Mac Server Processor Performance

I have been wondering about something for a while and it came up again in a recent conversation so I thought I would try to open it up to some of the people on the forum who are more computer savvy than I am.

From the beginning (all of a few months ago) it has been my impression, and stated in several instances, that the function of music server requires little in the way of processing power. At present I am using two servers, a current generation Mac mini and a current MacBook, both running Leopard. During a conversation about file formats Gordon Rankin told me that, to him, the difference between AIFF and Apple Lossless was more noticeable on a MacBook than on his bigger Mac, a Mac Pro, I think. It was a casual comment but the implication was that the better processing capability allowed the compressed signal to be uncompressed more correctly. I didn't worry about it too much since I had determined that I could hear the difference using AIFF, albeit fairly subtle, and had decided to rip everything in that format.

Yesterday I was talking to Dave Gordon with Audio Research about their new DAC7 USB piece which is in the pipeline. I have heard really good things about the sound of the prototype from a couple of people I respect. We should get our turn at it soon. While we were discussing it, Dave mentioned that they had used a MacBook and a higher end Mac computer, sorry but I don't recall the model. He said that the sound was noticeably better using the more powerful computer. I asked him about what file format he was using and he said lossless. I had to take a call and we were unable to discuss it further, but I am going to ask him to see if he hears the difference between the two if he rips in AIFF.

Any thoughts anyone? Obviously a mini or MacBook can be very good but I am very curious if this is indeed a variable when you are looking for optimum performance.

Sorry I don't have more specifics. I will try to get them.

Rick

You said: "It was a casual comment but the implication was that the better processing capability allowed the compressed signal to be uncompressed more correctly."

I have a hard time believing this. After all, we are talking about a simple format conversion. As long as the underlying algorithms are the same, the result will be the same, irrespective of what Mac is used. Now there is a faint chance that floating point operations are carried out with different precisions on different machines, but I almost doubt it. The proof would be to take a file from a CD, convert it into Apple Lossless, convert it back and compare it against the original file. If the "Lossless" is any indication, I bet the double-converted file will be identical to the original (except for some header info, maybe). It would be a disaster if that would not be the case.

If there really are any differences between different Macs, they must lie somewhere else. Many (most) people can't hear the differences between a high-bitrate mp3 file and the original on their own systems, so I would be interested to find out what equipment was used in the comparison. Differences between Apple Lossless and AIFF??? My, my. I wish I had your ears and/or equipment.

Best - MM

...between playing back the same piece in the two formats. I use a MacBookPro Core2 2.4GHz w/ 2GB RAM and I notice it. I have to agree with the pro's here.

There are one or two other threads on C/A that discuss this somewhat. I believe that one of them is titled "Uncompressed v. Lossless Compression."

What is going on is this: When you play back an AIFF or any other uncompressed source, the computer merely has to spin-up the data and pass it to the DAC. In the case of LOSSLESS compressed data, the computer must first uncompress that data and THEN pass it to the DAC. More system resources are being consumed in the case of playing lossless compressed files. This is NOT the case with lossy compressed files (though they have their problems too), such as MP3. I don't know what the exact ratio is, re: the higher amount of the computer's system resources that are being used (2X?) when playing back lossless. I do know that I can hear a difference in the fine detail on many (otherwise identical) files.

markr

Download FLAC, convert it to AIFF - it B better

The beauty of this is that you are not required to believe it. I am not stating it as a fact; that is why I threw it out there. Two very competent people have brought it up and I found it interesting. I do find that, for myself, it is helpful to experience something before I decide to believe or disbelieve it. If I do experience it, then I have to look for explanations. If not, I discard it.

My ears are adequate. I listen criticially every day. It's what I do for a living. I have never encountered someone who can't hear the difference between MP3 and the original, uncompressed signal on any kind of decent recording. The degradation seems to be even more obvious on good recordings than poor ones. I have run across people who could care less about the difference, but that is another story. This has always been in the context of a competent system.

The system I used for the comparison was my own. Briefly, a Wavelength Brick USB DAC, an Audio Research REF3 preamp, Classe CA-M400 monblock power amps, Vandersteen 5A speakers.

Thanks for the reality check, Mark.

This is certainly an interesting topic and one that will never be "decided" one way or the other. I think we need to make sure we keep the two parts of this discussion separate.

