Skip to main content ↑ ↓

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Transparent Gear and Testing (Read 20436 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Transparent Gear and Testing

Reply #25 –
Quote from: saratoga on
Quote from: theriverlethe on
Quote from: Arnold B. Krueger on
Quote from: theriverlethe on
You're still going to get background noise as high as -80dB or higher, which could be audible at high volume during silent passages.


As a rule, the noise floor of almost all commercial recordings is -70 dB or higher. There are a few exceptions.


The source of the noise is typically "room tone" in the recording studio or venue where the recording was made, and that usually turns out to be spectrally shaped  so as to put a lot of the noise at frequencies where the ear is not particularly sensitive.


Computer noise is not shaped to those standards.


Which page exactly are you getting those numbers from?


My own testing with an ALC889

Transparent Gear and Testing

Reply #26 –
Quote from: Arnold B. Krueger on
Quote from: theriverlethe on
Quote from: Arnold B. Krueger on
Quote from: theriverlethe on
Quote from: Arnold B. Krueger on
Quote from: theriverlethe on
You're still going to get background noise as high as -80dB or higher, which could be audible at high volume during silent passages.


As a rule, the noise floor of almost all commercial recordings is -70 dB or higher. There are a few exceptions.


The source of the noise is typically "room tone" in the recording studio or venue where the recording was made, and that usually turns out to be spectrally shaped  so as to put a lot of the noise at frequencies where the ear is not particularly sensitive.


Computer noise is not shaped to those standards.


Which means that there may or may not be a problem.

The fact that computer noise is another 20 dB down in most cases often effectively addresses that concern.


I don't know how you established "in most cases,"


Over 20 years experience measuring and liotsening.


Quote
but I'm planning to check my motherboard grounding...


May help if not done right the first time.


I checked motherboard connections and tightened screws...  Any other ideas?

This is not even 16-bit performance, and the computer is practically idling:

https://onedrive.live.com/redir?resid=5B2A7...int=photo%2cPNG

Transparent Gear and Testing

Reply #27 –
Quote from: theriverlethe on
I checked motherboard connections and tightened screws...  Any other ideas?

This is not even 16-bit performance, and the computer is practically idling:

https://onedrive.live.com/redir?resid=5B2A7...int=photo%2cPNG


Is this a self-loop using the onboard ADC? I know you have a Tascam interface in previous posts, try to switch the gain to GUITAR, turn the gain knob to compensate level differences and run the test in 24-bit again.

My VIA HD audio tests have much better results (recorded by X-Fi Titanium HD)
http://www.hydrogenaud.io/forums/index.php...ost&id=7804

Transparent Gear and Testing

Reply #29 –
Quote from: bennetng on
Quote from: theriverlethe on
I checked motherboard connections and tightened screws...  Any other ideas?

This is not even 16-bit performance, and the computer is practically idling:

https://onedrive.live.com/redir?resid=5B2A7...int=photo%2cPNG


Is this a self-loop using the onboard ADC? I know you have a Tascam interface in previous posts, try to switch the gain to GUITAR, turn the gain knob to compensate level differences and run the test in 24-bit again.

My VIA HD audio tests have much better results.
http://www.hydrogenaud.io/forums/index.php...ost&id=7804


You beat me by a minute!

Transparent Gear and Testing

Reply #30 –
One of the greatest contributions from RMAA to consumer audio world is that manufacturers are forced to use excellent ADCs even for $100 soundcards.

Transparent Gear and Testing

Reply #31 –
Quote from: saratoga on
Quote from: theriverlethe on
Quote from: Arnold B. Krueger on
Quote from: theriverlethe on
You're still going to get background noise as high as -80dB or higher, which could be audible at high volume during silent passages.


As a rule, the noise floor of almost all commercial recordings is -70 dB or higher. There are a few exceptions.


The source of the noise is typically "room tone" in the recording studio or venue where the recording was made, and that usually turns out to be spectrally shaped  so as to put a lot of the noise at frequencies where the ear is not particularly sensitive.


