Amarra

[audioengr]

Bit-perfect is not everything. I can give you two bit-perfect identical rips that sound entirely different, that is if your system is resolving enough.

 

Unlike some, I like to keep an open mind.

 

Steve N.

Empirical Audio

 

[cfmsp]

 

I keep an open mind as well, Steve, but you're claiming that there are data integrity issues in a product that is known for the ease with which it can provide bit perfect playback (iTunes on a Mac).

 

Surely, you must have something more to go on than what you've offered thus far?

 

Mind you, your response to my request to substantiate your data integrity claims is:

 

"I can give you two bit-perfect identical rips that sound entirely different"

 

Even if you could do that (provide diffferent sounding identical rips),

that wouldn't naturally point to 'data integrity' issues; indeed, it would seem to point to something else entirely, and yet you've made the leap of faith, indeed, you've stated that:

 

"This indicates that Core Audio is somehow changing the data"

 

...and apparently you've done so without actually checking the data or knowing what data has changed!?

 

 

 

 

 

 

 

 

 

[HarryHWombat]

This indicates that Core Audio is somehow changing the data as well as generating a lot of jitter.

Steve N.

 

I think it would help "unconfuse" everyone if you could either let us know what you think is changing or what you mean by data. As everyone points out iTunes in the correct configuration is bit-perfect which means that the data being sent is exactly the same as the data on the disk. Most people understand the data as the set of bits - iTunes is provably bit-perfect therefore the data is unchanged. yet you say that the data is changing. As a consequence, you must mean something different by the term "data". Maybe the total of the information sent including jitter, noise and maybe something else? Maybe a small increase or decrease in the average transmission rate of the bits?

 

[audioengr]

Nothing like that.

 

I'm talking about what comprises the frames of the data stream:

 

1) music data

2) padding bits - offset

3) control information, such as word-length, sample-rate, pre-emphasis etc..

 

Any of these can change the way the data sounds IMO.

 

Steve N.

 

[HarryHWombat]

ok now I get it.

 

But in this case Apple would not be conforming to recognised standards - unless there are different ways to implement the standard which would kind of make it not a standard (one would assume). One would also assume that conformance to this standard would be part of Apple's standard release testing and would be highly automated through a series of test suites.

 

I don't know much about this but if it is true that the standard is just that - one way to implement - then either Apple have a bug in their test harnesses (possible) or .. well, the other alternative is hard to believe, right, and also makes no sense.

 

Although I continue to believe that perceived differences in subjective testing are always to be treated skeptically there is also a small concern with the more objective tests which are usually computer-to-computer and not computer-to-dac therefore the testing path is different to the real path. A small delta difference in a stream which a computer is immune to a dac may not be .. I have no idea, though, what this could be although Peter has been doing his interesting objective tests.

 

 

 

[cfmsp]

 

"But in this case Apple would not be conforming to recognised standards "

 

Actually, in this case, Apple would not be confirming to THEIR OWN standards, as AIFF and ALAC were developed BY Apple.

 

 

 

 

 

 

 

[HarryHWombat]

I think Steve is referring to a transmission standard/protocol such as iec60958 (found thru googling) rather than AIFF and ALAC etc which are file formats and in the case of ALAC a codec. Basically it seems to be a model for transmitting PCM data but I still don't see what Apple could have done wrong.

 

 

 

[cfmsp]

 

Harry, I disagree.

 

"Offsets" are defined in the AIFF standards spec. That's what I think Steve is referring to.

 

The document I found states:

 

"offset determines where the first sample frame in the soundData starts. offset is in bytes. Most applications won't use offset and should set it to zero."

 

 

clay

 

 

 

 

 

[audioengr]

Okay, I just downloaded Amarra test software and it works with my Tascam US-144/PAce-Car 2 reclocker combo.

 

On to Amarra testing:

 

Pre-Amarra: Playing iTunes from the Mac Mini through the Tascam/Pace-Car combo is not quite as good as playing the same hardware using PC and Foobar0.8.3.

 

When I install Amarra, however, the result is identical. Both have highly focused, vivid presentations. Very live and 3-dimensional.

 

Because both are going through the Pace-Car 2, I know that the difference is not jitter. They are using the same identical clock in the Pace-Car 2, not the computer clock.

 

So, I conclude that these sound virtually identical:

 

1) PC with Foobar 0.8.3 (unmapped) driving Overdrive DAC or Off-Ramp USB with Axis or Polestar cable

2) PC with Foobar 0.8.3 (unmapped) driving Tascam US-144 into Pace-Car to good S/PDIF cable

3) Mac with Amarra/iTunes driving Tascam US-144 into Pace-Car to good S/PDIF cable

 

What this says is that not only is the jitter a problem with iTunes on Mac, but the Mac Core Audio is also mucking up the data somehow, like kmixer does on a PC.

 

Unfortunately it costs $1500 to get the the same place you can get to for free on a PC, but after all it's iTunes.

 

Next I need to test straight USB to see if the Mac jitter is reduced to the levels of the PC.

 

Steve N.

