you good luck on your endeavors. There are many that have tried before.
Strangely enough, when those that did just that turned to rigorous blind testing methods their abilities seemed to vanish suddenly for them and others.
Instead of that asr relies on old measurements that have been elaborated mostly for marketting.
Can you elaborate why ASR relies on 'old' measurements that are intended for marketing ?
Is it possible that manufacturers use similar test equipment and standard tests that show all the aspects technical people needs to asses operation of the DUT as Amir (or some other measurement folks) do?
In fact Amir's AP is often even 'better' than those of some manufacturers.
It's not complicate to have idea for new tests:
With no band limited measurements
IS it needed or do you feel it is needed ?
- thd vs freq at different levels
This has been measured many times.
What do you think the plot below shows ?
Add to this that it is common knowlegde amongst engineers that the highest amount of THD is at 0dB FSD (max level).
Do you see the upwards bend around 5 to 10dB ? That's where distortion starts to creep up again when nearing 0dBFS.
The lower you go in level the more the distortion products drown in noise. This is what is shown in the above graph.
Do you feel Amir needs to do full THD sweeps at different levels say 0dB, -3dB, -10dB, - 40dB and -90dB and between DC and 100kHz just to please a few readers ?
Do you have any idea how much time it takes to take good measurements and effort this takes and if showing all this is feasable in the forum format of ASR ?
If you want to see those plots just look on RAA and other websites that sometimes do this. Aside from some -3dB plots in case the DAC seems to distort heavily there is no added value for simple engineers to see these plots as what is shown here is worst case. It only gets 'better'.
Well... for most R2R implementations things get worse at lower levels though due to practical .
- same with different samplerates and bitdepths
This too isn't a 'new test' and exists. When an anomaly is found during tests Amir also publishes these. When you don't see them assume the best and don't expect an evil Amir to NOT measure it on purpose or not post it to fool audiophiles.
I suggest you use the time to educate yourself and read relevant threads on ASR and other websites that don't adhere to sighted evaluations.
Find scientific articles about digital everything and learn.
All of the above is shown in articles and measurements. It is common knowledge.
Then ... once you know all of this by heart and you have some interesting remarks or test methods just suggest them and come up with actual ideas for tests.
What you do not seem to grasp is that by showing 44.1kHz reproduction you are showing worst case. When someone decides that he will only publish the audible frequencies then so be it. Look for the measurements you need elsewhere.
Just because the forum has the words audio, science and review in it does not mean they are obligated to show all possible measurements and show them all. There are other sites that do. The auteur can decide which plots are relevant and show issues they encountered instead. That does not mean the device is not scrutinized properly.
- finding ways to measure dsd
What is it that needs to be measured 'more' about DSD ?
It is the most simple concept of digital and well defined. What measurements are shown is how the known aspects of it can be shown. Some even take it further for specific reasons.
- testing the noise produce by the PSU of the dac and by the usb cable up to ultrasonic frequencies
Why would one need to probe the PSU of the DAC ?
Also each separate voltage rail ?
Referenced to specific grounds ?
On the power pin of each essential component ?
You want capacitors measured as well ?
ALL these results should be published as well even if they do not reveal anything of concern ?
Should data like PSSR (PSRR) of the components where the noise is measured be included (data sheets or links to it) ?
All that matters is what signal comes out of the analog outputs. You see.. that includes noise on the various voltage rails and ground(planes).
Having said that you are correct that there is something that is never measured (or should I say almost never measured).
This is common mode noise.
Common mode noise is noise that is present on both signal wires (ground and signal or signal+ and signal-) that has no influence on the signal.
That noise can result in audible nasties/hum in downstream gear when something is not designed properly or not connected properly by a groundloop.
Groundloops are hard to figure out because they need to be measured in a different way and can frustrate folks to no end.
If you want to know then ask the manufacturer for their EMC tests (they SHOULD have had these performed).
Then you can 'see' how that device measures.
Alas ... it does not show how such a device will perform when improperly connected and how it interacts with other devices.