1a. The file format lossless v. uncompressed AIFF

2a. Realtime playback of lossless v. uncompressed AIFF

1b. To be honest I am somewhat bored of discussing the merits of lossless v. uncompressed AIFF. This is something that everyone with a computer can try for themselves free of charge and make up their own mind using their own system. I've made up my mind and I'm sticking with AIFF. Why do any conversion or compression when you don't have to? I actually just purchased 5 TB of disk and my goal is to fill it up with uncompressed music. Now that will be fun!

2b. This is the main issue when considering the actual computer used for playback. I think it is pretty much a fact that uncompressing lossless files does use more system resources than playback of uncompressed AIFF files. But, is this a difference that has an audible effect on the music? I could take the easy way out here and tell everyone to try it themselves and make up their own mind, but it is a bit harder to purchase the necessary equipment to conduct the listening sessions. On one hand I really find it hard to believe that computer power makes a difference when playing back lossless v. uncompressed music. Back in the late 90's I am sure it would have mattered but now any Mac can playback lossless and uncompressed while taking a nap. Does it really matter if uncompressing lossless music uses 10% of a 2.0 GHz Core 2 Duo, or 5% of a 2.8 Quad Core Xeon? It just seems pretty odd that it would matter enough to make an audible difference.

On the other hand I think the people Rick has discussed this topic with are some of the most knowledgeable in the industry and spend their lives listening to music both live and reproduced. Rick correct me if I am wrong but I bet your conversation with Dave discussed a certain event in Florida with some of the best known manufacturers in high end audio and the guys from Stereophile? I think their opinions must be respected. Without the equipment to test this I don't know that we can honestly say they are incorrect.

So, I am skeptical, but optimistically seeking information about this. If it turns out a Mac Pro helps reproduce audio better than a MacBook the only downside is cost and a little noise from the machine. But, we would all benefit with the knowledge of what we need to use to improve our systems. There is certainly no need to fight the results of anyone's opinion or startup the whole lossless is lossless conversation. This is just as counterproductive as telling art connoisseurs that red is better than orange. Hopefully we are on the road to better music reproduction and more information is never a bad thing!

to markr: you refer to more system resources being used for playing back compressed files. That is certainly true, but has nothing to do with the contention that these data are somewhat different from uncompressed AIFF, as the OP surmised. As I said before, any differences come from things other than the results from uncompression.

to rom661: Your system is quite nice. How many people do you think have a system that's comparable to yours? One percent? 0.1 percent? I would say probably way less than 0.01% That's exactly my point. The differences between Apple Lossless and AIFF files will not matter for the vast majority of listeners. Among those, you will find many who can't hear the difference between 320 kbps mp3 and AIFF on their systems.

to Chris: yes, some of these pros are widely respected, but they are also looking for tiny differences with an oscilloscope, have a system that may reveal tiny differences and that costs tens of thousands of Dollars, and some have a business that depends on convincing people that there are differences that anyone can hear if they only bought their equipment.

Perhaps it would be a great idea to conduct a little survey of who your audience is. Are most of them "audiophiles" who have the equipment and training to reveal differences in the amount of jitter of different DACs, and similarly subtle things? Or are most of them people who have good, but basic setups, no perfect room treatment, who enjoy listening to music more than listening to equipment (sorry for using that old line) and who are interested in storing their music on a computer?

Best - MM

Yes, your assumption is correct regarding the event. I agree and was actually not debating your 1a since I am comfortable with my conclusions. I tend to agree with your overall assessment and frankly would never have given it a thought if Gordon had not mentioned it. Having the second, unsolicited comment about the same issue made me start wondering. And Dave seemed to feel it was a more significant issue than Gordon. Maybe it is a Gordon thing and if you are a non-Gordon it doesn't matter...

In some ways this reminds me of the learning curve when CD first came out. It sucked but it took a while to figure out why when conceptually we were told it should have been glorious. The industry started to identify variable and issues like jitter, pre-emphasis (remember that?), etc.

Perhaps if Gordon is monitoring this he could help. He is certainly more qualified from a technical point of view than I am by a huge margin.

I would like to know. I'm trying to think of anyone I know who has a big daddy Mac I could borrow for a day.

Mark messed with me with his last line: Rip to WAVE and convert to AIFF. Is he messing with me? You really shouldn't say that to someone who just ripped 7,000 plus songs....

