The key is systematic, I assume the files on the disk are exactly identical in the sense of there are no extra or too few leading zeros as stored on the disk. The differences in the analog playback would then be randomly applied to either file.
The leading zero problem is real, there are rippers that fail to maintain the exact count only caring about the actual audio data. These files can fail a checksum.
Till he retires or quits the current job.
I've seen recognizable pattern by him and others who are in the same business.
This would be too easy, but I must admit there is plenty of the usual here.
Merrill Audio Advanced Technology Labs, LLC. "Audio Purity"
More to the point Jakob asked about ABX vs A/B. If the results were significantly different why would I not question both? That is if A/B produces more correct answers why not search for a hidden cheat or bias?
In any case questioning test methodologies against desired results seems to me to be the wrong thing to do.
Converging on anything here is going to be difficult.
In any case questioning test methodologies against desired results seems to me to be the wrong thing to do.
Converging on anything here is going to be difficult.
By all means, please do, or we all can.
Any serious study needs to take great care to make sure it done properly. It isn't a trivial matter, but an experienced researcher capable of acting as PI ought to be able to do it. I would sure hope they could.
Of course, it could also be that small changes to ABX, as I have described, could do a lot to narrow any difference with A/B.
There might turn out to be a perfectly plausible scientific explanation as we continue to understand more about how brains work, doesn't have to be cheating, but sure let's be very careful and not make mistakes by letting cheating slip through.
Last edited:
While I disagree with Jakob plenty about interpretations and magnitudes of things, I will continue to interact with him because he doesn't play games with discussion. And he backs up what he says often enough for my satisfaction. More than we can say for many members (myself probably included)
Other members, well, I have an ignore list for my own sanity.
Other members, well, I have an ignore list for my own sanity.
Last edited:
The word cheating has negative connotations, but it should simply be recognized that innocent flaws in test methodology can be exploited by very aware test subjects.
A while ago Dan posted the two bit identical files in another thread after much delay. I tried to differentiate the files with a half decent usb dac with headphone out (thorsten's iDAC) and three half decent headphones.
I wasn't able to tell the files apart using foobar ABX even thought Dan said that the sound/light crew he was setting up with at the time was able to tell the difference from casual listening while working.
I consider that some sort of data point.
Cheers,
Jeff
I gather that includes a few smokes, brews, and chatting up the ladies as well.
I wasn't able to differentiate them either.
Dan also said he would try to differentiate some files I sent him, by using Reaper as the playback engine. I went to the trouble of writing it all up with pictures and text to make it as easy as possible for him. After all the trouble I went to he never replied. Either he couldn't tell them apart or never bothered to try, is the way it tends to look from here. Needless to say, I was disappointed by the failure of Dan to ever say anything about it again.
I will get around to repeating uploading some content to dropbox and we can try the experiment again.
In the meantime, I have repeated the experiment with a number of files on various thumb drives and I still get the same result, and with others able to differentiate the pairs of files also, so yes more (anecdotal) data points.
Same PC USB socket, same Thumb Drive, HD file copied across, unplug/replug to clear Clipboard, second copy transfer of HD file written with filter mix on TD USB plug pins/pads.
Differential of files playback with goop and differential of files playback without goop, ie four sounds, does not matter which playback PC.
I don't know the technical reason for this seemingly 'odd' result, but it is true that multiple others can/have readily identified and described differences and preferences and similarly.
The cause must be at Flash Memory internal write/read/error correct level is the only rational/accepted explanation I can think of.
BTW, yesterday I had a closer listen to the MF/Foil pair of files put up a while back and I do find a readily audible difference that I was able to score 10/10 Foobar ABX first and last attempt after listening to the two files in the background on extended repeat while I did kitchen work......by then I had heard the passage enough times that I didn't want to hear it any more !.
Dan.
- Status
- Not open for further replies.