Computer noise is not shaped to those standards.


Which page exactly are you getting those numbers from?


I presume that the -80 dB number comes from spec sheets and actual tests of audio interfaces and players.  IME, most of the current crop of on-board audio interfaces are about 10 dB or more better.

My -70 dB number comes from analysis of a large number of commercial recordings.  A recording with this low of a noise floor would likely be very dynamic and uncompressed. For the longest time my benchmark for this number was Rickie Lee James eponymous album and its still a good number for a pop album.  Most recordings are at least 10 dB worse.  Highly compressed recordings are going to be far worse. 

There is a real issue here which is differences in the spectral content of the noise floors of digital gear, and that of real-world recordings.  There is Fielder's classic JAES paper which has information for some of the finest recording venues in the world, which are generally in a class by themselves. That paper does not seem to represent actual working situations, even for work that is done in them. The noise floor of an world class empty room in the dead of the night is not representative of what they are like with 100 musicians playing their instruments in it in the middle of a normal working day.

Not mentioned is the noise floor of typical playback environments which are generally far poorer yet, and have a spectral shape that is more like that of recording studios and the like.  How many complain about the noise floor of say Sansa Clips and Fuzes which are well known to be in the -85 to -95 dB range?

Measurements of generally well-regarded sources

Transparent Gear and Testing

Reply #32 –
Quote from: bennetng on
One of the greatest contributions from RMAA to consumer audio world is that manufacturers are forced to use excellent ADCs even for $100 soundcards.


I suspect that the chicken came before the egg. ;-)

My view is that DAC technology progressed to the point where it wasn't worth building a sound card audio chip without a halfways decent DAC in it.

It was competition among the vendors that made it happen.

Realtek's commanding market share didn't hurt.

RMAA allowed a large number of people to observe the process by which this happened.

In fact the popularity of USB DACs has seems to have led to a fairly massive disimprovement in computer audio interface performance. I've been testing a fair number of pro and consumer USB audio interfaces and many have relatively massive artifacts due to their USB interfaces.  The 7 KHz spike that Archimago has made famous seems to be pretty popular.

On balance their performance is probably good enough for human ears, but the working devices often vastly underperform their DAC chips.  USB DACs from a few years back such as the eMU 0404 don't seem to have a lot of modern competition.

Transparent Gear and Testing

Reply #33 –
I mean ADC and it is not a typo. People like to share their results on the internet therefore manufacturers need to use better ADCs on their products otherwise the loopback results will look ugly. I still remember tomshardware made a mistake in looping the Audigy in around 2000, which resulted in poor frequency response, then Creative uploaded a RMAA guide for users to test their cards. Since then Creative also have their RMAA guide included in later flagship soundcards and Asus also have their guide to test their flagship Xonar cards.

http://audio.rightmark.org/download.shtml

My X-Fi scored 119dB in loopback tests.
http://www.hydrogenaud.io/forums/index.php?showtopic=100481

But now manufacturers are not interested to make PCI soundcards anymore and there are many USB DACs without ADC function.

Transparent Gear and Testing

Reply #34 –
Quote from: bennetng on
I mean ADC and it is not a typo.


Thanks for the correction.

Traditionally, audio interfaces have had symmetrical performance.

The asymmetrical products probably exist due to budgetary considerations.

I don't see the Rightmark program, as much as I appreciate it, as being a strong influence simply because almost every time I google for a Rightmark test for a product, there is none.

Transparent Gear and Testing

Reply #35 –
Quote from: Arnold B. Krueger on
In fact the popularity of USB DACs has seems to have led to a fairly massive disimprovement in computer audio interface performance. I've been testing a fair number of pro and consumer USB audio interfaces and many have relatively massive artifacts due to their USB interfaces.  The 7 KHz spike that Archimago has made famous seems to be pretty popular.

