Can you hear the differences? TOSLINK / COAXIAL in your system

[bbp]

I test my TDA1541 OS DAC KIT FS:TDA1541 OS DAC KITs with my RMAA system based on E-MU 0404.

But I got some question after the testment

It seems that there's no differences with the two method
COAXIAL must be better than the TOSLINK in theory, for TOSLINK must not leave electrical-light-fibre-light-electrical conversion.

If the differences can not be test ? or the diifferences are too smallllllll to measure by RMAA test

Can you hear the differences ?

Test COAXIAL/24BIT ,,,,16 BIT,,,,TOSLINK/24BIT,,,,16BIT
Frequency response (from 40 Hz to 15 kHz), dB: +0.02, -0.13 +0.02, -0.13 +0.02, -0.13 +0.02, -0.13
Noise level, dB (A): -101.3 -101.1 -101.0 -101.1
Dynamic range, dB (A): 98.8 98.3 98.3 98.3
THD, %: 0.0022 0.0022 0.0021 0.0022
IMD + Noise, %: 0.0048 0.0047 0.0047 0.0047
Stereo crosstalk, dB: -99.4 -98.9 -98.5 -99.5

[dangus]

Jitter may be worse with Toslink, if the DAC doesn't include some jitter-reduction stuff like an ASRC.

[CBRworm]

Other than the noise level and dynamic range, I think you are at the limits of the 0404. If you change it to 96khz or 192khz you may see a small improvement in measurement (my noise level went from -101.4 to -105) and you can measure frequency response farther out.

Those are basically the same numbers I get with decent analog stuff using RMAA and 0404USB.

I am already trying to figure out how to get to the next level.

And no, I can't hear a difference between TOSLINK and SPDIF in my system. I think SPDIF has a length restriction compared to TOSLINK.

[Javin5]

I have a CD-player with both coax and TOS-link outputs and I tried both with a Buffalo 32 DAC. Both the coax and the TOS-link were about 1 m long. Changing between the two would take a few seconds with the setup I used at the time, so that may somewhat limit the comparison (I would have preferred instant switching between the two). Anyway, I could not hear a difference.

Kurt

[Pano]

On most DACs I hear a difference. Subtle, but real. Been that way since the 80s. You can also see the difference in the waveform on an O'scope.

Some circuits are better than others, and can "fix" the problems.

[Pjotr]

You can eventually measure jitter differences with the method of Julian Dunn. A nice primer can be found here: Stereophile: A Case of the Jitters Further reading: http://www.troisi.com/lit/jitter.PDF

What you need is a Dunn test signal, a good low jitter soundcard (i.e. M-Audio Audiophile192) and a good FFT program. ARTA can generate a Dunn test signal as .wav which you can transfer to CD eventually and doing the FFT. The "Audiotester" analysis program has longer FFT capabilities than ARTA, which allows you to measure deeper in the noise.

[PaleRider]

Depends on what coax you`re using. Years ago I stumbeled over the German TV-coax named Vivanco KX-710. This cheap coax makes the cleanest (not-)sounding ic`s & coaxes cI`ve ever heard. This cheap TV-coax even outperformed the AT&T optic reference cable.

[tritosine]

Originally Posted by Pjotr

You can eventually measure jitter differences with the method of Julian Dunn. A nice primer can be found here: Stereophile: A Case of the Jitters Further reading: http://www.troisi.com/lit/jitter.PDF

What you need is a Dunn test signal, a good low jitter soundcard (i.e. M-Audio Audiophile192) and a good FFT program. ARTA can generate a Dunn test signal as .wav which you can transfer to CD eventually and doing the FFT. The "Audiotester" analysis program has longer FFT capabilities than ARTA, which allows you to measure deeper in the noise.

1. Such test signals with 16 or 24bit precision give grossly different results.
2. "good low jitter soundcard" -this isnt one: M-Audio Audiophile192!
3. I just checked ARTA, despite setting 24bit , it gave me 16bit precision! Worthless.


Im going to try 24bit test signal thru the weekend (national holiday of 1956 revolution), and a reference grade external ADC.

[Pjotr]

Good luck! Curious about your results

[tritosine]

You don't have to be, here are some 'preliminary' results.
http://www.msbtech.com/support/JitterPaper.pdf

I also hope to test a 0404usb of a friend btw, and other so called 'pro' audio hw.

" 24 bit jitter test signal should always be used due to the low residual harmonic levels not found in the 16 bit version of this test signals used in years past, as shown in Figure 3. Even if signal subtraction is used other nonlinearities acting on the high level harmonics of the 1/192 fs square wave may be mistaken for jitter artifacts. This is virtually eliminated in the 24 bit version of the test signal since the square wave harmonics are mostly below –160 dbfs which is lower than the noise foor in any currently available ADC. The 24 bit version of this signal is also a more sensitive jitter probe than the 16 bit version because of the larger number of bits changing state. "