Have you tried the zero oversampling mod? It's fantastic.

[Dmannnnn]

I mean seriously, why would the engineers put in the extra cost for something that makes it worse.

Isn't this the basis for all mass-market audio gear? [:^)]

[Dmannnnn]

Regardless of what the data says, this mod improved Lorenzo and Erin's systems. Period. Unless they are lying to us.

If I have a choice between a system with less distortion or a system that sounds better, I will choose the latter every time. Is there really someone here who would make their system less enjoyable to improve the specs? As far as audio is concerned, I say use the science as a guide, but let your ears make the final decision.

[Batmans Robin]

I would be afraid to try this if I had one, which I don't.

There is one thing I would like to take issue with. Supposed "jaw-dropping" changes are usually more of a "hmmm" experience to me. Like, yes, there may be something different but no big deal.

[DrWho]

I was just thinking that there is actually less phase shift when oversampling...in fact, less of everything bad

 

Doc, are the digital filters in a CD player Infinite Impulse Response? If so, why not use Finite Impulse Response filters? Is anyone, in fact, doing FIRs in a CD player? It seems like phase shift would be greatly reduced with thier use. 

Just about all the DACs I've messed with are FIR, but your phase response is determined by your windowing. It kinds dissapoints me to see so many elliptical filters being used...injecting all sorts of ripple into the passband. They do it to hit insanely low noise floors. I'd rather raise the noise floor and have less passband junk. You might see as much ripple as 0.1dB in the HF...I can't help but think that creates the impression of harsh highs. Wolfson Micro has some ADC and DAC solutions that appear to have a 0dB passband ripple....just gotta make sure your analog filters are up to snuff. The same chip offers the typical elliptical filter if you want to be lazy about your analog section.

[DrWho]

Regardless of what the data says, this mod improved Lorenzo and Erin's systems. Period.  Unless they are lying to us.  

If I have a choice between a system with less distortion or a system that sounds better, I will choose the latter every time.  Is there really someone here who would make their system less enjoyable to improve the specs?  As far as audio is concerned, I say use the science as a guide, but let your ears make the final decision. 

2 Questions: What makes a spec good or bad? And would you paint a moustache on the Mona Lisa? I would suggest that by definition, a good spec is defined by that fact that it sounds good - so to suggest that a good spec sounds bad is to say that you are incorrectly correlating the spec to what you hear. I am also of the opinion that we need to preserve the work of the artist...so do you see your system as a reproducer of music or a producer? There's nothing wrong with the latter, but it's a totally different approach.

[Erinp]

My understanding of a digital audio recording is that the analog audio goes through an anti aliasing filter prior to going to the ADC and then stored on a recording medium.

So when playing back the recording it is not necessary to filter the output because the frequencies that cause aliasing have already been removed.

If they had not been removed proir to recording the audio then the alias frequencies would be present on the recording and then hence present during playback even if there was a filter.

Correct me if I'm wrong here but output filters are not even need to filter the 44.1Khz sampling frequency from the output because DACs like ADCs use a sample and hold circuit so the 44.1Khz has already been smoothed so that the highest frequency possibly present on the output is 22Khz (if present on the recording).

Class D amps on the other hand do need LC filtering because they will have a certain amount of Volts on the output @ high frequency because they use PWM and a product of the mark space ratio is that there will be volts present on the output during the "mark" albeit very briefly, hence the filter needs to be tuned to the frequency that the amp oscillates at to make sure that there is 0Volts DC on the output as well as no HF energy to blow the HF drivers.

When a DAC is reading the digital data and if there is data representing silence going into the DAC then the DAC outputs 0 Volts or 0 current.

[Don Richard]

Anti-aliasing, somewhat simplified:

http://www.maxim-ic.com/appnotes.cfm/an_pk/928

[PrestonTom]

Mike, Did I hear you correctly? Are you worrying about 0.1 db ripple in the HF band?

Goodness gracious,

-Tom

[DrWho]

Mike, Did I hear you correctly? Are you worrying about 0.1 db ripple in the HF band?

Goodness gracious,

-Tom

Well it's certainly not the magnitude of the ripple that I'm worried about. It's the ghosting you get in the impulse response - it shows up to the left and right of the main impulse. There's an IEEE or AES article on it - I don't remember which one, but I read it not too long ago. I can probably find it again at work if you want to read up on it.

