Bernhard,
In the post #27 you have two graphs showing the one DAC chip with two different clocks (and output stage and what not) and, as you have pointed out in that post, you have similar graphs. Shouldn’t you, if investigating clocks, measure jitter and not harmonic distortion?
Fellow member Ulas has nice small program which shows relation between jitter and THD+N. It is true it will not show the situation with HD only, but can give you idea about importance of jitter at lower levels where only a few bits are used. So, I really do not expect you will see anything watching harmonically related distortion.
When the harmonic distortion of -60dBFS signal has components more than 50dB below signal level, I do not call it/them strong distortion/components. But, whatever is the way we call it, what could be the reason for non-o/s to have the spectrum of harmonic distortion you claim to have? Btw, I can witness that at 0dBFS level (TDA1541A non-o/s, opamp I/V) it is possible to keep all harmonic components 100dB below the signal. It also worth to mention that your non-o/s with TDA1541A has dominant even order components but your non-o/s with TDA1541 has dominant odd order.
Pedja
In the post #27 you have two graphs showing the one DAC chip with two different clocks (and output stage and what not) and, as you have pointed out in that post, you have similar graphs. Shouldn’t you, if investigating clocks, measure jitter and not harmonic distortion?
Fellow member Ulas has nice small program which shows relation between jitter and THD+N. It is true it will not show the situation with HD only, but can give you idea about importance of jitter at lower levels where only a few bits are used. So, I really do not expect you will see anything watching harmonically related distortion.
When the harmonic distortion of -60dBFS signal has components more than 50dB below signal level, I do not call it/them strong distortion/components. But, whatever is the way we call it, what could be the reason for non-o/s to have the spectrum of harmonic distortion you claim to have? Btw, I can witness that at 0dBFS level (TDA1541A non-o/s, opamp I/V) it is possible to keep all harmonic components 100dB below the signal. It also worth to mention that your non-o/s with TDA1541A has dominant even order components but your non-o/s with TDA1541 has dominant odd order.
Pedja
Pedja said:
If different clocks with different jitter have same distortion, I will not care of that jitter.
That means if I measure the same with crystal on board, canned oscillator, Kwak clock etc. , I will put back the crystal on the board.
Pedja said:
Don't you think -60dB signal level is low enough ?
Pedja said:
They are not strong but measurable, even with my analyzer.
And I hear clearly the difference between stronger even or stronger odd orders.
I have no idea about the mechanisms that affect distribution and level of harmonics.
There is no doubt about what I measure, because it is the same in three different players.
Pedja said:
So what you wanna say ?
That we don't need S1 or S2 chip ?
And no low jitter clock ?
Because distortion is already so low that it is not disturbing ? 😀
It's the reason why I use -60dB signal as does Philips.
Somebody once pointed out in a forum that the behavior of an amp at low level is important, how the tone ends at low level, smooth or harsh.
ClassB ends harsh and classA ends smooth. Something like that.
If you listen to classic music at high level and there is a quiet instrument, you will hear the difference. 😀
Pedja said:
Bernhard said:
That shows clearly that 1541 and 1541A are different.
Hey, I think I'm the first who found out about the relations between different combinations of chips / modes, and distortion spectrum. 😀
What about 8 x oversampling ???
Then, if I may save you the effort, just leave the crystal.Bernhard said:
Sorry, I do not understand.
I have checked that and I must correct myself, in fact it was the TDA1541A/S1 in that particular measurement, but nevertheless, the answer is – if harmonic distortion performance that matter, no, we don’t. Note that with everyday audio stuff of today you can easily achieve such specs. However, there is much more about the good sound than it is the harmonic distortion.
Pedja
Thanks Pedja.😎
BTW Bernhard, just listen to the music with your ears, not with your scope.
Otherwise, why would a well implemented NOS Dac sound better than an oversampled Dac?🙄
And the TENT clock is very good and cheap, why bother with anything else?
BTW Bernhard, just listen to the music with your ears, not with your scope.
Otherwise, why would a well implemented NOS Dac sound better than an oversampled Dac?🙄
And the TENT clock is very good and cheap, why bother with anything else?
carlosfm said:
Please...
I listen to the music and I use the analyzer to find out why it sounds how it sounds.
Again: Non os 1541A sounds subjective better because of "tube like" dominant even order distortion.
I find it very questionable to ignore measurements.
If you just listen with your ears, you can only judge sound subjectively and there is no other way for you to improve sound than trial and error.
