[NewBuyer]

...of data correlated jitter. That is what makes SPDIF so bad. The method of extracting the clock contains modulation artifacts that is strongly dependent of the programme source. It shouldn't be.

That is why companies have gone to things like ASRCs, to clean up that modulation...

Pat, in your opinion please - does use of ASRC's succeed in ridding of such modulation, and how so? Just curious / trying to learn... 

[*Scotty*]

Daygloworange, while I can't claim that I have the highest resolution system it does seem to be adequate to reveal subtle differences between 16 and 24bit versions of these two recordings,
NAXOS Mussorgsky-Stokowski:Transcriptions 5.110101 and Close Encounters of the Third Kind HDAD 2005. Other DVD-A recordings from Chesky and Telarc are qualitatively better than 16bit recordings with lower grain and a  more real presentation than I have heard from CD's. I suspect that the format has not been fully exploited and probably won't be at this point in time.
On the jitter subject Pat mentioned that in a standalone CD player a good clock right at the DAC is the key to satisfactory performance from the onboard DAC. The current DVD player I am using has shunt regulation of the output stage as well as a FET buffer in the feedback loop of the op-amp which is direct coupled at it's output. The DAC used in the player is an AKM DAC. The player came with an external crystal controlling the internal clock in the DAC. It was explained to me that when the DAC is in operation variances in loading occur at the in and out pins that the external crystal is connected to which destabilize the clock frequency which degrades the sound. A substantial improvement can be realized by removing the external crystal and replacing it with a VCXO  of the correct frequency. The VCXO can have a RMS jitter of 3.5 pico sec. which is pretty good. It is also immune to the operating conditions that upset the external crystals stability. This part costs around $3.00 from Digi-Key or other vendors. All in all a pretty cost effective modification and one that is easily evaluated for it's actual effectiveness assuming that the output stage and power supply are upgraded first.
Scotty

[Daygloworange]

Scotty,

If you have a 16 bit and 24 bit version of the same performance, what you have are 2 different masterings as well. If the performance was recorded in 24 bit, it would have to be dithered down to 16 bit. There is debate about whether or not the random noise that is generated when you dither, is actually audible.

I've never done that, so I can't tell you what to expect, sonically. I believe Russell Dawkins has dithered from 24 bit to 16 bit. Perhaps he'll chime in and let us in on what he hears when he does.

I'm not an expert on dithering, as I haven't felt that moving up from 16 bit recording was necessary.

I know in the case of re-sampling to a lower sample rate, there are filters employed in the process. But honestly, I forget if that's merely truncating or dithering. I used to know all the technicalities years ago, but honestly, since moving from analog to digital recording(and being so pleased with it's sonics), I stopped obsessing about the technicalities of the how's and what's of what is actually going on.

[woodsyi]

I am going to be in Monroe for a St. Patrick's day party Saturday.  I can swing over on Sunday.

Do you mean next Sunday the 16th? I can't do it that day, but every Saturday and Sunday after are available. Let's you and me pick another weekend afternoon, and anyone else who's up for it the same day we pick is welcome to visit too.

--Ethan

It's the 16th.  Too bad as it's just me and my kid on this trip.  I kind of wanted to see your diffusers in action if I got to your place.  I will see if we can work out a date when I head up to Vermont next month but I am not sure my wife would understand the need for this excursion.   

[*Scotty*]

Daygloworange,obviously it's damn hard to compare apples to apples when buying two versions of the same recording released in both 16 bit and 24bit formats. It might be a more valid comparison if a 24 bit master was burned to a DVD-R and a properly dithered 16 bit version was burned to CD-R. Both could then be played back by a high quality player and something might be learned.
Scotty

[Daygloworange]

Scotty,

Unless you compare the same event recorded with a 16 bit recorder (non-dithered) on a 16 bit playback, and a 24 bit recording on a 24 bit playback of the same event, and can A/B them, it's not a fair comparison. The 16 bit format is manipulated by virtue of bits being rounded to approximate values. This is basically distortion. Because of the random algorithmic dithering, the distortion is randomized, and it becomes low level noise.

I have recorded and A/B 'd 16 bit vs 24 bit without dithering, and explained that I did not notice any difference. By virtue of that, it long ago became a non issue for me.

Cheers

[TheChairGuy]

It's not likely the bit length that matters - whether 16 or 24 - as almost no consumer amplifiers have a signal-to-noise ratio of over 100 dB.  Almost certainly no tube amps do. So, tho a difference may exist if we can hear it, it cannot be due to amplifier limitations.

It's the sampling rate, Denny/DGO.  Redbook's 44,100 per second simply is insufficient a sampling of the complex waveforms that make up music.  Further the very real 22Khz limit may be insufficient, as well, to capture the top-end of an event and record it for playback.

