Computer sounds flat!

[mudlark]

The variables when playing music from a computer be it PC or a MAC are quite substantial. Windows does its own EQ. which you need to bypass, not something I have bothered trying as a load of hassle. I did some work around recording and mixing some years ago at a very low level whilst on a music course, and with a PC the best thing to do was to migrate the music processing off the sound processor of the MLB (logic board) and onto a dedicated external processor that is set up for the job. This gives best results.

With respect to digital music it is important to understand its limitations in relation to what you are listening to.

Lets start at the top, a FLAC file of an Album is about as close you can get to the original digital master recording, remember that if this has been created from an analogue tape there may already be some loss or introduction of change from the original. So replay of a FLAC file through the right set up will give the best result. Ignoring amplification and speakers etc.

The next step down is a CD, this is 16 bit and is compressed from the original FLAC and is somewhere between a quarter and a fifth the amount of digital data compared to FLAC.
CD's have there own issues, jitter being well known and the ability of the Laser to retrieve the information from the disc, and the DAC to be able to accurately convert to Analogue. Meridian over the years have extensively developed CD players that extract as much as possible and get the best out from a CD, as others have done. There are trade offs on the conversion process, where something that is very accurate and fast can sound too sharp or clinical, whereas a slightly different approach can sound softer and more musical (dare I say it more akin to vinyl), the reason for so many CD players and DACs out there.

The next step down is AAC lossless or MP3 hi-res, now you only have to look at the actual file size and compare with a CD to wonder what is going on. Lets say a average CD is about 600 megabytes, the same in lossless will be about 100mB a factor of 5 yet again. My understanding of the compression that is going on here is that it is actually removing aspects of the musical content, to gain the size, and obviously that lower the size the worse it is. I was told that the compression of MP3 and AAC and its reproduction actually relies one your brain filling in the holes that are missing using your musical experience, a bit of trickery the brain does.

So this for me this raises several issues. I in fact went through a process of digitising all my CD's a few years ago to my computer and creating a digital music library, putting it on a drive and hooking up to various media players and finally ending up with an Apple TV (not streaming). Then one day a friend recommended some music for me to try out, so I downloaded and started listening, and I was shocked. It sounded weird and horrid. I thought I had a download issue and repeated it, still the same. Then I checked my system and listened to some other music, all OK there. So initially I thought rubbish recording/music, then I was lent the CD and listened again and hey it was good, not muddled, and quite good. So what was going on?

Well here we come back to the brain trickery bit, the music I was recommended and downloaded was outside my previous musical experience and my brain had no terms off reference to fill in the missing gaps, new instruments and style and so on, so it sounded weird. I investigated this further, by listening to MP3 albums of artists I had not listened to for a long time and once again I sometimes felt it was lacking somehow. But when I played the CD it was as I remembered.

So have stopped playing CD's, I decided to stop the rot and go back to playing CD's and brought a secondhand Meridian 508. I have even thought of going one step further a back to vinyl, but there are issues when you already have a large CD collection and Vinyl is not without it's own compression issues, but not the same as CD.

So the answer is not simple, in fact I fear for current generations who are probably completely unaware of the impact on reproduction quality, and hear music not as it originally recorded or intended to be heard, and just believe that what comes out of a computer, MP3 or iPod is right.

I know this explanation is not technical, I do not intend it to be, more an insight and a thought provoker, and I am sure that it will raise or sorts of arguments. But I will say this, the less information you have to input into a processor(CPU) from the original then the less likely you are to be able to re-construct the original, no matter how good the algorithms and programming is.

Much of this post is anecdotal and is wrong in some respects. At the end of the post I am wondering "what was the point of that?".

The first paragraph is of no consequence at all.

Flac files contain all the data from the "master tapes". The master tapes if analogue are to represent a varying continuous signal. Digitising these tapes at 24/96 will result in the loss of nothing that is audible. I digital in the first place, I don't understand your point at all.

CD quality is fine. When you use the term compression here you don't mean compression.

CDs can be mastered to produce a source with inaudible "jitter". In any case when lifting the data from the disc we know what the data rate should be so it is possible and routine to recreate the data that was put on the disk.

AAC lossless isn't a step down as it is as good as wav, flac and any other uncompressed file.

What is mp3 hi-res?

Etc etc.

If you are going to suggest you know what you are talking about I suggest you use some more accurate language when you present your "knowledge" to others.

I have to say that my first thought in reading your post is "what?".

 

[Pandapple]

all the editions. i own them all.

All 2 of them?