If you also do measurements and once you found the relationship between results and sound, you know the direction to go.
till said:
LOL*
No, it sounds different due to the richness of even order harmonics, and I like it too.
But odd orders are like with os and even orders are stronger.
This must hide some information in the music.
OS sounds cleaner.
I am not decided yet what I prefer.
i measured averaged white noise from CD played with the CD PRO build in DAC against the balanced nos AD1865 DAC, throug speakeres with mic. I got -2 dB or so at 20kHz with the nos dac ? Else response was same.
till said:
You don't believe that, do you ? Also ready built non os DAC's sound very good. The ready built 47 Labs DAC is the one that started the TDA1543 non os hype.
Besides that I have built things that I did not like when I heard them. That's normal, it really becomes strange when you think everything you've built sounds cool.
Bernhard said:
I have the impression you trust your instruments more than your ears. You also speak negative about the chips you seem to like, collect and work with for a hobby. Why do something when you don't like it ?
Never having heard or measured TDA1543 non os you DAC's you have an opinion about a DAC a lot of people prefer. The only thing you come up with is: "it distorts like the datasheet says". "Economy chip" and so on.
If you really want to do as you say you would not stick to the same measurements over and over again.
Speaking about questionable statements....
till said:
I don't think it is the case. I've built so many things that just didn't sound right (to me), and I made it diy.
sorry, i´m not able to answer that question with yes or no. There may be good ways to build os and nos DACs, and there may be bad ways.
But i´m absolutely sure that statement is very often true, if i read all those people who have 100% better sounding gear after changing a resistor, cable, or laying a stone under the CDP...
Its easy to fool yourself, and i´m with Bernhard what concerns measurements.
You both won't ever find satisfaction I think except for chatting about your negative analyzer results. Chip A measures worse than chip B.
Why exaggerate if you receive counter arguments ??
Why exaggerate if you receive counter arguments ??
do i?
I see measurements as a tool to a) see if it really works as it should, b) understand a little more.
I don´t say better specs sound better. I experience a measuring worse response of my speakers sound better than the better measured. I experience an amplifier with "bad" THD (Zen lite) sound better than a gainclone with fantastic datasheet specs... I do not "belive" in that measurements (maybe we all measure the wrong parameters...) but i take it as a helpfull tool.
jean-paul said:
Do you really read and understand my postings ???
You just repeat in a circle things like that you don't believe the measurements etc.
I will build, listen and measure 1543 non os dac, when I have time.
There is nothing to trust in instruments, they show what's up.
What do I speak negative about my chips ?
That I don't like some combinations ?
And ? It's the truth.
It is common knowledge that dominating odds sound bad.
I hear it and the instrument shows it and everything fits together.
But you doubt it.
Concerning 1543:
If the chip distorts in OS it is very unlikely that it doesn't in non os.
I just gess that from my experience.
The 1543 non os dac seems to be the mother of all tube sound dacs.
If lots of people like it, ok... this is subjective.
For my personal taste it is too much.
Bernhard said:
Sorry but I don't repeat in circles that I don't believe measurements. I really read and try to understand your postings, how irregular they may be sometimes. I do measure also but I don't have an analyzer myself and I don't feel the urge to show every result/image on the forum. I just believe that you trust to much in *your* *interpretation* of your measurements which is something different.
I also don't think that you are able to correlate your measurements with what you hear on a high rate level.
But I won't let you down: I will repeat in circles that you should read some more on the subject like jitter besides doing your measurements on THD and opening canned oscillators.
I give up.
jean-paul said:
Don't worry, I was satisfied for 20 years with my CD304mkII and still I am.
All that DAC stuff is not so really important for me.
But nevertheless I try to get the best results and I use whatever tool I can get my hands on, even an analyzer.
It is very simple: If chip A measures worse, it sounds worse and I sort it out, that's it. No mercy.
Use the search function and look for even & odd & distortion in the solid state forum.
I did it and it is clear that odd orders sound bad.
Nobody likes the odds.
And most like the even.
It confirms my listening experience with chips that show this kinds of harmonic distribution.
And: I first listened and did not like it and then measured and found out about odd orders...
Same with the beauty of non os ... *
So what do you see negative in it ???
There are good leftovers from the "I don't like this" and "that measures bad":
1541 + 7220A Os
1541A + 1541A Non Os
* It's like a girl with lots of make up 😀
Beautiful but not the truth.
- Status
- Not open for further replies.