24/96 or 24/192 with MLP is indeed a very real advance on Redbook...but it has little to do with dynamic range betterment.  It has to do with capturing the nuance of real live music and top-end/treble extension with higher sampling rates.

You need only an original DVD-A recording, or a DVD-A from a good analog master, to hear the very real betterment of higher sampling rates.

Most of the music buying and listening world don't care about this...but if your an audiophool, you ought to and probably do care and likely have a system capable of really discerning it's benefits to recordings 

Jitter ain't much of a substantial issue next to insufficient, 30+ year old Redbook digital recording standards and room acoustics (both in recording and playback),

One thing further: If you want to number-crunch, here's some food for thought... some years ago a well-respected audio technician named Barry Fox did a mathematical comparison of the relative bit-rates (information-transmission density if you will) of CD vs. vinyl.

We know that domestic CD audio samples at 44,100 Hz, so that its maximum bit-rate is 705,000 bits per channel per second. Mr. Fox then suggested that, if we take some nominal values for things like minimum cantilever excursion to produce audible signal and the contact area of stylus/groove interface, then by the same application vinyl calculates out to have a resolution in the order of 2 MILLION bits per channel per second!

The resolution of vinyl, plagued by non-linearities, noise and the like, can only be equalled or excelled in resolution by a higher sampling rate digital technology such as DVD-A.  Fortunately, Blu-ray adopted this audio standard so we all will have the proverbial cake and eat it, too, in one convenient format 

[BrianM]

It's the sampling rate, Denny/DGO.  Redbook's 44,100 per second simply is insufficient a sampling of the complex waveforms that make up music.  Further the very real 22Khz limit may be insufficient, as well, to capture the top-end of an event and record it for playback.

As far as I know people who can hear as high as 20Khz are the exception.  As for 44100 being "simply insufficient" for "complex waveforms" that sounds like argument by assertion to me.  Any evidence you can point us to?

[Daygloworange]

It's not likely the bit length that matters - whether 16 or 24 - as almost no consumer amplifiers have a signal-to-noise ratio of over 100 dB.  Almost certainly no tube amps do. So, tho a difference may exist if we can hear it, it cannot be due to amplifier limitations.

It's the sampling rate, Denny/DGO.  Redbook's 44,100 per second simply is insufficient a sampling of the complex waveforms that make up music.  Further the very real 22Khz limit may be insufficient, as well, to capture the top-end of an event and record it for playback.

24/96 or 24/192 with MLP is indeed a very real advance on Redbook...but it has little to do with dynamic range betterment.  It has to do with capturing the nuance of real live music and top-end/treble extension with higher sampling rates.

John, I know you think I simply prefer digital because of convenience. It really has to do with the fact that digital is more transparent and dynamic than analog tape as a recording medium.

Saying 44.1 and the filter cutoff is insufficient is mainly conjecture for logicians. It becomes a purely academic debate. The question is, can it be proved empirically?

Can you hear a difference between analog and digital? Many say yes. I say yes as well. But I say digital is better. I say it's better, because it is more linear, quieter and more transparent.

Many prefer the sound of analog. But is it that they prefer the (in theory only) better waveform reproduction, or is it because they simply like the sound because they subconsciously relate to the fondness they've developed for the warm and fuzzies that analog recreates?

Does human perception require a higher sample rate in order to perceive a continuous analog waveform?

I've read theories that assert that human hearing is in fact not a continuous analog function, but in fact a digital stream of neurons firing one after the other, and our brain fills in the gaps, just like digital.

Ok, so let's play logician.

Let discuss sample rates. One millisecond is one thousandth of a second. 44.1 is 44,100 samples per second. So, roughly a millisecond divided by 44.

One sample equals .0226 of a millisecond. So each 16 bit sample at a filter cutoff of 22k has a duration of 0.0226 milliseconds.

The sample rate for motion pictures is 24 frames per second. The duration of a single picture in a motion picture is  44.6 milliseconds.

That's 1973.451 times longer than Redbook's sample rate, yet human's perceive a continuous analog stream of movement while watching a movie( there are a few claims of people who can indeed see a strobe effect of motion picture).

So how important is it that we double the sample rate of the Redbook format?

One thing further: If you want to number-crunch, here's some food for thought... some years ago a well-respected audio technician named Barry Fox did a mathematical comparison of the relative bit-rates (information-transmission density if you will) of CD vs. vinyl.

We know that domestic CD audio samples at 44,100 Hz, so that its maximum bit-rate is 705,000 bits per channel per second. Mr. Fox then suggested that, if we take some nominal values for things like minimum cantilever excursion to produce audible signal and the contact area of stylus/groove interface, then by the same application vinyl calculates out to have a resolution in the order of 2 MILLION bits per channel per second!

