John Curl's Blowtorch preamplifier part II

[hhoyt]

AMI

Was that the plant over in Burlington, or thereabout?

Yep! You knew of it I guess?

Howie

Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org

ex-Dir of Engineering
AMI, LLC
Burlington, NC
www.ami-media.com
(for some reason the website is still up..?)

 

[1audio]

And the resulting mistracking will cause a loss of data and E32s, and a potentially degraded sound quality. Uncorrectable errors are not a parameter on a disc, although they could be caused by a disc problem, they are an error flag in the CD Player's error decoder indicating that the error severity exceeded the error correction circuit's ability to correct. It is the incidence of E32s which cause a change in sound, whether due to a mute or interpolation.

The CD system is really robust. I have a CEC TL1 transport. At one point I noticed a sort of grating noise coming from it. It seems the disk was rubbing on something on every revolution. On a record it would be very audible. Even though this should cause beaucoup errors there was no audible degradation, no sonic indication of a problem at all. I think it had been an issue for a while. The case is pretty robust so the noise was not audible with the door closed.

I do think there is a connection between the EFM rate and the output data. However its the reading of the disk that's is servoed back to the buffers and then to the master clock. Isolating the master clock and its supply from the rest of the transport/player does seem to help.

 

[jan.didden]

My presence has been requested at a cell phone manu that I can't name and OPPO to consult on audio quality on portable devices, the irony.

Scott, is there a tech reason that sound Q in a portable device or cell phone cannot be the best we can do? I understand that you must have a designer who knows his stuff, and good (active) devices/chips/DSPs, but power would probably not really be a limitation, assuming either headphones or streaming output to an audio system.
In other words, if someone wants to really do it, is there a real problem?

jan

 

[jan.didden]

No, the testing would be have to be done carefully to always ensure we never had an E32 - the aim would be to go as close to causing one to happen, without going there; the error correction should be forced to constantly be invoked throughout the playing, but always to be able to recover correct data - the real aim is to put the servo systems under heavy stress, force them to constantly demand current spikes from the power supplies to maintain good tracking.

I seems to me a more efficient way to test this is by imposing an error voltage on the supply and increasing the error until you see an effect on the output. That way you can clearly document the susceptibility in terms of error voltage, frequency etc. and you then can compare players. Also more easy to be replicated by others. Is that an idea?

Jan

 

[Kastor L]

My belief is that this capability of the mind is what allows convincing sound to be achieved - if a certain threshhold of quality is achieved in playback then the mind can fill in what's necessary, and discard the irrelevant - the listener accepts the illusion as being "real". However, any performance standard below that is not sufficient - the sound is obviously "fake"

I've never heard a system sounding real myself.

We can't hear everything at once so with rich information in stereo imaging and locality, if all we do is listen to spatial cues, that should sound pretty real.

Luckily I mostly use in-ear monitors so that will never happen, unless it was a binaural recording.

 

[simon7000]

In this case you would be trying to see if a player, when drawing more current from the tracking servo will introduce noise in the analog outputs? Its sad, but many players do this even with good discs, even some moderately priced players. Yet another reason to use outboard D>A conversion.

Cheers,

Howie

Howard Hoyt
CE - WXYC-FM 89.3
UNC Chapel Hill, NC
www.wxyc.org

I have seen CD players that use the 5 volt supply rail as the reference voltage for the D/A converter. The other issue is just how noisy the internal EMI can affect the output linear stages. So yes my vote is an external converter.

 

[fas42]

I seems to me a more efficient way to test this is by imposing an error voltage on the supply and increasing the error until you see an effect on the output. That way you can clearly document the susceptibility in terms of error voltage, frequency etc. and you then can compare players. Also more easy to be replicated by others. Is that an idea?

Jan

Certainly sounds like a good starting point, I would record the distortion spectrum of 19/20kHz mixed tones while varying the error voltage - and see if that pointed out a trend ...

 

[fas42]

I've never heard a system sounding real myself.

We can't hear everything at once so with rich information in stereo imaging and locality, if all we do is listen to spatial cues, that should sound pretty real.

Luckily I mostly use in-ear monitors so that will never happen, unless it was a binaural recording.

Then you're in for a treat, one day ... !

IME, when the information is reproduced cleanly enough at realistic levels then the "realness" pops into life - extremely heavily layered mixes unravel, and every sound element can be clearly "seen" in its own space, you're not submerged under the totality of the sound as is usually the case. Just like in real life, when there are multiple sounds occurring around you, coming from all directions, from totally different types of "things" ... you can, without difficulty, switch your focus to one instrument or source of sound in the recording and follow its 'line', without struggling to do so - the integrity of that "track" in the recording remains intact, doesn't falter.

 

[marce]

Sounds like somebody's into serious photoshopping..

jn

Lightroom5 and the Google add-ons...
Google Nik Collection

 

[marce]

Again we're locked into the digital domain - conveniently ignoring the fact that there is a mechanical system relying on electricals to do the recovery of the data. I would do some experiments: make multiple copies, bit-identical, of a CD and then create a series of physically unbalanced disks, by adding weight to one side of the edge of the disks, steadily increasing the unbalance; and on another series, shim up the centering hole on one side to cause the play surface to increasingly wobble vertically while spinning. Then we'll see whether we always get "no difference in sound", ...