You must be reading a different book or know a different Bob Katz to the rest of us anyway

Bob is Mr Golden Ears, and his book encapsulates this perfectly. This includes the chapter on jitter.

 

[CVentura]

CVentura, as has been said several times, jitter has stopped being a problem for many years. What you are hearing is in your imagination.

What I suggest is you head down to the pub, invite some nice lady back to your place and get her to swap thing around with you blindfolded. You won't hear any difference, but you never know how the evening may progress - and maybe you'll get over your hifi obsession, realise it's only ever been about the music, sell most of your gear and move forward in a deeply fufilling relationship.

Thantx mate, as I said before, been there done that... I am a lot happier now. Besides I don't want no boad messing with my toys are you nuts ???

 

[CVentura]

Macs possibly sound better because there's less RF on the USB signal lines and because they tend to have separate regulators for the 5v lines coming from the mainboard. Whenever I've trialled with and without a USB filter (galvanic decoupling) there was always appreciable improvement with the PC, but none I could hear with the Mac. draw your own conclusions.

I'm firmly in the camp that 'bits is bits' but there's an awful lot more to being 'in spec' than just the timing of the data.

Well bits are bits, if you cant have ones bigger than others the only other variable is .... Timing !!!

 

[CVentura]

Interesting- I've never heard that before. Either way A good argument for sticking to toslink Ethernet or wireless. Personally I've always thought that USB is a dumb thing to use for audio bearing in mind the unnecessary 5v connection.

I am with you on that. Adam must be a genius. If that idea was mine, I would make a fortune out of it

 

[CVentura]

That's a perfectly valid point of view but it doesn't sit very comfortably with the suggestion that regulation of the 5v line in the computer affects the sound of the dac. Obviously the 5v is potentially useful for bus powered devices, but for most hifi dacs it doesn't seem to serve any purpose at all.

Curiously I have got the impression that most dac receivers do have the 5v line connected which just seems to be asking for trouble. I have a feeling that this is because stricty if the line isn't connected it doesn't comply with the USB standard. I asked PS audio whether they used galvanic isolation and they said no- because galvanic isolators will not cut off hf noise just low frequencies like mains hum and they tend to increase jitter in the signal (although it's not obvious why this should matter with asynch usb).

I'm not suggesting this lot is insurmountable, but it arguably raises more problems than S/PDIF has, bearing in mind that the whole interfaced jitter thing is a bit overblown. It seems pretty odd unnecessarily to connect a dac to the power line of a noisy device like a computer.

Nooooo ... you said "Jitter" ? here we go again ....

 

[CVentura]

There's no logical disconnect between the two statements.

1. Cleaner 5v power is a good thing if it is used by the dac.
2. Cross contamination from adjacent wires during transmission probably isn't an issue that affects the data integrity.

There's nothing to say that noise on the 5v line might not make its way where you don't want it once inside the dac if it isn't correctly isolated/filtered.

Of the two dacs I've had that used async protocols, Young & a Weiss 202, both of them use galvanic isolation on their input and both of them couldn't give a rats ass what cables you connect up to the source with (assuming usb, firewire of spdif).

I do agree with you regarding attaching a dac to a pc, it's just crazy, that's why I think inline filters and things like the SoTA usb filter supply board can make a real difference. Much more so than **** wire and bullshit external clocks.

BTW, the bullshit external clock, has the responsibility to minimize time alignment errors while the signal is converted from USB to SPDIF. You still have the same conversion without the Bullshit Clock but being controlled by a less accurate one. But your the engineer not me...

 

[Teddy Ray]

All 2 of them?

You must be reading a different book or know a different Bob Katz to the rest of us anyway

Bob is Mr Golden Ears, and his book encapsulates this perfectly. This includes the chapter on jitter.

he spouts some bullshit, yes, but there are some nice explanations in there about digital audio.

I know about Bob Katz. Ive been interacting with him on fora for a decade, and email, phone couple times, etc

 

[sq225917]

BTW, the bullshit external clock, has the responsibility to minimize time alignment errors while the signal is converted from USB to SPDIF. You still have the same conversion without the Bullshit Clock but being controlled by a less accurate one. But your the engineer not me...

What you are actually doing is taking a USb data stream converting it to Spdif via the addition of an external clock then feeding that to a dac that then takes the incoming spdif and strips the timing info passing the data through a FIFO RAM buffer, and reclocking it with a local clock inside the dac. So it's pretty much an utterly pointless step to try and give it a better clock upstream as it is completely dismembered inside the dac.

