Good point Richard,
You are correct as well, the defects are pressed in at the time of the pressing. That process can add off-center and warp issues. Fun for the lens to try and follow in the center where the rotation is around 500 rpm. And you wonder how read errors can creep in?
If you can "record" those tracks in a 24 bit format, it will sound a lot better than a copy of a CD. The CD format would lock you in if you wanted to play your music in a normal CD player, like in a car for instance.
-Chris
You are correct as well, the defects are pressed in at the time of the pressing. That process can add off-center and warp issues. Fun for the lens to try and follow in the center where the rotation is around 500 rpm. And you wonder how read errors can creep in?
If you can "record" those tracks in a 24 bit format, it will sound a lot better than a copy of a CD. The CD format would lock you in if you wanted to play your music in a normal CD player, like in a car for instance.
-Chris
Richard, there is soo much interesting and sometimes rare material out there (new and old) that is available only on CD that will never see the day as HD download. Why limit the choice?
If one is really into music, as opposed to listenig to only what's on HP's top disc list or endlessly comparing various transfers of 'Kind of Blue', 'Dark Side' or 'Brothers in Arms'
, this is the valid reason to own as good CD playback device as one can afford.Here's one good example of rare material and no HD downloads:
Mosaic Records - Limited Edition Jazz Collections
Mosaic Records
(I have sevaral of their box-sets)
And then there's underground music, experimental stuff, small labels, CDs of lesser known artists from all around the world that you buy at concerts and so on..
Last edited:
And in fact that is a signature of a system working correctly. When I first got heavily involved in the audio game I did the rounds of listening to every system I could get an ear to, and essentially every one of them was flawed in this respect - when asked to deliver a higher volume the quality of the sound started to disintegrate at a certain point - in some cases at an amazingly low level.
Things have improved. At the recent audio show the majority still had this same flaw, but there were a handful that, finally, were well enough engineered to not show this behaviour in a completely obvious way.
Thank you Pavel for the drum sound uploads.
I listened to them through headphones, after midnight.
My first impression was that the 16bit recording was a bit louder.
After many listening rounds I wasn’t sure anymore.
Then, I compared repeatedly the two recordings in detail (sections of 1-2 seconds). Couldn’t perceive any difference either with certainty or at all anywhere else but the hat hit at 27s (the decay sounds to me as more abrupt on the 16bit recording).
After the listening sessions, I checked in detail the signal level of the files second by second through a wave editor.
At digital silence: 16bit is –90.xx dB, 24bit is –133.xx dB
At very low passages <-60dB level: 24 bit is 2 to 5 dB lower in level.
At passages >-50dB level, there are no differences (only at the second decimal of a dB)
George
I listened to them through headphones, after midnight.
My first impression was that the 16bit recording was a bit louder.
After many listening rounds I wasn’t sure anymore.
Then, I compared repeatedly the two recordings in detail (sections of 1-2 seconds). Couldn’t perceive any difference either with certainty or at all anywhere else but the hat hit at 27s (the decay sounds to me as more abrupt on the 16bit recording).
After the listening sessions, I checked in detail the signal level of the files second by second through a wave editor.
At digital silence: 16bit is –90.xx dB, 24bit is –133.xx dB
At very low passages <-60dB level: 24 bit is 2 to 5 dB lower in level.
At passages >-50dB level, there are no differences (only at the second decimal of a dB)
George
Attachments
Last edited:
Thank you George, I might point out that if properly dithered there is no quantization noise simply a noise floor. I might add that if you A weight it the 93db goes to about 95dB at 16 bits.
Have you ever seen an old Sony DiscMan in operation? You remove a CD, put it back
into its box, put the box back into the rack, and it keeps playing, playing, playing...
And a DiscMan is low-end.
Obviously, the jitter of the pits has ZERO influence; or there could be no music
if there are no pits/sec.
