I took the time to check L/R phase correctness by swapping speaker cables for each Ruby and Opal files and found everything was in order. (I didn't notice if one has better stereo quality than the other. Well, I actually listen in mono, apart from checking L/R phase correctness, also to make sure I'm comparing the same channel)
Prior to that I found Ruby was more analog. Roughly pre-guessed a turntable, but it didn't sound like turntable then I checked the L/R phasing. Dont know how "tape" could be out of my mind then.
Frank, have you got Audacity/Edit/Preferences/Quality set to Dither None?
This would give the artifacts you describe on low level sine.
Also if Windoz is doing its yucky Sample Rate Conversion.
Foobar2000 uses ASIO and bypasses Windoz.
You're not using Vista by any chance? I don't think there is ANY serious music software that works correctly with Vista.😡
Richard,
1) if he works with 24-bit files with no dither, the artifacts on low level sine are inaudible. Of course they are audible with 16 bit files.
2) regarding sample rate conversion, it must be avoided especially in soundcard itself. If the card is clocked only for 96kHz and 48kHz, 44.1kHz test files should be avoided for the reason that the card makes sample rate conversion automatically then. It is much better to use foobar SRC than card SRC.
3) yes, foobar is preferable for this sound games if one has not a serious sound sw.
I would say that it's impossible to set up a properly controlled test using this kind of "internet delivered" model. The test can never be truly blind because you are giving the raw audio data to your listeners and trusting them to administer the test honestly themselves.
So, don't sweat it and have fun. 😀 To those listeners out there, I would encourage you to use some sort of ABX testing software, and prove to yourself that you can reliably detect a difference before expressing a preference.
Last edited:
Pavel has not refused the advice, in fact he's been very understanding of what it takes to do this sort of test properly. He has just come to the conclusion that it's more trouble for him than it's worth, and I frankly don't blame him.
Not impossible for things like opamp comparisons, but not trivial to set up to foil cheating by self-proclaimed golden ears.
I don't know what the official definition is for noise and distortion but I think of distortion as being a function of the signal and noise being everything else where some of that is random with no information content and some is interference with information content.
I think it helps to differentiate between the random noise which will average out of an FFT and disappear and noise with information content which will show up in an FFT.
Thank you, Stuart. And, secondly, I have come to the conclusion that there are so many flaws in usual PC based audio chains used by members that it would be difficult to dig out ANY useful information from the test results. I believe that one of the postulates of the proper test is that participants have to use the same hardware and that the hardware must not limit the results by its flaws. Only in case we want to test flaws of the hardware used such test might make sense. It was not my intention, I wanted to test really very minor differences caused by high quality hardware. I have come to the conclusion that it is impossible in a public web test. Nevertheless, I thank all the participants for their effort and patience. I am ironic sometimes, but I see no chance in detailed and detailed explanations for the reason that more and more unexpected and bizarre questions result from explanations. Same or similar level of knowledge is requested in such discussions and web democracy and openness does not allow for it, though it is great that it attracts attention of so many people and it enables easy worldwide communication.
With such a huge difference in the noise floor, I assumed that that was what the test was about so answered according to my strong preference for random noise over warble.
I thought it was going to be an interesting test to see if the quality of the noise floor was significant...
If that was what the test was about, I would have to wonder why Mooly introduced 2 minutes of music to distract us when a clip of the noise floor alone would have done the job. 🙂
Glitches like this just show how hard it is to do a properly controlled test. I once tried a comparison between two high-end mic preamps, and one of them sounded "bigger" and "more spacious". It simply had more noise at low frequencies.
Glitches like this just show how hard it is to do a properly controlled test. I once tried a comparison between two high-end mic preamps, and one of them sounded "bigger" and "more spacious". It simply had more noise at low frequencies.
Last edited:
Interesting, and quite plausible. We probably judge the size of a space from the frequency of the lowest acoustic modes supported in it. We can do this in the absence of music as there is always ambient noise.scopeboy said:
LF noise adds 'space'. HF noise adds 'air' and 'detail'. Low order distortion adds 'warmth'. I will soon be able to design a 'high-end' preamp and impress journalists!!
Make sure the noise on the l&r channels is not correlated.
The more you uncorrelate the channels, the more spacious it will sound.
- Status
- Not open for further replies.