On balance their performance is probably good enough for human ears, but the working devices often vastly underperform their DAC chips.  USB DACs from a few years back such as the eMU 0404 don't seem to have a lot of modern competition.

In reply to your comments on recent USB interfaces, I agree there are some regressions. The only USB RMAA result I can find which is better than my PCIe X-Fi is Lynx Hilo but it costs $2495 while my X-Fi is only $170 back in 2013. Benchmark DAC2 is another possible candidate but I can only see Audio Precision results in the manual, so the numbers may be incompatible with RMAA. I am highly interested to see John Siau post some results by using RMAA test signals to loop DAC2 and ADC1 as an article in Benchmark's site said their products already outperformed Audio Precision analyzers.

http://benchmarkmedia.com/blogs/news/40489...de-measurements

Transparent Gear and Testing

Reply #36 –
Quote from: theriverlethe on
This is not even 16-bit performance, and the computer is practically idling:

https://onedrive.live.com/redir?resid=5B2A7...int=photo%2cPNG


I was more concerned about your claim of 80dB range or worse, but 85 dB isn't that bad, particularly if you did that as a loop back test.  Take a look at this thread:

http://www.hydrogenaud.io/forums/index.php?showtopic=92637

FWIW, the newer parts are a lot better:

http://www.anandtech.com/show/8797/asus-x9...rboard-review/5

Although I don't think it really matters much in practice, unless maybe you're using a headphone amp.

Transparent Gear and Testing

Reply #37 –
Quote from: bennetng on
In reply to your comments on recent USB interfaces, I agree there are some regressions. The only USB RMAA result I can find which is better than my PCIe X-Fi is Lynx Hilo but it costs $2495 while my X-Fi is only $170 back in 2013. Benchmark DAC2 is another possible candidate but I can only see Audio Precision results in the manual, so the numbers may be incompatible with RMAA. I am highly interested to see John Siau post some results by using RMAA test signals to loop DAC2 and ADC1 as an article in Benchmark's site said their products already outperformed Audio Precision analyzers.

http://benchmarkmedia.com/blogs/news/40489...de-measurements


I've seen a lot of RMAA results for the Hilo, but none seem to clarify that they were made using the USB interface.  Both the Hilo and the DAC2 can be operated using other sources than USB-2 and the issues that I see in other products are USB-2 related.

Transparent Gear and Testing

Reply #38 –
Quote from: saratoga on
Quote from: theriverlethe on
This is not even 16-bit performance, and the computer is practically idling:

https://onedrive.live.com/redir?resid=5B2A7...int=photo%2cPNG


I was more concerned about your claim of 80dB range or worse, but 85 dB isn't that bad, particularly if you did that as a loop back test.  Take a look at this thread:

http://www.hydrogenaud.io/forums/index.php?showtopic=92637

FWIW, the newer parts are a lot better:

http://www.anandtech.com/show/8797/asus-x9...rboard-review/5

Although I don't think it really matters much in practice, unless maybe you're using a headphone amp.


The -85dB number came from the Tascam.  I had to use 18dB of gain because of the onboard audio's laughable output.  -80dB might have been from a loop back, or just an exaggeration from memory.

Aren't those Tom's Hardware examples all $200+ enthusiast motherboards reviewed in enthusiast magazines, along with the motherboard used in the original ABX?  Hardly a typical end-user system, and I doubt mine is the worst example...

Of course I'm using a headphone amp.

Transparent Gear and Testing

Reply #39 –
Quote from: Arnold B. Krueger on
I've seen a lot of RMAA results for the Hilo, but none seem to clarify that they were made using the USB interface.  Both the Hilo and the DAC2 can be operated using other sources than USB-2 and the issues that I see in other products are USB-2 related.

I don't like USB interfaces because they generally have higher latency and CPU usage than PCI/e ones, and they usually don't support multiclient ASIO except for some expensive models. I am a virtual instruments user so these are essential requirements.

