AVS Forum banner
Status
Not open for further replies.
1 - 7 of 7 Posts

·
Registered
Joined
·
2,093 Posts
Discussion Starter · #1 ·
Hello,


I'm a little new to this HTPC bizness so bear with me. I want to preface my questions/assertions by letting you know that I'm really a vinyl fan, so... However, digital is the way now, so when in Rome I guess.


I've read in several posts about the noise (jitter) contributed by the lower quality components (power supply, etc) in an HTPC setup. I also know that there are different types of noise- jitter being just one of them. The IEEE has addressed this in various ways- twisted pair media, error correction, etc…


The idea in any computer is to allow consistent and reliable operation. As we all know, computer hardware does this quite well (in my experience, if there's ever a problem, 9 out of 10 causes are software-based or the result of user error).


The idea that I'm getting hung up on is the hardware quality issue and how it affects "the sound" of what is ultimately a digital signal. It seems to me that a PCM signal from a CD player, whether it is audio-based or just the latest Microshaft update- it must be delivered in the same manner, reliably. Again, jitter is just another form of noise. It affects all types of signals- not just PCM signals. I don't see how the power supply is a major factor here. If a computer can't pass a PCM signal without jitter-error being introduced because of a low quality power supply, then it doesn't make sense that any other signal would be passed reliably. What makes a PCM signal any different?


I have no problem with the idea that a sound card can affect the sound quality that is ultimately delivered to ones speakers. What confuses me is that a PCM signal from a CD player is no different than any other type of signal. Any computer must pass data from it's CD player through it's circuit traces reliably-irrespective of the type of signal. Ultimately a PCM signal is simply 1's and 0's. Is the interconnect from the CD player to the sound card header the cause of added noise? If so, wouldn't a better interconnect help resolve some of the issues that some people experience?


If it's a back-end issue (the sound card), then it seems the jitter problem may be addressed through better driver support, or hardware modifications to the sound card itself- thus the idea of buying a GOOD sound card to begin with. There are also outboard solutions (the Monarchy DIP 24/96 - Digital Interface Processor being just one), which correct jitter problems.


I just don't understand how the signal passed by a standalone CD player is any better than that passed by a well constructed CD player that you install into your PC. I understand that there may be differences because of the transport. But all things being equal, unless you get into the real high end with dedicated transports AND dedicated outboard processors, I can't understand how you would really hear that great of a difference.


What am I missing here? All responses are definitely encouraged- help me understand!



Thanks
 

·
Registered
Joined
·
6,660 Posts
The arena of A/V design has yet to reach out and touch HTPC and Computer design, simply because an exclusive design standpoint is not easily reached in this arena. Basically stated, the architecture is virtually open to outright intellectual theft. This is the way it has evolved.


In the computer arena, jitter has NEVER been an issue.. but it should be with the current speeds now being realized. They would function in a much more reliable and qualitive way if such issues were properly addressed.


High quality clocking systems can be very expensive. So, this would put them out of the area of standard computer fare.


In HTPC's, the integration and behavior of the soundcard is intimately tied to the behavior of the DVD software. If error correction of the sound card's clock was properly addressed, we MIGHT notice a increase in quality, but it would be minor, due to the fact that the wait states, bus handling, different components on the bus, clock dividers, different component response times..and the overall number of pieces of software,and hardware accessing the systems..as integrated systems themselves...WELL!! This certainly removes much of the effects of proper soundcard clock handling. It would have to be a WHOLE system addressing of the situation,and this is very, very unlikely to happen on a low priced, easily upgradable platform, due to the fact that new parts can be introduced.


The entire system as it works is merely due the delays, wait states, dividers, etc that are fundamental parts of the hardware level bus usage and control system. Without such measures, almost NO computer or HTPC would function at all period.


Engineers of digital hardware, in the computer front do NOT in any way consider these things to be an issue. Hiccuppy behavior is part and parcel of the computer as a working system.


When the very, very sensitive to timing issues human hearing and visual mechanism are engaged with a video image from a HTPC, these problems rear their ugly head. Most people attack this situation with high speed systems, to keep the internal 'wait states' and 'delays' down to a minimum, so as to keep the hiccuping as minimally noticeable as possible. I address it as clocking/bus handling/wait state issues,and have put together a few systems that are considered very slow..but have little in the way of intermittent timing issues.


basically, I have addressed the issue with 64 meg video cards, and messing with the bus handling, and achieved smoothness with 500mhz Celeron (563hz, a 500A running at 75Mhz). The 64 Meg video card as a frame buffer turns out to be essential, if the CPU speed is near the edge, which tells you something, at the very least.


