Why NON-OS?

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Kuei Yang Wang said:
As it so happens, I am failiar with a number of rather justly famous of pieces of gear using the THD1541 and which are pretty much UNIVERSALLY aknowleged to sound great (Marantz CD-7, Marantz CD/DA-12, Marantz CD/DA94, several earlier model Naim CD-Players). Other items exist that are equally universally acclaimed and use the TDA1541 including the Sugden CS-Player and the Zanden DAC.

This thread was about the non os lark and all your examples bar the Zanden include the digital filter. All you've shown is remove the filter and it all goes pear shaped.
 
On topic

The effect of removing the digital filter is most of the times tighter bass.
This indicates one thing to me: lower jitter.
Of course, you have to deal with some issues.
Unavoidably, high frequencies start to roll-off, beginning at the midband.
It's clearly noticeable, and that's why some people don't like it.
I don't like it too, it sounds too smooth and "dead" to my taste.
Not what I call neutral.
Now... the important point is: deal with that treble roll-off and then use a simple 1st order lowpass filter and the sound can be shocking.:bawling:
I love the TDA1541A.:hbeat:
 
A bit off-topic, but I think it's the best place to ask this:
The TDA1541s are popular for DIYers, but is it really the best DAC chip to use? What about PCM63 or AD1865? Is a Kwak 7 clock good enough for using these DACs with synchronise-able sources (like some CDP and computers)?
 
Re: On topic

carlosfm said:
The effect of removing the digital filter is most of the times tighter bass.
This indicates one thing to me: lower jitter.
Of course, you have to deal with some issues.
Unavoidably, high frequencies start to roll-off, beginning at the midband.
It's clearly noticeable, and that's why some people don't like it.
I don't like it too, it sounds too smooth and "dead" to my taste.
Not what I call neutral.
Now... the important point is: deal with that treble roll-off and then use a simple 1st order lowpass filter and the sound can be shocking.:bawling:
I love the TDA1541A.:hbeat:

Hi Carlosfm, A more intelligent approach would be :
How to deal with the aliasing noise problem and LP-filtering.
The TDA1543 does not sound "dead"; the TDA1541 and the TDA1545A do, in my opinion.
:cool:
 
Konnichiwa,

erozsolt said:
The TDA1541s are popular for DIYers, but is it really the best DAC chip to use?

Yes, absolutely, if you value natural, realistic sound reproduction, not artifice.

erozsolt said:
What about PCM63 or AD1865?

I did build my first Non-Os DAC ages ago from kusonoki San's first schematic (4 X TDA1543). It held much promise but had also severel issues, sonically. I later experiemented with PCM56, PCM63, PCM1702 and TDA1541 and I also had a chance to work on severe AN DAC's which use AD1865 Non-Os.

I found the TDA1541 best by far with PCM56 & PCM63 running close, the PCM1702 was a little behind and the AD1865 was IIRC behind that, but in someone elses implementation, so it MAY have been the implementation and not the DAC.

My recommendation would be to build a "universal" DIO Board which can provide the needed data format for any of the above and to make a few different DAC Modules. Any of the chip's mentioned are very cheap to buy in the general context of things and if you build a DAC that can a generic DAC module you can try them all.

Sayonara
 
I found the TDA1541 best by far with PCM56 & PCM63 running close, the PCM1702 was a little behind and the AD1865 was IIRC behind that, but in someone elses implementation, so it MAY have been the implementation and not the DAC.

interesting. I breadboarded DACs with TDA1543, TDA1541A, AD1865 so far, and like the Analog Devices chip best. There must be some more magic how to do the implementation for TDA1541.

For some reason i do not understand the interest in this http://www.diyaudio.com/forums/showthread.php?s=&threadid=31780&perpage=10&highlight=&pagenumber=1 seems to have gone.
 
Re: Re: Re: On topic

rfbrw said:

Neither. That takes you into the realm of musical effects.

Not really,

Have to dig up my books from 'digital signal processing' classes i had 10 years ago. Must be possible to put this into formulas.

Anyway, i dont hear any effects, other than the ones on the cd (if there are). But what i hear is the same as with non-os, just the high frequencies are there! Since it is "hot" switchable in the dac, it is easy to hear the difference. Even did measurements, which give the same result (obviously).

Some day soon, i'm going to start on a new version of this thing.
Together with better implementation of the 1541 setup, based on the info of HtP. Havent seen ANY implementation which is 'correct', specially around the powersupplies.:whazzat:

But for now, i'm quite happy:D
 
Re: Re: On topic

QSerraTico_Tico said:
Hi Carlosfm, A more intelligent approach would be :
How to deal with the aliasing noise problem and LP-filtering.

If you hear that.
You can measure, of course, and it's shocking.
But can you hear it?

QSerraTico_Tico said:
The TDA1543 does not sound "dead"; the TDA1541 and the TDA1545A do, in my opinion.:cool:

The TDA1543 has a more "upfront" presentation and on a NOS implementation it doesn't sound so "dead", but it's still not correct.
The implementation of the 1543 is different to the 1541.

Anyway, although it can sound very good, the 1543 sounds ragged and somewhat rude compared to the 1541.
There's a big difference in distortion between these two dacs.

Jocko Homo said:
Right. Jitter has to be eliminated. But to some, it is just an excuse for guys like Guido to sell clocks and for me to complain about reflections on SPDIF setups.

Back to the comedy at hand............

Jocko

My post was all about direct I2S interface, not SPDIF.:dead:
 
Re: Re: Re: Re: On topic

guido said:


Not really,

Have to dig up my books from 'digital signal processing' classes i had 10 years ago. Must be possible to put this into formulas.

Anyway, i dont hear any effects, other than the ones on the cd (if there are). But what i hear is the same as with non-os, just the high frequencies are there! Since it is "hot" switchable in the dac, it is easy to hear the difference. Even did measurements, which give the same result (obviously).

Some day soon, i'm going to start on a new version of this thing.
Together with better implementation of the 1541 setup, based on the info of HtP. Havent seen ANY implementation which is 'correct', specially around the powersupplies.:whazzat:

But for now, i'm quite happy:D

There is nothing equivocal about it. When you delay in units of one L/R frame or 64 BCK cycles to use your parlance, you are into musical effects. You certainly are not performing linear interpolation as you are creating no new samples.What you hear and whether or not you like it is neither here nor there.
 
Re: Re: Re: Re: Re: On topic

rfbrw said:


There is nothing equivocal about it. When you delay in units of one L/R frame or 64 BCK cycles to use your parlance, you are into musical effects. You certainly are not performing linear interpolation as you are creating no new samples.What you hear and whether or not you like it is neither here nor there.

no lin interpolation, sounds great to me! Must be an even better form of interpolation then :D

I know you dont really care much of other people's opinion, to bad for you. I don't know what i'm exactly listening to, but the 'musical effects' sound fine. Dismissing something because you don't understand what is going on is your loss.

Dont know why (yet?), but it works (and better than non-os).
 
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.