I didn't notice USB interfaces' sound quality issue is caused by the type of connection until you mention it and I am pretty surprised. In another forum a member asked for a solution to hook up an Onkyo soundcard with only one stereo line out and a toslink output to powered speakers and headphones with an ability to switch between speakers and headphones without too much hassle, and he doesn't want to give up using the soundcard completely. I recommended a $60 SMSL SPDIF only DAC with headphones amp while another member (A) recommended a USB-only one. I said using USB is inconvenient because the OP need to switch between audio devices in Windows control panel and/or audio/video player device preferences and A said that it is a common sense to avoid toslink since it is jittery. SIGH.

Actually toslink measured better than other electrical interfaces at my home because it can effectively avoid ground loop.

Transparent Gear and Testing

Reply #40 –
Quote from: theriverlethe on
The -85dB number came from the Tascam.  I had to use 18dB of gain because of the onboard audio's laughable output.  -80dB might have been from a loop back, or just an exaggeration from memory.

Aren't those Tom's Hardware examples all $200+ enthusiast motherboards reviewed in enthusiast magazines, along with the motherboard used in the original ABX?  Hardly a typical end-user system, and I doubt mine is the worst example...

Of course I'm using a headphone amp.


I hope the Tascam's gain is analog otherwise 18dB is a big penalty to SNR. My motherboard is only an el cheapo ASRock H55DE3 with VT1718S codec but I only have 5dB level differences when recording using my X-Fi.

Transparent Gear and Testing

Reply #41 –
Quote from: bennetng on
I didn't notice USB interfaces' sound quality issue is caused by the type of connection until you mention it and I am pretty surprised. In another forum a member asked for a solution to hook up an Onkyo soundcard with only one stereo line out and a toslink output to powered speakers and headphones with an ability to switch between speakers and headphones without too much hassle, and he doesn't want to give up using the soundcard completely. I recommended a $60 SMSL SPDIF only DAC with headphones amp while another member (A) recommended a USB-only one. I said using USB is inconvenient because the OP need to switch between audio devices in Windows control panel and/or audio/video player device preferences and A said that it is a common sense to avoid toslink since it is jittery. SIGH.

Actually toslink measured better than other electrical interfaces at my home because it can effectively avoid ground loop.


Toslink is very good for the reason you mention - its complete and total electrical isolation.

Ironically, USB 2.0, at least at the lower speeds used with stereo audio interfaces, can be totally electrically isolated because the signal wiring is transformer coupled.

The higher speed forms of USB 2.0 that are associated with products such as external hard drives appear to have lost the transformer coupling in the quest for speed.

The specific faults that I am seeing appear to be related to the packetizing of audio data.

This is further described here:

Apple TNL about USB 2 packet sizes and rep rates

with some user experiences here:

Archimago report of USB packet noises

Transparent Gear and Testing

Reply #42 –
Quote from: krabapple on
Quote from: Dark_wizzie on
Quote from: krabapple on
Quote from: Dark_wizzie on
Thanks for the replies guys. However, I was hoping for something of a larger scale. For example, when we do drug trials or crime statistics, we look for a much larger sample size than 2 people. I understand that doing thousands upon thousands tests like this will be impractical but I was hoping for maybe a test of 10, 20, 30, 40, 50 people.



You can't extrapolate from bench measurements + known limits of human hearing?

Do your own experiments if that's insufficient.  Because who, exactly, is going to perform (and pay for) the experiments you require?  DAC makers? Why would they?

This isn't about my own ability to discern a difference, it's about human beings' abilities to discern a difference in general, and whether we can just say that Odac is transparent, full stop and anybody that disagrees is wrong.



You seem to have ignored my first (rhetorical) question.

But at least now we have a better idea of the axe you're intending to grind.

Bruh, you don't even know me.

I personally believe Odac is transparent.  I hope it is possible to ask questions without people questioning each other's motivations.


