Netlist said:
Hugo, i did that test, but there was something odd...
The rmaa-measuring shows that the signal did not get degraded. I ran rmaa on its own supplied .wav and from cd (with this file burnt) through spdif.
But subtracting the 2 waves (the original and the captured) did not result in exactly zero. Some strange 2-bit noise was left, repeating every 2.7 seconds, having a gap of silence of ~0.7 secs. This noise was in no relation to the original signal.
I am paranoid enough to suspect that some software/hardware added a watermark...
Mike
AAAArgh !
Found the source to the odd noise... I did not use exactly the same wave-file rmaa puts out and the burned one. Rmaa adds white noise to each test pattern to "improve" fft ? This gives different wave file each time it gets generated.
Well, gladly i found the original wavefile that was burnt to cd.
The result: perfect digital zero.
So: Wave file generated -> burnt to CD -> played from player -> recorded via spdif. The result is a perfect digital copy, subtracting the recording with the original wave file gives complete zero, not a single bit lost !
So much about the voodoo, that error correction replaces samples, burning CDs changes the data and so on, bla bla bla. (burnt onto cheapest CDR available, one year ago)
BTW, my spdif coax-cable was a 2meter plain RCA-cable, nothing special/high quality.
The only real thing with the digital-voodoo is jitter, no data is somewhere lost/changed.
Hardware used for the test:
- Burner: LG GSA-4040B
- Player: Pioneer DV-525
- Soundcard: M-Audio Delta 24B96
- Cable: Standard RCA-cable audiophiles would never touch.
All mainstream, except the soundcard maybe.
The result: ALL data gets to the DAC completely unchanged/undegenerated, maybe just not in the exact nanosecond it should.
Mike
Yes, the soundcard is a nice one. I especially bought it for measurings and absolutely wanted spdif coax in/out. Measured thd of this soundcard is ~0.001%.
I hope this test can end many of the debates...
The results are plain and measured effects, no data gets lost or changed in the digital chain. I'm not sure whats left to argue about ? Identical data is identical data, even without high performance hardware.
Mike
I hope this test can end many of the debates...
The results are plain and measured effects, no data gets lost or changed in the digital chain. I'm not sure whats left to argue about ? Identical data is identical data, even without high performance hardware.
Mike
That's exactly what i wanted to justify, there is no sample error, if there is an audible difference it is not related to data errors, jitter is the only thing that seems to be responsible or this.
I intentionly used mainstream hardware, if this is able to deliver all data without a single error, high end hardware really should have no problem with that.
Jitter has null effect in this test, as only the datas transfered are compared. (Just like downloading from the internet)
Error correction typically takes place when bits are decoded from disc, but obviously no problem occured during my test. (with a cheap cdr)
Mike
I intentionly used mainstream hardware, if this is able to deliver all data without a single error, high end hardware really should have no problem with that.
Jitter has null effect in this test, as only the datas transfered are compared. (Just like downloading from the internet)
Error correction typically takes place when bits are decoded from disc, but obviously no problem occured during my test. (with a cheap cdr)
Mike
soongsc said:
There will surely be burning or reading errors, often 1000's on a CD. The point is that the error correction correct all this 100% (unless uncorrectable errors cause error concealment but thats really rare these days). The fact that Hugo saw exact copies is a testimony how good the CD system is even with many, many errors occuring.
Jan Didden
I_Forgot said:
I don't think that exists. You may have the same piece of music on LP and CD, but I would bet that the mastering and mixing and recording would be done different. In fact, it is my believe (which I cannot prove, but Hugo's and MikeB's findings support it) that the difference between LP and CD is SOLELY due to mastering/mixing etc differences and NOT due to any problems with either medium.
Jan Didden
Jan,
Thank you very much.
After further thinking - and having had a nice meal - the signal fed trough the whole chain as Mike did was only a 1Khz sine wave. Would the complexity of music still give the same results?
Because I think we (I only got the ball rolling) made a firm statement re. the complete digital audio world, more precise the cd player
Except for the jitter (pardon my ears but I still need to hear what jitter is at my age) there would be no difference between an average good cd player and a top high end model.