Again, that's just number crunching. That means analog has 2.83 times more bits than Redbook. And the extra analog bits are buried in noisefloor and non linearities.

Factor in all that is required to get a sound on and off an analog format. Lots of noise, and lots of distortion, lots of channel crosstalk.

How many people are even aware of how Dolby or Dbx noise reduction even works on analog tape?  Encode/Decode.

Digital = encode/decode.

Dolby or Dbx typ noise reductions have been used on nearly every professional analog multi-track recorder since the sixties.

Cheers

[Ethan Winer]

A lot of people have yet to be convinced by the inherent audible superiority of 24/96.

No kidding, and this series of blind tests proves that nobody can tell when high-res content has been reduced to 16 bits at 44.1 KHz:

http://theaudiocritic.com/blog/index.php?op=ViewArticle&articleId=41&blogId=1

More here from the authors of the study:

http://www.bostonaudiosociety.org/explanation.htm

It isn't the level of distortion that is point, but the distribution. Some is euphonic, and some folks like it. Others (like 7th and 9th order) are especially nasty, and folks don't care to hear them.

I guess you didn't read my posts in this thread either. Normal amplifier designs should not have more 7th and 9th harmonic distortion than lower orders. Moreover, Harmonic Distortion (THD) does not appear in isolation. Wherever you have Harmonic Distortion you also have Intermodulation Distortion (IMD). So saying that this or that higher order harmonic dominates the spectrum is wrong. And regardless of the distortion frequencies, if they're 80 dB or more below the music they are inaudible, period. This is the real issue, and this is very easy to prove for once and for all. I have proven this to my own satisfaction in my home studio by injecting varying amounts of artifacts into music, and I can prove it to anyone else who is willing to visit me to see and hear such tests in action.

As for level of audibility, I don't imagine that you ever measured capacitor distortion levels. They are typically down in that -120 dB range. I don't know of many folks here who won't agree that caps sound different.

Distortion that is at -120 is inaudible. Pat, please read my Believe article as it explains why people think they hear a difference even when no difference is possible. It's not a lot to read, and I promise you'll have a new perspective on this stuff afterward.

I have 24 bit capability in my recording studio. I use 16 bit. I've done many tests between 16 and 24 bit, and never found any difference that I could hear.

Exactly. By and large, professional musicians and recording engineers who actually do this stuff every day are less likely to believe in the kind of "magic" that non-pro audiophiles believe. I too have tested all of these things - high bit rates, high sample rates, dither, jitter, and so forth, and none of it makes any difference. It's all acoustic comb filtering I tell you!

--Ethan

[*Scotty*]

Ethan,in as much as distortion levels of better than -80dB were achieved more than 35 years ago your statement that amplifiers with -80dB of distortion are transparent implies that there are no sonic differences between amplifiers meeting this criteria. By extension, any component meeting this criteria is no longer a factor in the sound produced by the system,therefore there really hasn't been any progress made regarding amplifier design in the last thirty years. Is this actually your position on this subject.
Scotty

[BrianM]

Ethan,in as much as distortion levels of better than -80dB were achieved more than 35 years ago your statement that amplifiers with -80dB of distortion are transparent implies that there are no sonic differences between amplifiers meeting this criteria. By extension, any component meeting this criteria is no longer a factor in the sound produced by the system,therefore there really hasn't been any progress made regarding amplifier design in the last thirty years. Is this actually your position on this subject.
Scotty

I'm not Ethan, but I'll put it this way: the difference in sound between ALL amplifiers of a given type (and every other component, provided they share basic design principles) tends to be overstated.

[art]

I guess you didn't read my posts in this thread either. Normal amplifier designs should not have more 7th and 9th harmonic distortion than lower orders. Moreover, Harmonic Distortion (THD) does not appear in isolation. Wherever you have Harmonic Distortion you also have Intermodulation Distortion (IMD). So saying that this or that higher order harmonic dominates the spectrum is wrong. And regardless of the distortion frequencies, if they're 80 dB or more below the music they are inaudible, period. This is the real issue, and this is very easy to prove for once and for all. I have proven this to my own satisfaction in my home studio by injecting varying amounts of artifacts into music, and I can prove it to anyone else who is willing to visit me to see and hear such tests in action.

I did not say that 7th and 9th appear all by themselves, are more than the lower orders, or dominate the spectrum. I don't know how you come to believe that I implied that.

Frankly, what you, or any other listener, can or can not hear is not the issue. My job, as a designer of electronics, is to minimise all forms of distortion. It is my job to determine how far they must be lowered. And by what means to lower them.