Totally pointless experiment really...But one that could also be carried out using a LP, would we also hear a difference?
As said off board DAC and galvanic isolation is the best solution, this is pretty standard in any requirement where an analogue input or output needs to be isolated from the rest of the circuitry...it seems to be an endless loop regarding this 'problem' a problem that can be solved by some basic engineering. It all comes down to EMC engineering, often an afterthought in DIY and some commercial designs...whereas in life/mission critical designs it is considered from the start.
Most CD players I have seen the insides of (even expensive high end) frighten me, often cheap PCBs, crappy ribbon cables. One that retails for quite a few quid (midrange) shocked me when I saw a single sided CEM board inside.
CEM for those that don't know what it is, cheap!
http://en.wikipedia.org/wiki/Composite_epoxy_material

 

[Kastor L]

Fas42,

Yes I see what you're saying.

To me let's say Etymotic ER-4B or a STAX with a binaural recording sounds like an illusion, or you could call it lots of spatial information, I can no longer hear the transducers fine spatial character since it's masked with the spatial illusion. Crossfeed will cause an illusion as well. The point is normal stereo is more revealing in an IEM or headphone, of spatial character for sure.

For total audio/visual immersion, binaural is more realistic of course, if that's the intention.

The speaker equivalent is I suppose the listening room sweet spot in a rather anarchic setting.

 

[marce]

And then there was two

 

[Jakob2]

You excerpted my statement and left out the most important part, so I'll repost it: I said: "If on that single player, between two bit-identical discs there are no E32s there can be no difference in sound."

Analog contributions inside a player such as the ever-present tracking servo hash bleeding into the audio supply would be there in either case, since we are talking about the one player. Therefore, any differences between two bit-for-bit identical discs in that one player will indeed be due to E32s.

Change one variable at a time, please!

<snip>

I´m sorry hhoyt, but the most important part (at least to me) of your post was exactly that premise "...... there are no E32s there can be no difference in sound" .

"Change one variable at a time, please!" is spot on; "check if the premises are right" is the other point.

You take it for granted that there could be no difference between the "servo hash" (or any other parameter) but you can´t be sure.

So, the reasoning would be more like "if two cds give the same data, then any audible difference can´t be caused due to E32s.

 

[Jakob2]

The tests weren't blind, much less double blind. Done double blind, the correct way to evaluate by ear only (no peeking or preconceptions), the results are quite different.

No, just look at:

P. J. Duncan, P. Dodds, N. Williams 2008, Audio Capacitors. Myth or reality?, AES 124tth Convention Amsterdam.

P. Dodds, P. J. Duncan, N. Williams 2006, Assessing the effects of loudspeaker crossover components on sound quality, IOA Reproduced sound 22 Proceedings.

 

[fas42]

Totally pointless experiment really...But one that could also be carried out using a LP, would we also hear a difference?

Pointless ...? The digital is still working perfectly, but is the analogue affected, or not?

As said off board DAC and galvanic isolation is the best solution, this is pretty standard in any requirement where an analogue input or output needs to be isolated from the rest of the circuitry...it seems to be an endless loop regarding this 'problem' a problem that can be solved by some basic engineering. It all comes down to EMC engineering, often an afterthought in DIY and some commercial designs...whereas in life/mission critical designs it is considered from the start.
Most CD players I have seen the insides of (even expensive high end) frighten me, often cheap PCBs, crappy ribbon cables. One that retails for quite a few quid (midrange) shocked me when I saw a single sided CEM board inside.
CEM for those that don't know what it is, cheap!
Composite epoxy material - Wikipedia, the free encyclopedia

Yet many here might say there can't be a problem caused by that 'cheap' engineering - why not find out, for certain, ?

 

[SY]

No, just look at:

P. J. Duncan, P. Dodds, N. Williams 2008, Audio Capacitors. Myth or reality?, AES 124tth Convention Amsterdam.

P. Dodds, P. J. Duncan, N. Williams 2006, Assessing the effects of loudspeaker crossover components on sound quality, IOA Reproduced sound 22 Proceedings.

Those weren't the "tests" that were referenced.

 

[fas42]

The speaker equivalent is I suppose the listening room sweet spot in a rather anarchic setting.

The speaker equivalent is much better than that ... the whole of the listening space becomes "sweet". Something similar in headphone technology, from how people describe it, is the Smyth Realiser - one can walk around anywhere in the room, and the sense of the musical event and its location remains consistent, stable with respect to that physical space.

 

[Jakob2]

Those weren't the "tests" that were referenced.

Ahh, human communication- isn´t it fascinating?

Another try:

<snip>
Done double blind, the correct way to evaluate by ear only (no peeking or preconceptions), the results are quite different.

No, just look at:

P. J. Duncan, P. Dodds, N. Williams 2008, Audio Capacitors. Myth or reality?, AES 124tth Convention Amsterdam.

P. Dodds, P. J. Duncan, N. Williams 2006, Assessing the effects of loudspeaker crossover components on sound quality, IOA Reproduced sound 22 Proceedings.

 

[Kastor L]

The Smyth realizer is finely tuned binaural yeah.

If you combine that with the Oculus Rift, you will have a very convincing reality, in theory.

It's so convincing the DAC and amplifier will become less important I suspect, way too much information.

It's not like you're riding a rollercoaster and suddenly think "Hmm this sounds like LME49720 with Nichicon KZ".

 

[vacuphile]

Again we're locked into the digital domain - conveniently ignoring the fact that there is a mechanical system relying on electricals to do the recovery of the data. I would do some experiments: make multiple copies, bit-identical, of a CD and then create a series of physically unbalanced disks, by adding weight to one side of the edge of the disks, steadily increasing the unbalance; and on another series, shim up the centering hole on one side to cause the play surface to increasingly wobble vertically while spinning. Then we'll see whether we always get "no difference in sound", ...

Why don't you do it and inform us about the results instead of littering this thread with your brain farts. This test shouldn't be beyond your technical capabilities.