Do I miss something?
/Hugo
Thank you very much.
After further thinking - and having had a nice meal - the signal fed trough the whole chain as Mike did was only a 1Khz sine wave. Would the complexity of music still give the same results?
Because I think we (I only got the ball rolling) made a firm statement re. the complete digital audio world, more precise the cd player
Except for the jitter (pardon my ears but I still need to hear what jitter is at my age) there would be no difference between an average good cd player and a top high end model.
Do I miss something?
/Hugo
If the data could not match up, the reast of the computer world is coing to be in trouble because all software are distributed by digital means.
I think the data will always match up once it gets digital, but when it gets burned, will the data location on the CD cause the player to read in a matter that will cause different modulation in the PSU is the main question, and probably the main cause of sounding different. Different burning speeds will cause different effects.
I do wonder what burning speed will be the best match most CD and DVD type players.
I think the data will always match up once it gets digital, but when it gets burned, will the data location on the CD cause the player to read in a matter that will cause different modulation in the PSU is the main question, and probably the main cause of sounding different. Different burning speeds will cause different effects.
I do wonder what burning speed will be the best match most CD and DVD type players.
janneman said:
Jan, SPDIF is a digital signal, only containing ones and zeros with varying length. The problem with jitter is, that the clock sync supplied with the spdif get's unaccurate, permanently beeing randomly a few nanoseconds too early or too late. If this clock is used to feed the DAC, you get some kind of FM-distortion. This causes ugly distortions for higher freqs, and these seem to be large enough to be audible.
My test only proved that no data gets lost when transfered via spdif. Jitter had no chance to take effect, it does not matter if data is received at uncorrect timings, when a spdif word was received, it will be dumped sequentially to harddisk/memory, loosing all jitter information.
In theory, reclocking the signal removes the jitter, but there are hot debates that this introduces new problems.
Hugo, the rmaa test signal also contains a part with complex noise, used for the freq response test. But, as the transfer is digital, the content of the digitized signal is quite unimportant.
You're right, an average good CD-player and a high end player will not make any difference for the data delivered. You can get differences with damaged CDs, depending on the error correction capability of the drive.
I've burnt the CD with 8x speed, the lowest speed my burner allows. Burning speed is relevant, and can be easily proved. Simply burn a DTS or DD encoded audio cd with 48x, and try to play that !
For DTS/DD any lost bit in the stream is a disaster, resulting in dropped audioframes (replaced with silence) or decoding errors giving very audible glitches.
About ripping speed, if the CD-drive/burner is of reasonable quality it seems to make no difference. I compared a track ripped at 1x and at 24x speed, no difference. Again, jitter can't take place here as only data is copied. I observed that if the drive gets into problems decoding the audiotrack, it does reduce rippingspeed by itself.
AFAIK, there are 3 problems with reproduction of digital sound:
- Quality of DAC
- Aliasing (because of too low samplerate)
- Jitter (Increased by bad/long spdif cable)
About the aliasing, it is visible with a scope. For example, play a 10khz sinewave through the CD-player and look at the signal with a scope. At first look, you see a nice sinewave (impressive regarding the fact that it was generated from ~4 samples !) If you look closer, you see some edges walking through the sinewave, even with 8x oversampling and brickwall outputfilter.
You'll need a good old analog CRO to see this.
There is one way out of jitter, it's a bidirectional digital connection (AES as i remember). The clock signal is generated by the dac and supplied by a 2nd cable back to the player. This way the player sends data "on demand", eliminating any jitter caused by cable/drive. The only jitter left is the accuracy of the clock from the dac.
Mike
MikeB said:
Mike, I know that ...
I had the impression that using the S/PDIF you ran the data through a DAC-ADC combo.
I call S/PDIF an essentially analog signal since it is used that way: the zero crossings determine the period, while in a digital signal you would for instance count the pulses or the byte value or whatever. A bit tongue in cheek maybe, but some people call ANY signal essentially analog.
MikeB said:
Agree.
Jan Didden
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.