Yes, the noise floor was about 17 bits (as he said (its a litte more now)) but his own plots show a resolution better than -120dB
Thanks for the response, I think we can leave it at that then. Thank you for taking the time to respond to my questions. Itâs not everyday that you can ask questions of a products designers and get responses! I hope potential PS Audio customers such as myself find this information as useful as I did.
Hey Ted, letâs get that firmware done instead of wasting your time on replying this non-sense. 
4 Likes
I do not post on ASR.
1 Like
No worries, I am sure Amir will point you in the right direction on your next DAC.
1 Like
I have been dealing with electronics professionally for 40 years, currently wireless microphones and I am equipped with an RF generator, spectrum analyzer, oscilloscope and many other instruments. Years ago, I repaired audio equipment and also had a distortion meter, but I got rid of it a long time ago, because I found out that harmonic distortion has little to do with sound quality.
Sorry for my bad english.
3 Likes
The not so veiled dig at the DirectStream DAC is noted. You really need to buy that Topping DAC that measures so well for Amir. Of course, he canât hear the difference between a decent DAC and a great DAC so advice taker beware. You deserve each other.
No one at PSA (or anywhere else) needs to defend anything Armoire has written in one of his hit piece measurement spooge fests he calls a âreviewâ. This isnât a high school science contest. You say âI fail to see how the objective performance determined from all 3 different sources justify the marketing for the DS which I stated above, which is also described as being able to âfinally reveal all the music and subtle low-level details buried deep within its core.â What measurement would satisfy you exactly? Maybe you should consider taking a step back and putting a system together first and listen for a while instead lobbing passive aggressive insults. Starting with the sprout wouldnât be a bad idea or perhaps one of Armoireâs latest fav POS $80 DACâs. Iâm sure itâll measure great. Better yet just buy some expensive test equipment then you too can trash the hard working designers, engineers and artists that put their heart and soul into designing these products.
At this point sufficient insults have been exchanged.
Please everyone move on to again addressing the topic.
4 Likes
Iâm actually a little confused: the (pretty standard) plots show the correct output for a -120dB sine. How can someone not see that that shows the DS resolving to better than CD resolution? I can understand that not everyone knows the formula for how FFTs show signals and noise differently, but if they donât understand that then how did they think they understand any FFT plot?
8 Likes
I suspect this is the crux of the issue.
Most audiophiles pick up bits and pieces of audio science as we read reviews, etc. But there are a lot of holes in our understanding and we additionally make lots of assumptions.
This also makes it easy to mislead many of us.
3 Likes
An annoying example from my past was the claim that SACDs had a rising noise floor in the audio band based on JAâs plots of the noise floor of a 1/3 octave filtered sweep across the frequency band. They didnât seem to realize that every point was the average of the values over 1/3 octave and hence the value displayed at, say, 20k had some of the (higher) value at 25k mixed inâŚ
An excellent example and illustrative of why one must carefully read what you are being shown.
Yep, with FFTs you can make the noise floor appear as low as you want (if itâs white.) But the signals stay at the correct levels, so in all of my plots the apparent noise floor is irrelevant (they are all artificially low.) What matters is the levels of the signals compared to whatâs going into the DAC. For someone familiar with FFT displays saying that the input is a -120dB sine and seeing a -120dB sine says a lot about resolution and, with a little more context, linearityâŚ
2 Likes
Ted,
I donât pretend to know a lot about these measuring devices or what they show. Itâs fairly clear that this individual doesnât either. The difference is that I donât pretend to be an expert and tell a real expert what his data shows. I do know that spec sheets donât tell you how a DAC is going to sound. I also know that your DirectStream DAC sounds incredibly good, whatever the measurements.
Amir thinks he can measure a DAC and tell you how good it is. That may work with delta sigma DAC chips. I donât know. But, I do know it does not work with R2R DACs and your DirectStream DAC.
Here is a simple example that better illustrates my point. Below are some sine waves output from two different DACS, using the same digital input signal (1khz, 24 bit at various levels), and captured using the same device. All measurements were made using a 22.4 kHz bandwidth/filter, meaning that anything in the output is audible to humans in terms of frequency. As you can see, one is able to accurately reproduce the digital signal at -90 dB, the other is not.
One of these is the $6000 DS, the other is a $1400 DAC. I think anyone looking at the two graphs above can determine which DAC is better at converting a digital signal to analog, but Iâll let you all decide which DAC is which. Keep in mind this test is giving the DS its best chance to perform well, as Ted stated earlier that the DS performance better with lower level, higher frequency signals compared to a full level 50 hz signal.
Being able to see the expected signal in the output via FFT does not mean you are able to accurately reproduce that signal. It is the ratio of the the expected signal to all the other information (added via the imperfect conversion from digital to analog) in the output that determines how accurate the conversion from digital to analog is.This is why SINAD is useful when evaluating a DACâs performance. In the first example graph at -90 dB, there is still a 1khz sine wave at -90 dB in the output, however you cannot tell because there are many other tones of similar magnitudes and different frequencies that mask the expected tone, meaning anyone listening to the output cannot distinguish the expected tone from the additional tones.
I suspect the output from Tedâs FFT graph showing the -120 dB signal would like similar to the -90 dB chart above. Ted, could you show us the output in the time/voltage domain so we can see what the output looks like compared to the example shown above? I think it would be helpful to understand the effect of distortion and noise on a signal.
You are creatively missing the point⌠Thereâs noise at the output of the DS, if you donât want to see it because you donât understand FFTs or donât trust your ears add a reaonably steep filter at, say 25k and look again.
I didnât read most of your post, Itâs a waste of my time. If you ask some polite questions (one at a time) I may answer, but if you donât I wonât.
3 Likes
You are one easy going person to withstand all the nonsense directed your way. If It was me I wouldnât feed them. Regardless of the vagaries I stand-inline for the TSS without hearing it. Your skill designing the DS is testament enough for me.
7 Likes
Yep, his agenda was obvious with his first post, I figure that if he makes a big enough fool of himself it ought to be obvious to everyone here how ASR works. (Of course the many great posts above describing ASRâs character probably have already done that.)
7 Likes
Hi Ted,
So, from the charts you posted, thereâs noise at lower frequencies, and a few spikes at higher frequencies, WHAT causes those? and is there anything that could lower that noise down? or it does NOT matter in the grand scheme of the DAC?