 

[cfmsp]

 

 

Steve,

 

You keep suggesting that 'broken' iTunes causes jitter, yet software has nothing to do with jitter.

 

As Gordon Rankin said elsewhere in response to your earlier post (which you duplicated on another forum):

 

"First there is no JITTER in a computer or USB for that matter. The only jitter there is in a system like this is the intrinsic jitter of the USB receiver to I2S (or other protocol).

 

If that's the case and it's your design then you have some serious problems."

 

Steve, If I'm not mistaken the USB Overdrive DAC (used in your 'tests') is YOUR design. Why do you continue to blame iTunes / CoreAUdio for issues that an authority such as Gordon claim is with your own device?

 

:0

 

clay

 

 

 

 

[Dennis L Jacob]

Clay you hit the nail on the head with that observation. I've been thinking the same thing. At RMAF, I'd have to say of the folks using a computer music server for their source - 95% were using an Apple MacBook as their computer system with iTunes as the software used. I'm betting that USB DAC he is using is the reason for the jitter and other sonic problems. That and not wanting to hear his gear is the problem, not the software. That coming from a guy who has been a 25 year PC user. I switched earlier this year to a family of Mac computers. Wish I'd done so years ago.

 

[audioengr]

I guess you dont get it. Okay, let me explain.

 

The reason that My Overdrive DAC is such a GOOD tool to listen for jitter is that it has no intrinsic jitter reduction circuits, at least on the I2S input. The S/PDIF input is said to do some jitter reduction, but the I2S sounds better, so I dont really believe it.

 

Even Gordon says that reducing the jitter at the source, before the stream gets to the DAC, as early as possible is superior, and this is exactly what the Pace-Car does.

 

And good luck finding a DAC with ASRC that does not sound different with every digital cable you try. They are far from being intolerant to input jitter. Even Gordons Async USB interface DAC customers use expensive USB cables because they sound better.

 

Why do you guys piss me off like this? One day, you will get it.

 

Steve N.

 

[IamKirk]

I don't have the technical knowledge to dispute anything these hardware makers say but they all likely make viable solutions for different audiences. I think most of the flack Steve takes comes more from his personality than his actual product. I'm not debating whether expensive solutions are worth the money because I likely can't recognize the small gains achieved after you move beyond a certain echelon of audiophile gear. The article about Michael Fremer posted here seemed to put forward a fairly rational argument about this whole price/performance debate. Those willing to pay large sums of money for small gains in sound quality may very well be able to hear the difference. It's really become more of a character/(background in the industry) argument for a lot of folks here.

 

 

On another little side note...

I really dislike the whole idea of high end audio designers talking down to people who don't understand their product. Granted, it can be flipped the other way, with people bashing higher priced products without actually trying them. There usually seems to be the argument between the cost/performance side and the sound quality at all costs side.

 

[Dennis L Jacob]

Well if most of what Steve is saying is true; about his gear playing through an Apple computer with iTunes is inferior to what he is hearing through his PC makes for one conclusion. If you are using a Mac and iTunes, don't use Steve's gear. Does that mean his gear isn't good? NO it doesn't. It just means use his gear with PC's instead. End of discussion. Back to Amarra discussion.

 

[Opusover21]

This thread is weird...! I think what we are all strugglig with is the fact that the Amarra software somehow changes the data....what, how...is all a mystery...that coupled to the fact that we all assumed it was bit perfect therefore its either right or wrong? it should be black and white? taking out all the extra's suach as volume control, enhancements etc? So in these confusing times Steve when you state “two bit-perfect identical rips that sound entirely different” & “these sound virtually identical”....then it all sounds a bit vague..is it identical or not? possibly, definitely maybe.

 

I think we really need Sonic/Jon to give us more insight to exactly what this software does....and exactly how it works as this is going round in circles.

 

Jon?

 

[Audio_ELF]

I have a question which isn't about Amarra, but is about the direction this topic has gone. Steve (and others) have stated that iTunes, Amarra, Foobar, etc have sonic signatures beyond what can be explained away with jitter. I'm not going to argue one way or the other on this matter. I do however have a related question.

 

Do CD transports also display a similar sonic signature? (As I say ... another can of worms)

 

Eloise

 

[One and a half]

it's about time we had a label on the side like food labelling, so we can make an informed choice.

 

Music: 98%

Plastic: 1.2%

Defective pits in the CD: 2ppm

Jitter: min. 200ps

May contain traces of nuts

 

 

[Audio_ELF]

No, really I am

 

[cfmsp]

 

opus,

 

I think Jon/Sonic have already given as much 'insight' as they can (or possibly, as much as they are willing, for competitive reasons). We've heard that they do manipulate the data, but with great care and great precision (e.g., more sophisticated, more accurate algorithms), and we've heard that they think they do not manipulate the data any more than (they think) is required by 'modern' computer systems.

I honestly don't think they know any more than they've already told us.