I feel it would be a stronger case if there was some sort of study out there that pulled 50, 100 people and did a double blind study about transparency of say, the Odac. There are many ways to ignore the results, but I think it would convince some people. I only came up with this thread because somebody I knew felt this would be the nail in the coffin for him.

Transparent Gear and Testing

Reply #43 –
Quote from: Dark_wizzie on
I personally believe Odac is transparent.
I feel it would be a stronger case if there was some sort of study out there that pulled 50, 100 people and did a double blind study about transparency of say, the Odac.

What are you waiting for bruh?

Let me guess, non-belivers to spend their time doing all the heavy lifting searching for yet another witch effect, so that believers can ignore the results.
Loudspeaker manufacturer

Transparent Gear and Testing

Reply #44 –
Quote from: ajinfla on
Quote from: Dark_wizzie on
I personally believe Odac is transparent.
I feel it would be a stronger case if there was some sort of study out there that pulled 50, 100 people and did a double blind study about transparency of say, the Odac.

What are you waiting for bruh?

Let me guess, non-belivers to spend their time doing all the heavy lifting searching for yet another witch effect, so that believers can ignore the results.


You seem a little pessimistic about the whole situation. It ain't all that bad.

Transparent Gear and Testing

Reply #45 –
Quote from: ajinfla on
Let me guess, non-belivers to spend their time doing all the heavy lifting searching for yet another witch effect, so that believers can ignore the results.

What are you saying exactly?

Transparent Gear and Testing

Reply #46 –
Sane people might not have time nor desire for a large scale double blind test, to see whether an ODAC et al, is "transparent".

cheers,

AJ
Loudspeaker manufacturer

Transparent Gear and Testing

Reply #47 –
Quote from: ajinfla on
Sane people might not have time nor desire for a large scale double blind test, to see whether an ODAC et al, is "transparent".


Sane people might look at the ODAC's excellent measured performance and see no need for that large scale DBT.

Why shouldn't ODAC nay-sayers be obliged to follow the same stardards as ODAC proponents?

Transparent Gear and Testing

Reply #48 –
Quote from: Dark_wizzie on
I feel it would be a stronger case if there was some sort of study out there that pulled 50, 100 people and did a double blind study about transparency of say, the Odac. There are many ways to ignore the results, but I think it would convince some people...

For the group of people who claim the piece of gear is not transparent, all they have to do is find one single person who can hear a difference, using scientific protocols to preclude  bias such as level matching and DBTing, using whatever music they wish, and they win the argument.

For the group of people who claim the piece of gear is transparent, all they have to do is test every single person on the planet, with every single song ever recorded, through every price no object audio system, otherwise when their 50, 100, or 100 million test subjects fail to hear the difference it will simply be dismissed by the Stereophools with the exact same three reasons they always use, and which I listed earlier:

- you used the wrong people with inadequate hearing and/or training
- you used the wrong, non-revealing music
- you used inadequate playback gear which obscured the subtle differences

Which group of people has the easier task of providing the burden of proof that their rivals insist upon?

Transparent Gear and Testing

Reply #49 –
Quote from: mzil on
Quote from: Dark_wizzie on
I feel it would be a stronger case if there was some sort of study out there that pulled 50, 100 people and did a double blind study about transparency of say, the Odac. There are many ways to ignore the results, but I think it would convince some people...

For the group of people who claim the piece of gear is not transparent, all they have to do is find one single person who can hear a difference, using scientific protocols to preclude  bias such as level matching and DBTing, using whatever music they wish, and they win the argument.


Except if a single instance is 'not transparent' , that does not mean that a listener can be expected to be 'blown away' by the difference, 'even my wife could hear it', 'veils were lifted',  etc.  And THAT is the rhetorical norm in audio difference reporting.

Simply passing the 'not transparent'  threshold doesn't cut it.  Let's not let audiophiles get away with such bluster (e.g., wild extrapolations from the Meridian report on digital filter difference).