These tests merely show how much 'garbage' it emits on all its in and output ports measured under strict conditions both 'conducted' and 'radiated' (radio waves) and where the acceptable levels are that once have been established.
Another unfortunate thing is that these are not 'audiophile' levels but practical ones.
There will be no pleasing you I'm afraid.
- testing how the filter of the dac deal with the nyquist image and finding at wich frequencies it's located
I think you missed all the published plots that show this. When measurements show there are issues that are deemed important it will be elaborated on and shown. You demand even when there is nothing of concern it still must be shown and not for you alone but for the rest of the world.
This has been discussed over and over again and you keep on bitching about it.
Either seek the info elsewhere or use arguments that will appeal to the measurerment guys that are compelling.
- trying to find tests with complex signal to see if there are differences between dacs
More complex than what is out there already ? Noise, multitone, nulling music, phase, time related ?
Do you have any suggestions what meaningful different tests (that do not already exist) are really needed and say something about aspects that are not covered ?
Instead of trying to improve scientific tests asr try to fight against people that need more proof and asr call them audiophiles.
ASR is not trying to fight against people that want (I would not say need) even more proof. It is trying to show relevant measurements and on the side explains and tries to answer questions and suggestions.
What is often questioned is HOW some folks that have 'tested' equipment and have drawn their conclusions based on test
methods that have been known (for decades already) to be flawed.
Audiophile is not a negative word... Audiophool (or audiofool) is.
And asr is just arguing that these tests have no need.
How do they now that these tests are not needing without doing them?
Experience ?
Also how do they know if a dac has a failure in design without doing them?
It's not scientific at all.
Who says tests weren't done ?
One could decide not to do additional tests when it is not deemed necessary based on preliminary tests that suggest there may be something wrong.
It seems they are fooled by the marketing of classic tests that don't show all that can be showed.
A: all people can be fooled. Technical guys as well.
B: I think Amir tests equipment as to verify claims made by manufacturers and does not rely, nor is he fooled, by marketing at all. Consumers are.
Tests done show only the better aspects of the dacs.
To be revealing they should show worst aspects of the dacs.
Revealing the worst aspects is exactly what is being done here. Those plots are the ones that are shown.
For this reasons i don't trust much on asr.
Then move on to other sites that you do trust.
Their measurements show only one aspect of dacs where all dacs perform about the same.
the measurements show relevant aspects where all DACs perform the MOST differently and closest to borders of audibility. One can also test and publish aspects of DAC's where the results are far outside the audible range. Most DAC's perform more similar within and well outside the audible range at higher bitrates/depths so would do the opposite of what you want.
44.1 and DSDx1 are the worst case scenario and show the most problems.
That's why asr believe they all sound the same.
That's just misinformation.
Where does ASR believe all DACs sound the same ?
Where is the misinformation about this ?
DACs are measured far, far, far beyond any audible limits.
This means that certain measured differences are well below audible thresholds and thus won't change the sound.
Audible limits are often debated. Some place them at higher or lower limits than others.
THAT part can be discussed and debated. There are listening tests for it to act as personal evidence.
Not everyone has had similar training and hearing abilities for certain aspects or equipment that is capable of reproducing all there is.
The problem is here HOW the 'listening tests' are performed. It is infinitely more easy to f' this up than it is to test in a proper way.
THAT is the root cause of the discrepancy between measurements and listening tests.
Along with experience/abilities and understanding of measurements.
Those days people should be careful on what they read. Lot of things are made for money not for what they should be.
Indeed people should be wary of what they read.
Especially websites that merely 'test' sighted and with obvious motivations for endorsing a product.
All products are made to make a profit for the owner(s) of the company that makes/sells what they sell.
When only the best of the best would be available everything will be crazy expensive and durable and have looks that please every one on the planet.
Also It would have to support all formats. There would be little to choose.
This is why ASR tests the cheaper stuff. To see which products perform well (enough) for the price. It can even be so that a $9.- product measures good enough yet shows measurements that are clearly inferior to more expensive items but can also be the other way around.[/QUOTE]