Hi Waly,
Data is clocked in off the disc using a PLL locked to the CD's data stream. It is clocked in to an FIFO memory (recorded for all intents and purposes). The data is clocked out of the FIFO memory via the system clock. The minimum amount of memory is determined by whether the memory runs out of data or not. Not enough memory means you run out of sound (noise, music) and hear brief blanks. To much memory creates a delay in the beginning of playback. This can be leveraged to create a large buffer (memory) that renders the player "shock-proof", which is most certainly is not the case. To the user, shock has no effect on the operation of the device. The effect that shock has is identical to a player without the buffer. The large buffer allows the player to hide the problems from the user. A little lie, or a big lie? Hopefully this makes it clear. No great mechanism, just a great big place to hide the problems from you.
-Chris
You are correct in that the data is re-clocked in, error corrected, then sent out as valid data that only suffers from the jitter in that system. It may or may not be bit perfect.
Well, you are correct in that what you hear, if anything, is how the system may have dealt with the errors coming in. That means your play back equipment from whatever the medium is. Jitter does play a role in how severe reading errors are. Jitter ads to the problems, the sum of which can create data errors.
Data is clocked in off the disc using a PLL locked to the CD's data stream. It is clocked in to an FIFO memory (recorded for all intents and purposes). The data is clocked out of the FIFO memory via the system clock. The minimum amount of memory is determined by whether the memory runs out of data or not. Not enough memory means you run out of sound (noise, music) and hear brief blanks. To much memory creates a delay in the beginning of playback. This can be leveraged to create a large buffer (memory) that renders the player "shock-proof", which is most certainly is not the case. To the user, shock has no effect on the operation of the device. The effect that shock has is identical to a player without the buffer. The large buffer allows the player to hide the problems from the user. A little lie, or a big lie? Hopefully this makes it clear. No great mechanism, just a great big place to hide the problems from you.
-Chris
All that's needed is to make the analogue area of the digital replay impervious to amplitude or phase noise originating elsewhere, to make jitter a non-issue. This is achieved by reducing the level of that noise originating elsewhere, or increasing the isolation of the analogue from the noise, or a combination of the two methods - the latter would be the smart way to go. End of story.
Our favorite Toshiba 2SJ74/2SK170 do show some thd improvement with
20 to 50 ohms Source resistance between the complementary parts (and I
do tend to work them, so this is probably worst case). If you like your
sound "undegenerated" or are simply too cheap to add 4 resistors, it
doesn't seem to carry much of a penalty.
If you don't have to kill me to tell me, what are the typical standing currents used?Our favorite Toshiba 2SJ74/2SK170 do show some thd improvement with
20 to 50 ohms Source resistance between the complementary parts (and I
do tend to work them, so this is probably worst case). If you like your
sound "undegenerated" or are simply too cheap to add 4 resistors, it
doesn't seem to carry much of a penalty.
And is LIS actually getting some devices with 30mA Idss in their quasi-170 process? Of course that could be the "6-sigma" limit point in the distribution.
I consider it a good day when I find an over-10mA Idss 2SK170BL (the only kind I have).
Last edited:
And this thread never sleeps.If you take 3 days off this thread, it takes 3 days to catch up.
Our favorite Toshiba 2SJ74/2SK170 do show some thd improvement with
20 to 50 ohms Source resistance between the complementary parts (and I
do tend to work them, so this is probably worst case). If you like your
sound "undegenerated" or are simply too cheap to add 4 resistors, it
doesn't seem to carry much of a penalty.
The 2SK170/2SJ74BL has a Yfs of 25mS/33mS.
By adding 20~50R source resistor, the effective Gm is being halved.
And the Idss will drop (for a 2SJ74V) from something like 16mA to 5mA.
I wonder whether John uses such large values in the BT .....
Patrick
George, 16bit version is made by me from 24bit version, format conversion, with TPD dither. Think about it, what it makes to most silent parts of music. You can also "see" it if you zoom silent part in time and also in Y amplitude axis, to see the samples. 24bits still follow shape of signal, 16bits make a kind of pulse position modulation with harsh step of 1LSB of 16bit resolution. When amplified, it is clearly audible, even with analog, low noise amplification. Where 16bit version is noisy, 24bit version remains almost noiseless, keeping only the recording chain noise. I have already shown it in time record images.
With regard to that drum track, I tried changing the format of the 24b file to 16 bit, in Audacity using 4 types of dither, of the section from 19 to 27 secs. At 21.5 secs the drummer makes a noise, possibly taking his foot of the cymbal pedal, a distinct "tap" - a good reference sound. This comes through clearly in the 16 bit version, and the following record chain noise then varies in quality depending upon the dither applied, as compared to the original 24 bits. But the degradation, of the noise, is not dramatic - more a change of character, take your pick of which is preferable.
Last edited:
You are right I combined your quest for 'identical CDs sound different'(or vice versa) with Dicks quest to where in the chain the jitter was most damaging.
I retract that you were trolling. This time
Anyway, I hope this little exercise helps to realie that whatever happens upstream, it's the DAC that determines the final jitter performance.
Jan
Last edited:
Specifically, what I refer to is the amp having the same sound and tonality no matter which speakers were connected to it, and maintaining its say 1W tonality on all of them, whether it was 1W or 100W, at 100W there was just more of the same. And all the time, it feels like the question is can the speaker deliver what is passes on to it.
