[ROBSCIX]

You don't have to use a loopback test though. You can use another card to measure the output from the card being tested.
You see people posting RMAA results all the time and many have no idea if they are correct or accurate or even possible.

RMAA is a great tool if the person with the tool knows how to use and interpret them.

 

[Zaubertuba]

Quote:

Separate high quality DAC's for L & R channels Improved jitter correction Higher quality Clock generators EMI/RFI shielding Possible Balanced output such as XLR Different output section TRS, TS Improved output sections using higher quality components Different daughter cards allows specific features.

Some cards actually already have at least some of these features--my 1212m has a few, and it made a pretty outstanding difference from my onboard Realtek audio.

Part of the problem (or perceived problem) is once you put a pci card in a computer, it really becomes a part of a larger system, both hardware- and software-wise. As many people have mentioned, a $3,000 external DAC and an onboard card could have the same DAC chip (or chips), but differ greatly in the supporting componentry and implementation.

So, following this logic, I think you need to look at more than the card itself, but the supporting MB, power supply, Operating System and installed software components.

I've been wondering myself: Some cards like the Essence STX pull their power right off the power supply, skipping the power running through the PCI bus. But I got to thinking, couldn't you make some improvement to your computer source just by upgrading to a better PSU for the whole computer? And why wouldn't you, considering better componentry on this side might translate into better stability/performance for your computer source in general?

People spend a lot of time debating about power supply cables for their outboard gear, but what about that cheap wiring to the MB main and the Molex connectors that's ubiquitous in the industry today? I'd be interested in seeing what, if any improvement you might see if someone improved their PSU unit (I've been looking at some of the quiet units by Seasonic myself), and modded their own power supply cables for the internals inside the computer case. *If* there's any issues with EMI/EFI (which some say is debatable), it seems you could make some notable improvements in this area by working on the problem closer to at least one of the sources, I would think.

I know this is a bit broader, perhaps, than your interest in improving soundcards, specifically. To come to a point, I guess my thought is there are a number of soundcards out there that are already pretty darned good, and it doesn't make a lot of sense improving them much more until we look at improving what we're plugging them in to.

Some very theoretical musings, I'm afraid, but you started a very theoretical thread. I'm very interested in hearing some other people's comments.

 

[royalcrown]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif RMMA results use loopback, which can hide any groundloop/EMI/noise problem you could/would encounter w/ a third party input. these tests are worthless

Any noise or EMI is meaningless unless they actually affect how the sound card performs, and the rmaa test shows that the EMI and noise is handled to the point where it doesn't affect the performance enough to be even close to audible. As for groundloops you may have a point, but it's a non-issue if (in my setup as well as presumably amb's) there are no ground loops to begin with even when hooked up to headphones. Besides, a "better" DAC won't remove ground loops.

The only reason a loopback test is flawed is because you won't be able to tell if the ADC or DAC is at fault if you get bad measurements. However, it's not an issue when the test results come back really good, because if either the DAC or the ADC are flawed, it will show up on the loopback test. You can clearly see, however, that both are so good that it doesn't even matter, because they both have audibly perfect measurements.

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif IMHO, this Japanese site did it really well, I just dunno if they used a bit-perfect player...but even if they didn't their results are still plenty more meaningful than some random RMAA results

Results off of that site for the EMU 0404:

THD：0.00098％ S/N：84.69dB

that's even better than the rmaa testing, but largely unimportant because they both show that the specs are WELL beyond the threshold of hearing. If anything the rmaa tests are more comprehensive because they test more parameters than the Japanese site, but they both largely come to the same conclusion.

 

[royalcrown]

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif Part of the problem (or perceived problem) is once you put a pci card in a computer, it really becomes a part of a larger system, both hardware- and software-wise. As many people have mentioned, a $3,000 external DAC and an onboard card could have the same DAC chip (or chips), but differ greatly in the supporting componentry and implementation. So, following this logic, I think you need to look at more than the card itself, but the supporting MB, power supply, Operating System and installed software components.

There's absolutely zero evidence that two different (properly functioning) motherboards will change the output of a discrete PCI sound card, and it's even more nonsensical to say that the OS will make any difference given bit-perfect output. In fact, there's really no evidence to support the case that even the computer power supply makes any difference at all. No blind-testing shows a positive result, and the measurements indicate that they make no difference whatsoever. If computer components mattered, or even if a sound card being internal/external mattered, the rmaa testing would have shown those differences, but there's no difference to speak of (I think it's a prudent assumption that my own and amb's computer setups differ substantially in components).