It might help remove RF noise coming from the USB output but it sure as hell doesn't reduce the jitter seen at the time of conversion inside your Naim dac- because all timing info is stripped out of the spdif data stream by the internal buffering and reclocking.

 

[CVentura]

What you are actually doing is taking a USb data stream converting it to Spdif via the addition of an external clock then feeding that to a dac that then takes the incoming spdif and strips the timing info passing the data through a FIFO RAM buffer, and reclocking it with a local clock inside the dac. So it's pretty much an utterly pointless step to try and give it a better clock upstream as it is completely dismembered inside the dac.

It might help remove RF noise coming from the USB output but it sure as hell doesn't reduce the jitter seen at the time of conversion inside your Naim dac- because all timing info is stripped out of the spdif data stream by the internal buffering and reclocking.

Your statement is almost accurate. This "Bullshit" clock, is only user on clocking the conversion from USB to SPDIF inside the converter, it has nothing to do with the reclocking inside the NDAC... As you know the Naim NDAC does not accept a direct USB connection as for instance my DCS Puccini does.

 

[adamdea]

Your statement is almost accurate. This "Bullshit" clock, is only user on clocking the conversion from USB to SPDIF inside the converter, it has nothing to do with the reclocking inside the NDAC... As you know the Naim NDAC does not accept a direct USB connection as for instance my DCS Puccini does.

AFAIK the conventional wisdom is that external clocks are only sensibly employed to synch two or more devices so that the clock governs both/all of them.

Using an external clock to time just one device makes little sense because of the risk of talking in jitter in the line between the clock and the device being driven. You want the clock to be as close as possible to the thing being clocked.

I assume from sq225917's post that what you are using is an external clock driving a USB to S/PDIF converter . I think that the bullshit clock points are about the bullshitiness of the external clock not the usb-spdif converter. His point is that the clock driving this is not critical, my point is that if it is critical it should be onboard.

 

[Plutox]

I assume from sq225917's post that what you are using is an external clock driving a USB to S/PDIF converter . I think that the bullshit clock points are about the bullshitiness of the external clock not the usb-spdif converter. His point is that the clock driving this is not critical, my point is that if it is critical it should be onboard.

To compound the issue, there is no point converting from USB to SPDIF unless you really, really must!

Asynchronous USB is the best development to have occurred in digital audio in recent years as it completely bypasses the need for any critical clocks other than in the final DAC. SPDIF lacks any kind of flow control, hence the general need to lock to the incoming stream. Once you have implemented flow control, the clock performance isn't compromised by the need to lock to anything else - a significant performance booster.

And the best place to put that clock? As close to the D to A converter device as modern micro-electronics can physically get it.

 

[sq225917]

Cventura, your comprehension of my post is almost accurate. I never stated that the external clock affected the internal conversion. I'm well aware that it can have no beneficial influence and that the Ndac has no clock input- those were my points.

There's no point reclocking externally that which will be stripped of its timing data and then reclocked internally later in the Naim dac. You do understand this, surely? It's a redundant step, the hiface evo has a bit perfect output, so as long as the signal remains bit perfect the level of jitter on the incoming signal to the Ndac is of no consequence due to the large buffer.

The best place for the best clock is next to the dac, nowhere else.

 

[Pandapple]

there are some nice explanations in there about digital audio. I know about Bob Katz. Ive been interacting with him

Here are the main points of what my edition of Kat'z book says about digital audio.

Tell me, which ones do you agree with?

1. 24-bit sounds better
2. 96khz sound clearer
3. 16-bit dither noise gives a veil over music at normal volumes
4. Different dither algorithms sound different
5. Different high quality filter designs in DACs sound different
6. Pre-echo filter designs sound especially bad
7. 24/96 masters sound better than lower-res
8. 24/96 still sounds better even when the mastering is done all-digitally from a 16/44 digital source without performing any A/D or D/A conversion. Such releases should not be scoffed at, and the 96khz sampling rate is a central reason for the improvement.
9. Jitter artefacts as low as -117dB can be heard
10. Even some of the best jitter-performing professional "jitter-immune" DACs have audible jitter effects from AES/EBU interface jitter.
11. Jitter effects are always audible unless you use a top-notch DAC with a transport that can be externally clock slaved, or buy one of the handful of state-of-the-art DACs that can attenuate incoming interface jitter to the noise floor or below
11. The sound quality from most CD players can be improved by using quality gold discs, better quality burners or $100k mastering cutters, and slow write rates (by way of improving jitter).
12. Most DAT players sound worse than other formats, probably due to motors inducing jitter.
13. Firewire is jitter-prone and makes it difficult and expensive to enginner an audio device that sounds good
14.Most audiophile DACs have better sounding analogue circuitry than most professional DACs.
15. Every single professional D/A converter used to sound so bad they left Bob feeling cold, until a few new designs came out