MM

You are making some incorrect assumptions about me and what I do and I would prefer that you didn't. We see the "listening to the gear" syndrome on a regular basis and it gets in the way of what we are trying to accomplish for our customers. I don't suffer from it, thank you. I got past it a long time ago. Music is my passion. If a perceived difference isn't musically rewarding I don't care about it. You can buy a good, musically rewarding two channel system in my store for about $1,500.00, just for the record.

The percentage of the population who have high resolution systems has nothing to do with the topic. I do have a number of customers who own systems capable of very high resolution. I care about the question for that reason and also because I am simply curious.

I mentioned in my response to your first post that the vast majority of our customers can hear MP3 on moderately priced systems. Again, though, that is not the topic here.

Some very expensive gear is superb. Some isn't. You paint with entirely too broad a brush with your comment about manufacturers. It certainly is a business and you have to pay the bills, but many high end manfacturers are as passionate about what they do as their customers. In general it is not a field you go into to get rich. I don't know of any of these companies who determine whether or not something is significant based upon a scope. They certainly use measurements but they also do a lot of listening.

We presently demonstrate a system in the store using a $1250.00 DAC and a MacBook. It is competitive with a $4,000.00 CD player. We do plan to show something that we hope will sound as good as our reference $10,000.00 player, which is musically the best I have ever heard. I got involved with this forum to learn and help our customers pursue the dual goals of good sound and the convenience of server based music.

I suspect that most of the people on this forum care about music and sound quality, regardless of the price of their system. That qualifies them as audiophiles. I am sure that Chris had that in mind when he named the forum Computer Audiophile.

I would really appreciate it if you would respect the topic of this thread. If you would like to express your feelings about other issues, you can start one about them; that is the beauty of the forum.

Thanks for your consideration. Best wishes to you, as well.

rom661: With all due respect, but I don't think I have made any assumptions about you, let alone incorrect ones. Please read my comments directed specifically to you. I have not denied that there might be differences between different Macs. I simply don't agree with the explanation for why that could be. And I honestly wish I had a system as nice as yours that might reveal such subtle differences.

I didn't know that you are in the business of selling audio equipment, so I can understand that you felt addressed. That was not my intention. My comment regarding the pro's was directed at Chris, not you. I respect vendors who work with their customers to give them the solution that works best for them. I'm glad you're following that mantra.

Finally, concerning respecting the topic of the thread, to avoid any confusion and to get back on track, I will summarize my opinion, which I made in my first response: I cannot believe that Macs with more processing capability uncompress compressed data more correctly than Macs with lower processing capability. I think the burden of proof is on the ones claiming that's the case.

Best - MM

Exactly Rick.

I really enjoy hearing everyone's opinion. We all have much more to offer to each other than arguing about what is better red or orange and then suggesting someone is color blind if they don't agree with you. I'm not saying that is what's going on here, but the whole tone of this is really taking the fun and interest out of the topic. We are all after excellent sound aren't we? Who really cares if one guy likes to add 1+1+1+1+1+1 to get to 6 and another guy multiplies 2x3 to get to 6. If their goal is 6 then woohoo!

So, keep the personal opinions coming and keep the topic pleasurable.

Hi xvimbi - Thanks for the collegial response. It really keeps this forum much more enjoyable than many others.

Now back to the topic at hand!

(and Chris) .... that it isn't the 'processing power' that makes the difference that I hear between playing lossless vs. uncompressed audio. It is merely the most convenient argument - the one most often heard. I posted my computer system basic specs to emphasize that point. I suppose that I could have expanded on the point a bit, but I was on my way to work, and was in a hurry. I should have taken more time, and I apologize for not doing so. You and Chris are correct in your beliefs that the "processing power" of todays machines shouldn't be a problem.

I didn't use sophisticated test equipment, to come to my 'conclusion' about the difference between the two formats. Just my audio system, computer and my ears. True, my audio equipment's quality is probably in the top 5-10% of those available. I know that you addressed that point to rom661, but it is true in my case as well.

I didn't go looking for this opinion. I stumbled upon it by having a Mac (where FLAC playback is not fully developed yet) and having a need to play back FLAC source material in iTunes, which does not support FLAC on Mac. The best option seemed to be to convert the FLAC to AIFF for easy play in iTunes. It was only after comparing those two formats that I concluded that there was a difference. Only my 54 year old ears helped me form the opinion that I still hold: that the two formats sound subtly different - though not disagreeably so. However, I absolutely believe that properly done lossless audio contains all the audio data intended by the artist.