Any way, I have drifted off course on the situation at hand. Better clocking.


Not seen as an issue at all, in the computer arena. Very much so,and properly at that, seen in the AUDIO industry as an issue. So, it might be fun to create better clocks for the HTPC, but the effect of such would be largely lost due to the fact that such disparate pieces of hardware from so many different vendors have to get along in the same box.



Almost a waste of time,unless it is a custom system, but then...it's no longer upgradable or open.
 

·
Registered
Joined
·
3,102 Posts
In terms of any noise, there are always extraneous circumstances one has to understand. The enviroment of the computer is very important. Where the computer is placed in terms of other equipment can cause noise. What I have found out from my past work experience ( troubleshooting many of Disney's show computers in the control room) is that the cabling is the most important factor of controling noise. The internal cables (most are not suitably sheilded) can cause an audio nightmare. same with the cables that are run from the computer to either speakers or external reciever. The PC manufacturers saw this problem years ago. However, since nobody back then used computers as home entertainment systems all they did was offer sheilded speakers. Try to run the PCM through the system's internal BUS (forget the cables that run from CD/DVD to the sound card). Try to buy the best sheilded cables (IDE, speaker, coax) that will be near the PC. And test different powersupplies. Again, not all manufacturers take into account that the PC is going to be used for audiophile quality sound. KBK was correct about clocking. however, remember that all music and movie sound is now compiled on a computer. Those computers just have better cables.
 

·
Registered
Joined
·
164 Posts
From what I understand, the sound card is the important aspect for the HTPC. The SB line of cards (Live, audigy), from what I understand, are not as good as the M-Audio cards for HTPC applications. This has mainly to do with how the card/driver/software sends the digital signal to your decoder (presumably in your receiver or pre/pro.)


I had heard that the SB cards don't send the digital signal out directly, that is, it's processed, and then sent out. (don't know this for a fact.) However, most folks like the M-Audio cards better for HTPC applications anyway.


Now...one place jitter is of concern is when encoding music as mp3. There it's very important that the bit-rates are high enough, you're using a good encoder (i.e., Lame), and your PC can keep up.


I'm not in the HTPC world yet, but will be soon. Heh...I'm still trying to understand fully what sound card to use. :)


Enjoy entering the HTPC world :)


------

AMD Athlon XP 1700+

Abit KG7 Raid.

Seagate Barracuda IV 60gb (2 of them on Raid -- man they are quite!)

1gb ram

Nvidia geforce3
 

·
Registered
Joined
·
2,093 Posts
Discussion Starter · #5 ·
Sorry guys, I was a little confused on the jitter issue. You see I come from a networking background were jitter definitely is an issue in signal transmission. Jitter is considered a form of noise (at least within the realm of Cisco internetworking). However, I will say that I cannot completely agree on the degree to which a PC is inherently bad in terms of latency and wait states.


I think you are correct in many regards. If a PC is not set up/maintained properly one can experience many hiccups. But the fact that a PC can be configured to minimize these things does say something IMHO.


It appears to me that the whole PC industry is obsessed with the issue of latency and wait states in the modern PC environment. Why introduce a (closer to true) multitasking OS such as OS X, or Windows 2K/XP? Sure, a multitasking OS probably is more significant in terms of reliability, but one reason we have to "wait" sometimes is due to poor OS implementation. For instance, while Windows 98 was called a multitasking OS, it was at the expense of reliability given the access that device drivers, and apps had access to the kernel. Hopefully, you get what I'm trying to say, the idea is to create a more reliable computing environment. Part of this equation is the way a OS handles CPU calls.


Another way that this hiccupy behavior is being addressed is via the use of faster memory- DDR/Rambus. As well, Intel and AMD have focused considerable talent and resources on faster chipsets to accommodate these DDR/Rambus modules. Of considerable note, is the development of faster bus speeds and I/O. One example of this is Hypertransport, which has wide-ranging implications in the IT industry.