Either the drive gets all information, then everything is fine, or it gets most,
then the true information is recovered from redundant info from a distant corner
of the disk and everything is fine.
If there is not enough, you can hear that quite easily, everybody here knows
what a worn-out laser sounds like, or the result of cleaning the CD with a steel brush.
Burning a properly recovered content to a new CD propagates no old jitter;
the result depends only on the medium and the CD writer. The recovered data
image to be burnt is immaculate, a boot CD for my laptop cannot argue about
some flipped bit; it simply would not work.
Last edited:
Last night you said your normal listening space is 15dBA which is in the range of anechoic chambers, moving target?
My answer will probably generate some controversy.
I noticed that complex assemblies (I mean things like Cascode, Super pairs, diamonds etc....) always help to reduce distortions or other evils. BUT they often add a "signature". I'm referring to Cascode, as the most obvious for me.
It is totally subjective, and hard to generalize.
Some "signatures" are nice ! Since I first heard an old luxman hawksford
jammin' through an Sanken EF3 .... I fell in love.
OS
Ummm... yes, except jitter. Who told you otherwise?
DSP 101. Jitter effects don't propagate over a bitwise copy. Bit errors are either correctable, or not. You may hear uncorrectable errors, but jitter plays no role in that.
I wonder where you take the energy from to stay here continuously;
I usually need recovery time after 3 days max.
In Audacity, amplified by 30dB it's easy to see why this is so: the 24b decay trail is still clear of the noise, but not so for the 16b. The ear "follows" the sound down as it fades, and with the 16b the noise accompaniment starts to intrude noticeably on the last audible segment of metallic ringing - the ear rapidly has to work harder to filter out the musical component at the very tail of that bit of sound, and subjectively it seems to "disappear".
There is noise in the optical domain that can cause error, decreasing the SNR such that the bit information over laps. We call it optical phase modulation error. Here is a link to a random scholarly paper on it ( Reducing bit-error rate with optical phase regeneration in multilevel modulation formats. - PubMed - NCBI ). I have maintained, that the electronics are beholden to what happens optically until the signal hits the photo-detector. This is why the old Spatial filter worked. Theta, PS Audion, Cal Labs etc used them. A qouoe on the optical signal with a filter; " a laser researcher and non-audiopile, took a look at the
laser and optics in CD players and noticed that the optical aperture
(opening) of the lens seemed larger than optimum. He also noticed a
consequential high level of noise (jitter) in the recovered EFM (eight
to fourteen modulation) signal from the disc. The EFM signal is a series
of nine discrete frequency sinewaves that contain the information encoded
on the disc. These nine frequencies correspond to the nine discrete pit
or land lengths embedded in the CD. Although the EFM signal is composed
of sinewaves and appears to be analog in nature, the sinewaves zero
crossings delineate the digital code transition.
A tiny donut-like disc was machined that would fit over the lens, thus
making the aperature smaller. Looking at the EFM signal recovered by a
CD transport fitted with the Spatial Filter, it was markedly and quite
measurably less noisy, presumably due to the increased depth of focus.
However making the len aperature smaller also reduces tracking ability.
The CD's designers apparently decided that reliable tracking was more
important than a cleaner EFM signal"
laser and optics in CD players and noticed that the optical aperture
(opening) of the lens seemed larger than optimum. He also noticed a
consequential high level of noise (jitter) in the recovered EFM (eight
to fourteen modulation) signal from the disc. The EFM signal is a series
of nine discrete frequency sinewaves that contain the information encoded
on the disc. These nine frequencies correspond to the nine discrete pit
or land lengths embedded in the CD. Although the EFM signal is composed
of sinewaves and appears to be analog in nature, the sinewaves zero
crossings delineate the digital code transition.
A tiny donut-like disc was machined that would fit over the lens, thus
making the aperature smaller. Looking at the EFM signal recovered by a
CD transport fitted with the Spatial Filter, it was markedly and quite
measurably less noisy, presumably due to the increased depth of focus.