It this case, we are talking about jitter. Again, unless we are talking about SPDIF specifically, jitter is not a major issue. It is with SPDIF. How low it needs to be lowered, and how, is the question.

There is not a lot of controversy on the "how" part. Except that some designers do not like ASRCs. The only controversy seems to be on the "how much" part.

I know from my experience what point the "how much" gets to the "that's enough" point. That is what I believe the folks here want to know. You seem to have your mind already made up that you already know the answer, and think that anyone who says otherwise is is making stuff up to sell equipment.

I don't make stuff up, or sell jitter reduction gear. You can chose to believe my position or not. That is your choice. I am not altering my stance and throwing out >15 years of work simply because you have differing beliefs in distortion.

Are you sure that you don't belong to the BAS? You sure sound like you do.

Pat

[*Scotty*]

Pat, the BAS you are referring to doesn't come up until page 6 of a Google search, you may be guilty of obscurantism here. Do you have a position on aftermarket clocks, it seems hard to justify the expense of some of these clocks when a less expensive alternative like a $3.00 VCXO can be used.
Scotty

[Daygloworange]

Ethan,in as much as distortion levels of better than -80dB were achieved more than 35 years ago your statement that amplifiers with -80dB of distortion are transparent implies that there are no sonic differences between amplifiers meeting this criteria. By extension, any component meeting this criteria is no longer a factor in the sound produced by the system,therefore there really hasn't been any progress made regarding amplifier design in the last thirty years. Is this actually your position on this subject.
Scotty

I think what Ethan is getting at is whether or not you can hear the loose change in the symphony conductor's pocket that is 80 db lower in amplitude than the music he is orchestrating.  

What about 120 db lower?

Cheers

[Ethan Winer]

Ethan,in as much as distortion levels of better than -80dB were achieved more than 35 years ago ... Is this actually your position on this subject.

There's more to what we hear than just distortion. There's also noise, frequency response, and time-based errors. Though that last one won't appear in an amplifier. There are also many types of noise, and at least two types of distortion. But Yes, that is more or less my position. If a device passes audio with noise too soft to be heard if you turn it on and off, and has a response within, say, +/- 0.1 dB from 20 Hz to 20 KHz, and all other artifacts are at least 80 dB below the music, that device can by definition be considered audibly transparent. Whose definition? Why, mine of course! But I still have that $100 for each person who can identify very soft artifacts.

Let me modify my wager a bit, as stated in this post, just for my own safety. Earlier I said I have $100 for anyone who can hear jitter that's 120 dB below the music. But 80 dB is getting a little close to what I consider the threshold in some contrived examples. I doubt anyone can hear "normal" distortion artifacts that are 80 dB below the music because I couldn't hear 3 KHz when mixed at -80 with a 100 Hz tone. So for the purpose of this new wager I have $100 for anyone who can hear amplifier distortion that's 100 dB below the music. This is still louder than jitter, and it's only a few dB below the noise floor of a CD. Since Pat said that capacitor distortion of -120 dB is audible, anyone who wants to bring such a capacitor to my house is a perfect candidate for snagging the $100 off me. Pat, see if you can find someone near me and send them a cap to test.

--Ethan

[Ethan Winer]

My job, as a designer of electronics, is to minimise all forms of distortion.

Of course, and I agree fully. I discussed this once with Bruno Putzeys. He told me he agrees that jitter at -120 dB is inaudible, but he still designs circuits with lower jitter because he can. Ain't nothing wrong with that. But at least Bruno understands that nobody can hear anything that's 120 dB below the music. More people need to try tests like these so they can see first-hand how it's impossible to hear anything that soft. Especially in an untreated listening room. Do you know how far below the room's own ambient noise level -120 dB is?

I know from my experience what point the "how much" gets to the "that's enough" point. That is what I believe the folks here want to know.

Yes, this is exactly the issue and it's trivial to test, as I keep saying over and over. If you believe distortion from a capacitor at -120 dB is audible, the only reasonable conclusion for me is that you need to test this stuff again using the methods I described. I don't know how else to say it.

anyone who says otherwise is is making stuff up to sell equipment.

I do believe that some vendors make up stuff to sell gear (Shunyata is a perfect example), but I'm sure many others really do believe it. Doesn't make it true.

--Ethan

[miklorsmith]

So, are we back to jitter being solely a noise artifact with a dB-below-signal quantifier?  Higher jitter doesn't affect the primary signal?

[jhm731]

16/44.1 verses 24/96 files for your listening pleasure:

http://www.soundkeeperrecordings.com/format.htm

[JeffB]

Can jitter really be considered noise or compared to noise?
I guess it depends upon how you define noise.
I think of jitter as skewing the time domain.
I think of noise as extra information that shouldn't be there.
Perhaps the ear is more sensitive to time smear than noise.