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif I've been wondering myself: Some cards like the Essence STX pull their power right off the power supply, skipping the power running through the PCI bus. But I got to thinking, couldn't you make some improvement to your computer source just by upgrading to a better PSU for the whole computer? And why wouldn't you, considering better componentry on this side might translate into better stability/performance for your computer source in general?

I don't see how performance of either the sound card or the computer will benefit from this. A DAC isn't exactly a power hungry device, and computer voltages on a proper power supply (i.e. one that is UL certified and well-built, which can be had very inexpensively) are more than stable enough. But again, even if it could make a difference, it would've been reflected in testing.

Quote:

Originally Posted by Zaubertuba /img/forum/go_quote.gif People spend a lot of time debating about power supply cables for their outboard gear, but what about that cheap wiring to the MB main and the Molex connectors that's ubiquitous in the industry today? I'd be interested in seeing what, if any improvement you might see if someone improved their PSU unit (I've been looking at some of the quiet units by Seasonic myself), and modded their own power supply cables for the internals inside the computer case. *If* there's any issues with EMI/EFI (which some say is debatable), it seems you could make some notable improvements in this area by working on the problem closer to at least one of the sources, I would think.

Modifying the power cables inside a computer is beyond extreme, and I highly doubt it would matter. Considering the low voltages going through a computer, I doubt it needs anything better than what the UL certifies, but then again if you're under the impression that power cables make an audible difference... see the huge thread about that on this subforum.

 

[leeperry]

Quote:

Originally Posted by royalcrown /img/forum/go_quote.gif Results off of that site for the EMU 0404: THD：0.00098％ S/N：84.69dB that's even better than the rmaa testing, but largely unimportant because they both show that the specs are WELL beyond the threshold of hearing. If anything the rmaa tests are more comprehensive because they test more parameters than the Japanese site, but they both largely come to the same conclusion.

well yeah, they used a top of the line Roland rack....these results don't mean anything out of their context.

but when you compare all these results, you see that the 0404USB yields the best results...the less "spikes" the better

the most important data for us is SNR & THD I think, and you can see how BAD Creative fares, and same for the Prodigy HD2+....AudioTrak boasts about the DAC specs on their commercial stuff(OMG 124 dB SNR

), but its PCB shielding simply sucks(plus it's got some lousy JRC4580 for the headphones output that are not on DIP8 holders)

 

[royalcrown]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif well yeah, they used a top of the line Roland rack....these results don't mean anything out of their context.

Whatever reference ADC they use is of no consequence - the tests aren't limited by context in that manner unless you actually plan to use the ADC (which 99% of people here aren't concerned with). The DAC performance are all we're interested in, and rmaa tests usually underestimate this (loopback, because the reference ADC is just whatever came with the DAC) - but even when underestimated, the DAC's performance is past the boundaries of transparency.

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif but when you compare all these results, you see that the 0404USB yields the best results...the less "spikes" the better

That's not true in theory as well as in practice. In practice, all that matters is audibility, and all of the spikes are WELL below audibility. But even looking at the spikes, just because there's more of them doesn't necessarily mean that it's worse. The actual figure matters more than how the graph looks at the levels we're talking about.

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif the most important data for us is SNR & THD I think, and you can see how BAD Creative fares, and same for the Prodigy HD2+....AudioTrak boasts about the DAC specs on their commercial stuff(OMG 124 dB SNR ), but its PCB shielding simply sucks(plus it's got some lousy JRC4580 for the headphones output that are not on DIP8 holders)

A high SNR and low THD won't matter if there's a lot of IMD. Granted IMD is usually very low, but SNR and THD aren't the end-all determinants of sound quality - it's about the whole picture. Furthermore, SNR will be limited by your audio 99% of the time rather than by your equipment, especially when it comes to the soundcards we're looking at - 124db is beyond overkill.

The PCB shielding, on the other hand, is a nonissue unless it affects the actual output of the card, which it seems to not be doing given its SNR. If its shielding is actually harming the sound quality, you'd be able to tell quite easily by just looking at the measurements.

 

[leeperry]

so the lousy results on that site from the Prodigy HD2, or worse the Creative X-Fi Xtreme Music are meaningless to you?

the low SNR/high THD and all the spikes around the 1Khz sine test tone don't mean anything?

I actually bought my soundcard after reading these tests, I couldn't afford the 0404USB....and mine wasn't too far from it, and I love it

I actually thought that the X-Fi Xtreme & the HD2 yielded poor results in that test because of lousy onboard components(cheapo op-amps) and bad PCB shielding....coz you can see that some PCI cards do give goood results like the Razor

 

[royalcrown]

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif so the lousy results on that site from the Prodigy HD2, or worse the Creative X-Fi Xtreme Music are meaningless to you?