However making the len aperature smaller also reduces tracking ability.
The CD's designers apparently decided that reliable tracking was more
important than a cleaner EFM signal"
But as the BAS survey showed these circuits have up to .05% range THD's at full scale even worse than the PONO.
What's one to think? It is a good sounding circuit block. I guess I can't enjoy listening anymore; measurements don't lie.
Any suggestions for the purrfect replacement? Just kidding
Hi canyoncruz,
Yes, this is correct in that the tracking is more important than increasing the depth of field for the lens. The E and F diodes lay off to the sides, so these signals need to get through as well.
The size of the aperture isn't the reason what the RF signal is noisy. Not the only one. The optics inside the head assy. have a great deal to do with recovering the beam while also cutting off the signals further from the sides that are simply noise. The other key location where noise comes in is the I-V converter for the 6 diodes (4 for a single beam type - Philips. NEC made the best optical head for CD players, and it was a single beam system. Sony has made some very good die cast head assemblies as well. The optics and physical alignment of the head in relation to the disc is where the worst of the trouble comes in. Fix that and you have fixed most of the problems.
The electronics used in amplifying the recovered laser beam is usually not done as well as it could have been. I am experimenting with filters in this area. The quality of the eye pattern is the one critical thing current manufacturing seems to miss.
-Chris
Yes, this is correct in that the tracking is more important than increasing the depth of field for the lens. The E and F diodes lay off to the sides, so these signals need to get through as well.
The size of the aperture isn't the reason what the RF signal is noisy. Not the only one. The optics inside the head assy. have a great deal to do with recovering the beam while also cutting off the signals further from the sides that are simply noise. The other key location where noise comes in is the I-V converter for the 6 diodes (4 for a single beam type - Philips. NEC made the best optical head for CD players, and it was a single beam system. Sony has made some very good die cast head assemblies as well. The optics and physical alignment of the head in relation to the disc is where the worst of the trouble comes in. Fix that and you have fixed most of the problems.
The electronics used in amplifying the recovered laser beam is usually not done as well as it could have been. I am experimenting with filters in this area. The quality of the eye pattern is the one critical thing current manufacturing seems to miss.
-Chris
Anatech. study up on High Numerical Aperture lenses. Like I said, the quote wasn't completely accurate, he was paraphrasing a stereophile article from 1992. The focus depth (or I guess he meant working distance) of the lens doesn't changes nor is any part of the the information for tracking. The optics inside have no sort of spatial filtering. It is signal to noise of the optical signal. This is why we were able to measure a significant difference at the photo diode. That is important, before any electronics, other than the PD.
The outside noise level drops to 25 dbA. Inside it runs 15dbA. There are enough noise sources inside to keep if from dropping any lower. I did get rid of a battery powered clock as it was too noisy.
An anechoic chamber not only has to have isolation in requires amazing amounts of absorption. I don't have large amounts, just normal ones.
As anyone who has built quiet spaces can tell you the details are where it counts. The hotel I am working on at the moment has a high quality noise isolation specification. Basically double layers of drywall on each side. No gaps in the drywall at the floor. Alternate wall studs and concrete floor and ceiling. Rugs on the floor to reduce impact noise. No electrical wall boxes on either side near each other. Conduits to the boxes are also isolated.
The walls will easily reduce noise by 45+ dB. It is not very hard to build a quiet room.
Gerhard,
The issue is does copying CDs increase jitter. The smart money is that it does not. It would be nice to have an actual measurement to verify that. It is quite possible there is some unsuspected mechanism that would cause jitter to increase. There certainly are enough folks who believe CDs duplicated at different speeds sound different. Even claims that running a PC in safe mode during duplication produces better results.
So I asked Jan if he could do a quick test. This certainly has gotten drawn out. I don't have any blanks or I would have done it.
- Status
- Not open for further replies.