 

[Pandapple]

add this one too.

http://www.amazon.com/gp/product/02...mp=1789&creative=9325&creativeASIN=0240821009

The less said about Ethan Winer, the better. He is wholly unqualified, and has made huge calamities over and over in articles and talks in the past. I personally would never recommend anything penned by him to anyone, ever.

Katz, Winer and Pohlmann - it's funny you cite them together. You couldn't have picked 3 different chalk n cheeses!

Ken Pohlmann's texts are rigorous, well-written and superb IMO.

 

[avole]

Pandapple, am I correct in thinking you are a qualified audio engineer who works in the production side of the industry?

 

[Teddy Ray]

The less said about Ethan Winer, the better. He is wholly unqualified, and has made huge calamities over and over in articles and talks in the past. I personally would never recommend anything penned by him to anyone, ever.

Katz, Winer and Pohlmann - it's funny you cite them together. You couldn't have picked 3 different chalk n cheeses!

Ken Pohlmann's texts are rigorous, well-written and superb IMO.

are you a recording engineer? Can you qualify your statements?

He is very, very qualified. The intelligent company he keeps (James Johnston, Poppy Crum) respect what he has to say.

In general the people who talk bad about him are the subjectivist, fringe audiophile (or pro audio) "golden ears" types, because he destroys their castles made of cards.

He has been proven right over, and over, and over. Having your fairy tales destroyed simply isn't fun.


its the same reason audiophiles hated Peter Aczel. No nonsense, no bullshit.


I suspect you wouldn't fare too well at Hydrogen Audio.

 

[Teddy Ray]

Here are the main points of what my edition of Kat'z book says about digital audio.

Tell me, which ones do you agree with?

1. 24-bit sounds better (in some cases SOME)
2. 96khz sound clearer (hell no)
3. 16-bit dither noise gives a veil over music at normal volumes (nope)
4. Different dither algorithms sound different (in some cases)
5. Different high quality filter designs in DACs sound different (no)
6. Pre-echo filter designs sound especially bad (
7. 24/96 masters sound better than lower-res (nope)
8. 24/96 still sounds better even when the mastering is done all-digitally from a 16/44 digital source without performing any A/D or D/A conversion. Such releases should not be scoffed at, and the 96khz sampling rate is a central reason for the improvement. (nope, nope, hell no)
9. Jitter artefacts as low as -117dB can be heard(no)
10. Even some of the best jitter-performing professional "jitter-immune" DACs have audible jitter effects from AES/EBU interface jitter. (no)
11. Jitter effects are always audible unless you use a top-notch DAC with a transport that can be externally clock slaved, or buy one of the handful of state-of-the-art DACs that can attenuate incoming interface jitter to the noise floor or below(no)
11. The sound quality from most CD players can be improved by using quality gold discs, better quality burners or $100k mastering cutters, and slow write rates (by way of improving jitter). (LOL HELL NO)
12. Most DAT players sound worse than other formats, probably due to motors inducing jitter. (LOL HELL NO)
13. Firewire is jitter-prone and makes it difficult and expensive to enginner an audio device that sounds good NO NO NO
14.Most audiophile DACs have better sounding analogue circuitry than most professional DACs. (****ing horseshit)
15. Every single professional D/A converter used to sound so bad they left Bob feeling cold, until a few new designs came out

*who cares*

 

[sq225917]

13. Firewire is jitter-prone and makes it difficult and expensive to enginner an audio device that sounds good.

I'm sorry but that is just flat out incorrect. The Weiss 202 is one of if not the best measuring dac made, and it uses Firewire and doesn't cost half as much as many top end studio DACs and a hell of a lot less than most high end consumer dacs. He may need to update his viewpoint in light of more up to date information.

He's right about Weener though, he's an annoying, incessant little gobshite, lacking in technical knowledge about the design and implementation of digital audio products. he should stick to talking about acoustic room treatments that you can design with a $30 Radioshack level meter.

 

[sq225917]

*who cares*

Well you did apparently, enough to quote his a learned source. But now it would seem that either you haven't read him or you were reading another book. Can we expect a similar falling out with your other sacred cows soon?