The Prodigy HD2 got a THD of 0.00201%, well below audibility - it tests just fine. The Creative X-Fi Xtreme Music, however, seems to be an anomaly. They may have had a bad card, or turned on the "crystalizer" or some such thing creative bundled with the card, because all of the testing done on multiple other sites show a much more reasonable THD and noise level. Their testing had THD and SNRs much worse than most onboard sound, so there's reason to suspect something's up there.

Quote:

Originally Posted by leeperry /img/forum/go_quote.gif I actually thought that the X-Fi Xtreme & the HD2 yielded poor results in that test because of lousy onboard components(cheapo op-amps) and bad PCB shielding....coz you can see that some PCI cards do give goood results like the Razor

Lousy components, sure, but I'd need to see evidence that the lack of SQ is due to PCB shielding before I buy that, because it's much much more probable that poor quality components are at work. Heck, if you take a crappy DAC and shove it in a box and sit it away from your computer you'll still have bad results.

 

[12Bass]

There's nothing wrong with internal soundcards. True, the inside of a computer is not without various challenges. However, a well designed card can achieve better than 110 dB SNR without much trouble, a feat which would make it quieter than most amplifiers, as well as put the card's noise well below the ambient noise in almost all listening situations. My internal Echo Gina24 has no problem meeting this specification.... THD is 0.0005%, IMD+N is 0.0014%.

Edit: What happened to the 'delete' function?

 

[olblueyez]

The answer is, "Make them external so better components will fit". I know surface mount solves lots of the size issues but some components cant be reduced without some compromise. Anyone know if someone makes a discrete output section on an internal card? If not then is it because of power issues?

 

[leeperry]

well Onkyo sound cards pack a lot of hi quality capacitors apparently, plus copper shielding and stuff

 

[obobskivich]

blueyez:

most of that board is the power stage, which is provided by the host computer

ultimate answer:
theres nothing wrong with internal soundcards on the whole, except the mythos surrounding them

 

[olblueyez]

Quote:

Originally Posted by obobskivich /img/forum/go_quote.gif blueyez: most of that board is the power stage, which is provided by the host computer ultimate answer: theres nothing wrong with internal soundcards on the whole, except the mythos surrounding them

So they do make sound cards with a class A discrete output section?

 

[null_pointer_us]

Well, if internal sound cards are so bad, why not just use them for gaming support while directing all the sound through digital audio output to a dedicated amp? That ought to give you much better quality -- look at how much bigger a normal amp's case is

-- than an external sound card.

This is assuming of course that all the interference stuff is true.

Personally, given that my internal sound card is capable of ruining my hearing at 2% volume w/ Sennheiser HD595, without any noticeable noise/distortion, I can't imagine why going external would help in any practical sense.

My sound card is just the $59 X-Fi PCI-E Extreme Audio {rebranded from Audigy SE [rebranded from something earlier (Live! 24-bit?)]}, which I've read is not even a good card by modern standards.

 

[obobskivich]

Quote:

Originally Posted by null_pointer_us /img/forum/go_quote.gif Well, if internal sound cards are so bad, why not just use them for gaming support while directing all the sound through digital audio output to a dedicated amp? That ought to give you much better quality -- look at how much bigger a normal amp's case is -- than an external sound card. This is assuming of course that all the interference stuff is true. Personally, given that my internal sound card is capable of ruining my hearing at 2% volume w/ Sennheiser HD595, without any noticeable noise/distortion, I can't imagine why going external would help in any practical sense. My sound card is just the $59 X-Fi PCI-E Extreme Audio {rebranded from Audigy SE [rebranded from something earlier (Live! 24-bit?)]}, which I've read is not even a good card by modern standards.

oh, and if RFI/EMI really bugs you -> TOSlink

blueyez:
I don't know to be honest, the SE200 has discrete outputs, but I don't know if its full class A (it does have separate L/R output stages, but I'm not sure if they're full class A, I would suspect yes (given that we're talking very low output power), but I can't find any documentation on Onkyo's (japanese only) website to confirm this)

my point was simply that in your example, a huge chunk of that box is for the power supply, which is provided "externally" for a soundcard (it recieves power via PCI, or PCI + extension), and many features are handled by the DSP itself (that require multiple components in your external box)

I'm gonna venture that power is NOT an issue in providing discrete class A outputs, as a modern graphics adapter will easily consume more power than that D/A (consider that an average external/discrete D/A box is around 20W consumption in normal operation, and an average graphics adapter is closer to 200W, I don't think power supply is a problem, probably more along the lines of market demand (and given that we've seen "headphone output" cards get quite popular recently, I wouldn't be surprised if we saw a soundcard that meets your fancy in the next year or so))