 

We've also heard that the HDCD indicator light is NOT necessarily a good indicator of bit-perfect-ness - for the reason that anyone who is using 'care' (i.e., knows what they are doing) in manipulating the data would do so without trouncing on the bit that HDCD relies on, as opposed to how the HDCD bit would be over-written by unintentional up/down-sampling, for instance. It is this latter instance which the HDCD indicator can report on, in other words, HDCD indicator being 'off' for an HDCD disk indicates lack of bit-perfectness, BUT, HDCD light being 'on' for an HDCD disk does not guarantee bit-perfectness.

 

What I think is missing is a bonafide bit-perfect test of the data output by Amarra compared to the input data. Unfortunately, that test can only be done by someone with the full version of Amarra, given the silence interpsersed in the demo version. Jon says that his clients do this regularly (which I take to mean they do so with Soundblade). Neither Jon, nor Vincent, nor Tuckers, nor Alan, nor any of the other guys who have been touting Amarra for months claim to have done a test that compares output to input for bit-for-bit match.

 

To get to the bottom of this, ideally, as I outlined in an earlier post, you'd want to be able to prove (e.g., with ABX) that the instances in which Amarra 'sound' better than other software are the same instances in which the software is outputting bit-perfect data (if Amarra are actually capable). I'm relatively convinced that Amarra (should) sound better than garden variety players when 'the sliders move' (as Peter says).

 

just my two cents

 

 

 

[Opusover21]

I guess taking a step back this is the as any other audio product...because one cd player sounds better than another we dont immediate jump into a lengthly technical debate as to why that it. The "bit perfect" historical statement is where I feel a bit miffed in the fact that if its producing a perfect output in simplistic terms then the DAC's are where the differences or at least 90%+ of the differences should be heard. So the whole challenge to iTunes bit perfect playback is where I go looking for answers....it's that niggling doubt thats caused me to think whats going on? I always thought iTunes/Mac playback was as good as it gets? maybe its just time to sit back and enjoy the differences without the analysis....but we know thats not in out nature now is it?!

 

[cfmsp]

 

"Why do you guys piss me off like this? One day, you will get it."

 

Steve, we're not trying to piss you off.

 

I will admit to not yet 'getting' why one would want a DAC with no jitter suppression.

 

Admittedly, it may be a good tool for someone who sells reclockers, as you say, but it seems disingenuous (to me) for you to state that iTunes is broken, when a 'normal' DAC with jitter suppression would handily address the 'jitter', never mind that software (such as iTunes) doesn't even produce jitter in the first place.

 

This is the kind of issue that drove Tim/Ashley/et al absolutely nuts. For the vast majority of people, iTunes on a Mac with a decent DAC (WITH jitter suppression) is all that is needed for computer-based musical enjoyment, and yet your posts would seem to indicate the opposite, indeed, you seem to be on some sort of rant/vendetta against the Mac/iTunes (as Gordon claims to have predicted).

 

 

 

 

[The Computer Audiophile]

"HDCD indicator being 'off' for an HDCD disk indicates lack of bit-perfectness, BUT, HDCD light being 'on' for an HDCD disk does not guarantee bit-perfectness."

 

Well said Clay. Right now it's a very easy indicator of problems and perhaps the best one we have for 99% of applications.

 

[The Computer Audiophile]

Steve - Your comments here are somewhat rude. They may be exactly how you feel, but they are not adding anything to the forum.

 

Also, I'm beginning to think your new slogan should be:

 

Empirical Audio: Where There's Mystery There's Margin

 

It seems like you are creating confusion and drumming up possibilities of sonic degradation that only your products can reveal or prevent.

 

New Rule: No more mentioning your products unless asked a specific question about a specific product. Period.

 

 

 

[blueixus]

Chris

 

It is your forum so you make the rules, but I strongly feel that Steves comments are some of the most grounded and sensible on this forum.

 

He has a real understanding on the issues involved which are I believe are considerably more complex than many think. As a forum user I want to hear from Steve et al because they are at the cutting edge of this new world.

 

I also suspect that there are no easy answers here and that far from making matters more complex Steve comments show the shortcoming in many other Maunfacturers products.

 

Specifically the interface between computers and Dacs is a fraught area. Interestingly the much lauded Amarra has limited the equipment it works with and I suspect this is because they realise that this is not quite a plug and play market.

 

I do not own any of Steve's products but the Off Ramp attempts to work in this area. I currently use a DDDAC USB to I2S interface (and a hidden SPIDF output) which i think is a (very) poor mans Off Ramp! as well as a very highly spec'ed Audionote Dac.

 

I bought the DDAC specifically to begin understand this interface and it has made me aware that their are many highly significant variables beyond 'bit perfection'.

 

Steves comments may have been strong but I feel if we are all to progress they are very important.

 

Please don't take this as criticism - the site is fantastic, but made all the better by many opinions

 

 

 

 

 

[The Computer Audiophile]

Hi blueixus - No offense taken at all. Thanks for the feedback. I'm only looking to curb manufacturers from pushing their own product either directly or indirectly through comments that lead one to their products only. I agree progress is made with opinions from all sides whether simple or very complex.

 

Thanks again for the feedback.