[DrWho]

My understanding of a digital audio recording is that the analog audio goes through an anti aliasing filter prior to going to the ADC and then stored on a recording medium.

So when playing back the recording it is not necessary to filter the output because the frequencies that cause aliasing have already been removed.

If they had not been removed proir to recording the audio then the alias frequencies would be present on the recording and then hence present during playback even if there was a filter.

Correct me if I'm wrong here but output filters are not even need to filter the 44.1Khz sampling frequency from the output because DACs like ADCs use a sample and hold circuit so the 44.1Khz has already been smoothed so that the highest frequency possibly present on the output is 22Khz (if present on the recording).

Class D amps on the other hand do need LC filtering because they will have a certain amount of Volts on the output @ high frequency because they use PWM and a product of the mark space ratio is that there will be volts present on the output during the "mark" albeit very briefly, hence the filter needs to be tuned to the frequency that the amp oscillates at to make sure that there is 0Volts DC on the output as well as no HF energy to blow the HF drivers.

When a DAC is reading the digital data and if there is data representing silence going into the DAC then the DAC outputs 0 Volts or 0 current.

Class D amplifiers and DAC outputs do not contain the same frequency content at the output.

Without a lowpass signal at the output of a DAC, you do not end up with the original signal. Just run through the math if you think I'm crazy []

[lorcoll]

Enrip,

I have now about 14 Cd Players in NOS mode and all of them are simply fabulous (my personal favourite is the Philips CD-960 in zero oversampling - nothing can reach it).

I know that instrumental measurements are not very good but the sound is absolutely first class and I hope that people would speak about the Zero oversampling only after they have heard it.

Zero oversampling and Klipsch speakers (I have two pairs of Khorn, Cornwall, Forte II and Heresy I) are a sublime match.

Many thanks again.

[cuneyt]

Hi,

I have modified my Marantz CD80 to non oversampling mode by success. Now it is sounding very good and ear friend to me. My remain question in my mind, after bypassing of SAA7220 dig. filter (as described in lampizator and here), it can not anymore deliver the clock frequency 11.2896mHz to the TDA1541A chip. In this case, what if i connect the clock out of SAA7310 chip directly to TDA1541A leg no:2 ??? I have already checked the datasheet of SAA7310 receiver and saw that pin no:18 is crystal oscillator output. Or, perhaps trought non os modification, the SAA7310 chip leg 18 is already connected to SAA7220 (and TDA1541A) ??? I saw a crystal very closer to SAA7220 chip in Marantz CD80. It is not functioning after non os, if i am not wrong. Any comment ???

Thank you,

[Erinp]

Cuneyt,

firtsly you must understand the fow of data in the CD player.
Very basicly the data from the laser is sent to the SAA7210.
This is the chip that generates the i2s signals.


When we perform the non oversampling mod. We are bypssing the SAA7220.(which does the oversampling)

We are effectively taking the i2s output from the SAA7210.

The 11.2896Mhz crystal is connected also to the SAA7210 and also used by the rest of the CD player to make the motor spin and data to be transfered correctly etc.

The TDA1541 does not have a direct connection to the 11.2896Mhz crystal.
it uses word clock, data and left right clock information derived from the SAA7210.

Dont touch the crystal or your CD player wont work.

Of course you could upgrade it with a 1ppm clock upgrade kit []

[cuneyt]

Thank you for your explanation. Now, i understood how it works.

Regards,

Cuneyt

[4wuz]

Just did the NOS MOD to a PHILIPS cd380

I also soldered the pin 23 from the SAA7220 to the pin 11 on the SAA7210.

Bought it 5 euro in a flee market... I was going to trash it because of the sound quality...

 

My god, now the sound is absolutely crazy. 🙃

Before it was flat and a little low on details in the high frequencies... really mainstream flat sound with no taste.

Now i can hear every little detail... The sound is incredibly more detailed... What i hear IS ACTUALLY ON THE CD. This is not some false details or artefact...

My Goodness, everybody should try this before everything else.

 

I had already done that mod to a Philips CD104 (actually a Radiola CD1104 which is a french copy of the Philips CD104).

They now sound almost identical.

 

THIS is the way it was meant to be listened, before 4x 8x 16x oversampling went to mess everything.

 

 

[chassell]

18 hours ago, 4wuz said:

Just did the NOS MOD to a PHILIPS cd380

I also soldered...

 

 

Thank you for providing a purpose in life for someone.