I realize that the mainboard and cards themselves can be considered as a form of radio transmitter. In fact, I think one of the problems of running higher bus speeds, et al, is that after a certain point a PC can become a radio transmitter (per the FCC). Therefore, the use of the oft' mentioned MAudio sound card might mitigate some this problem as well due to it's good isolation. But, board makers realize this problem and limitation, and engineer around it so to speak. For instance, some VIA/AMD based boards where prone to poor (EMI/RF) control on the chipset circuit traces. This problem was overcome with better internal filtering and in some instance rotating the chipset to minimize the length of the circuit traces.


The idea is that yes, noise is an issue, and data corruption can result from a poorly engineered board. However, board manufacturers are aware of this, and continue to address this problem- at least to the degree to which it affects the PC environment. I still don't know how well the PC environment is for audio, but I have a feeling it will only get better.


On the issue of cabling, I wonder how effective 3M copper shielding foil would be in mitigating some of the EMI/RF problems in a PC? For instance one could wrap the main power supply harness, the power cord, or VGA cable. Perhaps you could even wrap the IDE cables with this stuff. Any cable that might emit EMI/RF (could) potentially benefit from this treatment, no?
 

·
Registered
Joined
·
274 Posts
Well, let's not confuse two different types of noise here. Digital noise (i.e. jitter et al.) is not really relates to RF/EMI interference. Unless, that, is, that your RF noise is so huge that it can alter a signal enough to turn a 0 into a 1 (very unlikely). I'm utterly illeterate in the digital circuit design world, and I can't help anymore there.


What I'd like to add is that if you use an all-digital path, you should be pretty much immune to RF/EMI interference. So if your sound card proccesses and outputs the signal digitally, RF/EMI can't harm it. Same goes for a hard drive connection. For this reason, I'd really like to see a projector with a true DVI digital connector. That would make the connection virtually noise free (talking about added RF noise, not digital noise at the source), and very easy to match pixel-for-pixel.


However, analog video and sound outputs are quite suceptible to EMI interference, and those should be protected. Thankfully, video cards are mostly digital up until they reach the area of the connector. So there's very little room for EMI to leak in and doo actual dmamge and thus a small area to connect. Internal analog audio cables, especially flat ribbon ones that go from the motherboard to a back panel (like in the Soyo Dragon boards), are very vulnerable, and should be protected.


TV tuner cards typically do a decent job of shielding the RF Tuner, but extra shielding couldn't hurt. The big issue with them is that the S-Video and composite inputs are usually not shielded at all (some people tend th think those are digital...). I'm sure that a grounded housing aroung a TV tuner card would cut down on noise a lot.


Good external cables are also crucial. You'd be surprized how much 60Hz radiation is present in a room, especially if you have some fluorescent lighting. That radiatino will creep into cables and components and do all sorts of nasty stuff. But since you too work in electronics, you certainly know that good cables are worth the price.


RFMan
 

·
Registered
Joined
·
283 Posts
Mobius, I assume you are talking about outputting a digital signal (SPDIF) to an external decoder here. I too was confused as to how different sound cards could produce different quality sound, after all as long as the 1's and 0's arrive correctly, what can go wrong?


The problem is that SPDIF is in computer terms a very primitive interface, rather like the way that computers drove peripherals 20 years ago, SPDIF is a real time data stream with no flow control, error correctiojn or anything else, and the majority of receiving equipment is actually very primitive in digital terms.


The core problem arises because the DAC in the receiver usually derives its timing directly from the SPDIF signal, and the timing can have jitter because of poor quality output stage in the soundcard, AND because of a 'noisy' electrical environment, or just because of a long cable run. This doesn't result in the digital data becoming corrupt, but the jitter in the DAC D/A conversion can very easily result in audible differences. As I recall 10 nanoseconds jitter in the clock is reliably detectable, but I can't find the thread that discussed this in detail last year.


Really expensive receivers use a flywheel clock to reclock the timing signal to the DAC which enables them to do a much better job than run of the mill receivers, as the flywheel will remove most of the jitter.


Of course you can avoid this particular problem by doing the D/A in the soundcard, but that just introduces a whole new set of problems about handling high quality analogue audio within the PC environment. Some folks have gone as far as using a delta 1010 sound system, which gets the D/A outside the PC, and do claim very good quality sound.


Then we get to the claims that different DVD players produce different quality sound, even though in theory they are all just demultiplexing the SPDIF signal from the DVD and passing it to the soundcard, I haven't yet read of any plausible (to me) explanation of this, but I have an open mind.....
 
1 - 7 of 7 Posts
Status
Not open for further replies.
Top