In the end, all that matters here is if you like what you hear. Stick with that, and you cannot go wrong.

markr

Oops. Sorry Rick. I definitely didn't know about the 7,000 ripped thing....... in the end, t is probably just a Gordon & Mark thing.... It really is a very subtle difference, and as I said: not really disagreeable.

I ripped them in AIFF but the ripping to WAV and then converting is new to me. Has there been a topic on the forum about this? If worthwhile, I can consider that for future ripping or selective re-ripping.

Thanks

To be honest ripping and converting does not sound like a good idea to me. These are both uncompressed formats and converting only introduces potential problems in my opinion. But if it sounds good to people then hey I'm all for it! I just won't be doing it until I can hear the difference :-)

mark - I'm very skeptical here :-)

You write: "I think that you are probably correct xvimbi (and Chris) .... that it isn't the 'processing power' that makes the difference that I hear between playing lossless vs. uncompressed audio."

It might be, though. What I said was that I don't think that the data from the uncompression process will be different for different Macs. I think the result is exactly the same and independent of the processor.

Now, how fast that conversion occurs in a given machine may well have a bearing on the resulting sound, and there might be a difference between a busy laptop and a high-end Mac-Pro. It would come down to how well the data stream is synchronized when it leaves the machine. Perhaps this is where re-clocking might make a difference. There could be a lot of reasons for imperfect timing, e.g., differences in how fast data are read from the hard drive, how fast they get delivered to the processor, how fast they get passed on to other chips in the pipeline, which port they are getting routed out of, etc. The situation is very similar to writing a CD or DVD, where it is imperative to have a steady, uninterrupted, constant-speed data stream to the writer. If writing CDs/DVDs may cause problems on certain Macs, I can believe that streaming music may be imperfect as well. That would be some sort of buffer-underrun error. I don't think the processor itself will be a bottleneck, unless the machine is really overwhelmed. I would surmise that issues like how much general I/O and network traffic is going on at the time play a bigger role.

I'd be interested to hear if these differences disappeared when a suitable re-clocker is added to the system.

Best - MM

Well given all the fun of ripping a big pile of cds, I can only imagine the ecstatic joy of ripping to then convert. That ought to rank right up there with yanking hangnails or slamming fingers in car doors!

.... but now I'm confused.

I don't know where AIFF --> WAV came from here. Help?

markr

Hey Mark - You know I was a little confused when the conversation urned to this aiff to wav conversion and I probably got caught up in it like the others. We probably misunderstood a comment from you and the snowball only got bigger.

Anyway, I will be happy to forget about the conversion part of this topic!

After your signature on your first post on the thread you added the line

Download FLAC, convert it to AIFF - it B better

That's why I asked if you were jerking my chain. Not that, you know, I would suspect you of chain jerking or anything....

Rick

You are among many here who understand the whys of the computer technology far better than me. I'm good at the audio. I am working only by instinct and what common sense I could muster. I used the term processor power or something to that effect generically to describe a more upscale mac than a Macbook. I could understand the potential for a difference more if my Macbook was busy with other processes, but it was purchased and used solely for the server function. The only other program loaded is Signal. I have not personally had a chance to make the comparison. I have plenty of stereo gear on hand but not so many computers, especially Macs.

As I said early on, my assumption would be that there wouldn't be a difference, although I hadn't really considered it at all. The two people who have mentioned experiencing are very good at what they do and neither has a motive to encourage someone to spend more on a PC. Just the opposite in fact. I would much rather minimize the server expense and free up budget for a better DAC. One of the two is also a very experienced computer engineer in addition to his audio expertise.

I didn't get a chance today but I will see if I can get more details from one of them tomorrow. Even if it is true, my follow up question would be if there is still a difference when ripping in AIFF instead of Lossless.

That 'signature line' was meant to indicate that lossless compression *files* - FLAC - will convert to uncompressed audio *files* perfectly. FLAC, or any lossless, doesn't sound to me to be as perfect when played as native FLAC (or whatever...) though. The 'sound' of lossless ain't that bad, but I like uncompressed better. Uncompressed audio is what I am putting on my newest hard drives for serious listening purposes. I prefer the open-source sort of lossless too - That is just because it IS open source. I have no other reason for that.

