May be not a bug, but my not understanding of the software.
So you can download the Norah1.asc and Norah2.asc...
What if you rename the files secretly into something like A.asc and B.asc... Then can I tell you which one is which? Of course, I know which one is louder between Norah1 and Norah2. So by listening to A and B I should be able to tell which one is which simply by knowing which one is the louder one between A and B, correct?
If I can do that, what kind of bug do you think is possible? If the file is similar, how can a buggy software agree on the louder one?
I would have thought keeping it digital till the last moment would be best, even better active speakers with amps near the drivers, convert the digital at near the amp as possible, a short analogue path is lass prone to noise pick up etc. That's the advantage of digital signal transmission your data arrives unchanged.
Well I never had them do anything other than amplify at home but they are frequently used to run 4 8Ohm subs at or near full power in parallel.
I suspect heat may become an issue than. Most of the protection is there to avoid the amps shutting down completely when used in PA (clubs mostly, the things are heavy) systems. Also they never dynamically compress the signal, the output just gets quieter over a few minutes but at no time is the dynamic range of the input signal restricted in any way.
Point of all this being that there is nothing in the signal path except the actual amplifier, no limiters, relays or switches of any kind.
Also the fade up on switch on avoids the pop that is usually just hidden by a relais making the whole amp less likely to fail in the long term.
I love Foobar ergonomics, just I don't trust it.
I find VST Wrapper and ABX applet have issues.
Dan.
The files are bitwise identical, and so are wav files produced from them. There is no data in them to render into analog at different volumes.
I suppose the gain of foobar could be slightly off for files depending on which order they are specified for input. Maybe the 1st file you enter into to it gets played slightly louder. Maybe it sorts them by name alphabetically and the volume is affected by the sort order. I don't know what it is, but its something. Maybe a bug only for MP3 files, such as only if 2 instances of the MP3 decoder are running at once. Obviously, with a computer program, any kind of bug is possible. Some can be extremely obscure to find.
Last edited:
That's what I do, digital all the way, then DAC, then amplifier. At least if the source material is digital, such as from CD or DVD. But my DAC-1 does happen to have an analog volume control. To minimize any possible reconstruction filter inter-sample overs in the DAC, its actually better to attenuate the digital signal before the DAC by about 3.5 db. That leaves plenty of headroom for worst case reconstruction filter clipping. Make the gain back up later, with the DAC volume control, or the amp input balance trim pots.
How about a "test" I mentioned earlier. Can it be used to check if our understanding about the file is wrong? Because I'm not as sure as you are that they are identical... (because, I'm the one in the ABX window, knowing which one is which, approved by the software...)
You can make several set of files from Norah1.asc and Norah2.asc, for example:
Norah1 --> A
Norah2 --> B
Norah1 --> X
Norah2 --> W
Norah1 --> 2
Norah2 --> 1
and so on
If I can do the matching, what is the chance that the file is identical? Are you not interested to solve this puzzle?
BTW, I just tried to create 2 files with 2dB difference so I could post here for anybody to try. I wanted to make sure that the process works flawlessly. Surprisingly, the file size is different... May be when the difference is too small, the file size doesn't change? (I will try again later with smaller dB difference).
😡 What confuse me was, why the SHA was the same if the file is different???!
Attachments
The file hash should be different for the file you modified. If not, that suggests a bug in the software calculating the hash.
Regarding trying different named files, first you could try it yourself. Give the files new names, and in different alphabetical order. See if the same file then still seems louder.
Or make a copy of one of the files and rename it. Check it in WinMerge or some other free file comparison utility. Same, right? Do they play at different volumes? If you made them the same yourself and checked them yourself, why would you still trust foobar?
Finally, its easier to make two files that are different sizes, but the same volume, than different volume and the same size. If you want to see if file size has something to do with the problem, you could add some extra space to the end of a song, or truncate a song a bit early. Just don't edit the volume of the song in this case. You could perform these edits in Reaper. Just be aware that Reaper has to decode and re-encode mp3 format (but not if you use wav or other non-compressed format) in such a case, so you would be changing more than length and there is a chance re-encoding a file of different length could produce small, unknown changes depending on encoder behavior. Or just use wav, and see if you still have problems. You might not if the problem was with the MP3 decoder.
There are other experiments you could do yourself. Another option would be to contact the developers of foobar and report the problem to them. They would probably want you to send the test files to see if they can reproduce the problem. While this last option would take the longest, it is the only way the a bug in the software might get fixed.
Once you have tried some things, I can help you test the one theory you have suggested so far, but don't get stuck on that one theory.
Last edited:
I just made a smaller DB difference to see if the file size and SHA changes or not. I made 0.51dB difference and the file size didn't change! So is the SHA! But they sound different in FoobarABX...
I haven't read your post, I couldn't focus on the ABX (my son was pulling my hand for the walk around routine) so the result was 7/8 but it should show that I heard differences.
I'll walk around first...
Attachments
File size might not change, since is has to be written in some discrete number of blocks, often something like on 4kByte boundaries. But the file hash is a different matter. You could try checking the SHA hash with another tool: https://www.google.com/webhp?sourceid=chrome-instant&ion=1&espv=2&ie=UTF-8#q=sha hash calculator
I just checked your latest test files in WinMerge. They are identical, so there is no volume change, and SHA is the same. It's possible the mp3 encoder is normalizing the files to be the same, so your volume change is then cancelled out. You might need to leave a tiny part of the file, maybe the first measure at the beginning, the same high volume, then reduce the volume of the rest of the song by some small amount, and then encode to mp3. The encoder can't normalize the whole thing up then.
Of just use wav files, which you have more control over.
Of just use wav files, which you have more control over.
Last edited:
Can you save the files into something else (secretly) and let me guess which is which? I'm confused...
But the file-size and SHA changed when I made a file with not-too small differences? I didn't really pay attention to the file size, but during ABX I can perceive a perfect correlation between dB difference and ABX difficulty as well as subjective feeling about their difference...
I don't understand about this audio software algorithm...
I'm not sure I have a wav file. My audio files in this laptop are from blind test threads posted by XRK. I will check later.
Here are two files that should differ in level by 0.2 db. That's the way I set them, but haven't double checked. Change the file type to wav.
Edit: I just double checked and one file is 0.2 different from the other.
Edit: I just double checked and one file is 0.2 different from the other.
Last edited:
It's interesting, since I have the delusion that I (and others who enter my world of delusion and hallucination) can hear differences in opamps.
So, I'm assuming that I am using them "improperly" and I want to not do that. If you could explain how to not use them "improperly" then I can do that and perhaps revise my thinking on this matter??
Perhaps an example for two given opamps of your choice would suffice?
So, I'm assuming that I am using them "improperly" and I want to not do that. If you could explain how to not use them "improperly" then I can do that and perhaps revise my thinking on this matter??
Perhaps an example for two given opamps of your choice would suffice?
Skip this exercise.
It's futile.
If you want to hear "0.1dB" I suggest you pipe some pink noise into your speakers, and adjust the resistor pad on the tweeters by that amount. Of course if you have commercial speakers, you'll have to do a work around, unless the HF section comes out at the back for bi-wire...
- Status
- Not open for further replies.