Until I bought it, I didn't really know what my ancient AR94s were capable of, and the rumour that the JBL Ti600 speakers were bass biased was exposed as just that, a false rumour. They just need decent drive, and are not very fussy as speakers. They are no wonder of the audio world, but are quite resonable speakers with a reasonable tonal balance and can play fairly loud if pushed without changing their tonality until you push them near to their limits. By that time, there will probably be cops at your door.
Last edited:
16/44.1 isnt better technically in any way to 24/96K and a CD is just as limited as carrying a bunch of LP's around with me.
Do you know how much space a thousand CD tunes/songs takes?? And, I might not like every track on a CD, also. I have most all the older music I want via CD's... plenty, believe me.
Since those olden days of CD buying, I have now 600 new individual tunes via downloads... all of which i chose and like. And, maybe a hundred HD (when do I get the time).
And, I can carry them all with me in a small portable device (memory stick, iPOD or something like or better than the PONO). Or, put it all in the cloud and store it there.
I cant see any down side to it.
THx-RNMarsh
Last edited:
Richard, there are no down sides to the concept except the time it takes you to make your cmpilations. But that's also a lot of fun making.
I went throught that last year and some this year as well, since my Chevy Cruze does have an USB port. On the other hand, the prices of high capacity, say 16 GB memory sticks, has gone down to very affordable levels. so now I use uncompressed (as in MP formats) compilations, which is a joy all unto itself. All the more so since the audio system built into that car is actually rather good, frankly better than I expected.
I went throught that last year and some this year as well, since my Chevy Cruze does have an USB port. On the other hand, the prices of high capacity, say 16 GB memory sticks, has gone down to very affordable levels. so now I use uncompressed (as in MP formats) compilations, which is a joy all unto itself. All the more so since the audio system built into that car is actually rather good, frankly better than I expected.
Last edited:
George, are you sure that your DAC or soundcard is able to switch to 24-bit resolution? Many soundcards do not with WDM drivers ....., and ASIO is needed. Other built-in soundcards have so high intrinsic noise that are not able to tell the difference. This is, in fact a system resolution test.
Please find my ABX result, it took a short time and I would make 100/100 if I was not lazy. The difference in background noise is huge.
foo_abx 1.3.4 report
foobar2000 v1.1.6
2015/05/16 10:49:33
File A: C:\Documents and Settings\Administrator\Dokumenty\Hudba\16b.wav
File B: C:\Documents and Settings\Administrator\Dokumenty\Hudba\24b.wav
10:49:33 : Test started.
10:50:19 : 01/01 50.0%
10:50:26 : 02/02 25.0%
10:50:36 : 03/03 12.5%
10:50:43 : 04/04 6.3%
10:50:57 : 05/05 3.1%
10:51:05 : 06/06 1.6%
10:51:11 : 07/07 0.8%
10:51:23 : 08/08 0.4%
10:51:32 : 09/09 0.2%
10:51:40 : 10/10 0.1%
10:51:52 : 11/11 0.0%
10:52:05 : 12/12 0.0%
10:53:02 : Test finished.
----------
Total: 12/12 (0.0%)
Attachments
Jan
The expectation is that jitter does not travel well. But unless tested it stays in the land of opinion. There still could be hidden gremlins.
In a very large system I did get horrible results until the bad fiber optic jumper was found and replaced. Now on a digital system some would say as long as the output lights up the fiber is good. Really wasn't the case.
I still have some test CDs that I use. These are for STI and other similar measurements. Spares are kept on my server. So next time I make a set I'll try different speeds and even safe mode to see if there is any difference. Don't expect one, but a surprise will be interesting.
I did buy a fiber optic cable meant for the audio fool crowd. It has gold plated connectors! Well noted on the packaging. My opinion is that gold plating makes no difference, but I won't open the package to test it as I keep it intact for humor.
The expectation is that jitter does not travel well. But unless tested it stays in the land of opinion. There still could be hidden gremlins.
In a very large system I did get horrible results until the bad fiber optic jumper was found and replaced. Now on a digital system some would say as long as the output lights up the fiber is good. Really wasn't the case.
I still have some test CDs that I use. These are for STI and other similar measurements. Spares are kept on my server. So next time I make a set I'll try different speeds and even safe mode to see if there is any difference. Don't expect one, but a surprise will be interesting.
I did buy a fiber optic cable meant for the audio fool crowd. It has gold plated connectors! Well noted on the packaging. My opinion is that gold plating makes no difference, but I won't open the package to test it as I keep it intact for humor.
Last edited:
I guess I'm lucky. My original point was, I had tried it long ago, is that a 1 LSB TPDF dither signal at 16bits (~-95dBA) is not audible to me sitting and listening free field with speakers at 3-4'. Members of the Boston Audio Society (audio pros) have made the same observation.
First principles are not subject to opinion. Maybe the capacitor "sound" is really destroyed charge.
YOUR totally unfounded expectation is that jitter 'doesn't travel well'.
I DID test it and showed the graphs. Various amounts and types of jitter in the source didn't make a difference to the DAC output.
I challenge you to come up with a technical sound method how jitter 'can travel' through the system in such a way that it will impact on the analog output of any non-pathologically-defect DAC.
No requirement for measurements yet - just come up with a good story
Jan
- Status
- Not open for further replies.