Also, I have no preference between the sound of WAV vs. AIFF. They sound the same to me. Back in-the-day though, the audio programs on one computer platform didn't necessarily read and write all of the types of audio from another platform - and I usually have more than two types of platforms lurking around here - a quick count tells me that I have 5 different ones available to me right now.... Conversion was a necessary evil for me, and I always had good luck with it. It isn't anywhere near as necessary nowadays. For instance, iTunes reads and writes both WAV and AIFF on either a Mac or PC. Virtually all pro apps do the same now, with the exception of the (originally) Digidesign 'SDII' format (Sound Designer), but that is changing too.

I know the album art thing is a biggie to some with regard to WAV and metadata handling, but I figure that: 'HEY! I am using a computer (computerS, in my case) here - I can just surf for any info I want about a piece of music, when I want it. And I can get much more detail from fan or band sites than could be fit into the format offered by any jukebox program.' .... But then, I am sort of a maverick in most regards. ...and that's not just my opinion ;^)

MAN! What a messy post! It is too late, and I am too tired to change it now. So be it.

markr

nothing controversial

For a minute there I felt like I was travelling cross country and just realized I missed my turn about 500 miles back.

I have to confess some of the formats still confuse me, especially regarding what is possible with a Mac re: higher resolution formats. Rhetorical question, don't want to divert things.

Best

Rick

Just a few thoughts on the comparison between a Macbook and a Mac Pro. There are quite a few differences in the overall architecture of those machines that could explain differing sound quality besides the CPU. I believe there was mention on this site that the USB ports on the Macbook use a shared bus, which could introduce jitter to the DAC. The internal clock in the Mac Pro is likely higher quality than the Macbook's, although I think that's a less likely explanation.

As to differences between AIFF and compressed, I'd have to assume they're not buffering it optimally (optimally would be to uncompress to AIFF in memory). One test that would be interesting is:

Play an AIFF file on the Macbook with no other programs running.

Play an AIFF file on the Macbook while a browser is open playing a flash movie (or something to increase the CPU load on the machine).

If you can hear a difference, then a faster CPU probably reduces jitter when passing over USB.

Ask Chris for my email. I can spare some time here and there!

markr

Be patient, the SPAM filter will kick it back to you but will give me the opportunity to accept your email address after a day or so.

... here before! Welcome.

These are all valid suppositions, and I have one more that you missed.

BUT my MacBookPro doesn't share USB ports. They are all independent. Confirmed. That supposition is moot in my case anyway, as you will read in a second. I believe its MB to be the last rev before Apple upgraded the CPU, video and trackpad recently.

I too find the clock argument 'less likely'. The reason: I do pro audio and MIDI, and then syncing that data to video running on this thing and have had *no* issues.

The buffering issue is the likely 'hit' in my mind: There aren't any fully optimized FLAC playback programs for Mac yet IMO & I don't and won't do Apple or Windows lossless if there is any way to avoid it - Open Source is my preference when possible and practical. I haven't experimented with my FLAC files on my PC yet though. There is a GREAT audio setup on that machine too...... I'm a pretty busy guy and converting compressed audio downloads to AIFF (sometimes WAV too) is what I'm doing anyway. Hard Drives are CHEAP. Hell, I put AIFF on my iPod. It just isn't an issue with me. I MANUALLY sync my stuff to and from the iPod.

- Also re: optimization - I have had as many as 27 tracks of 16/48 audio with video running all at the same time in Logic Studio -HUGE program- (or as a 7.1 mix and 720p video using Soundtrack Pro) and have experienced no issues.

My AD/DA box is firewire and also does not pass jitter by design. It totally ignores it. It reclocks at the output. .....I'm sure there is SOME in there, but I can't hear it...

The one thing (you missed) about laptops is the (usual) 5400 rpm internal HD. BUT, I download all my FLAC to the external firewire HD (7200 rpm) and the "issue" happens when I use that HD too. Maybe my ears just are sharp in a particular area that causes me to hear this..... 'flat/closed' thing that I hear (it doesn't sound like it has as much 'air' as playing uncompressed audio does to me.) I don't know -

Just remember that if you like lossless played back as lossless, that is a GOOD thing. Heck I LOVE lossless (FLAC). It has afforded me some GREAT 24/96 downloads that I now listen to as AIFF!! Lossless is the wave of the future. Only just for downloads and archiving in my case.

markr