Topic: Archos Gmini 120

[JasonSobel]

hey folks, I haven't done too much with archos recently.
However, I do have the latest firmware installed, and I'm taping a show tonight, so I'll run into the gmini for a least one set, probably both. (and of course, I'll run into my D8 as well.  both from the v3)... I'll post the results shortly afterwards.  

I am going to try and figureout how to do a bit accuracy test this eve, but I'm just happy my recording sounds good

to do a bit accuracy test, you've got to also record on something which you know is bit accurate.  I'm not sure how many digital outputs the ua5 has.  perhaps you can record onto your computer simultaneously.  or run a DAT.

then it's simply a matter of trimming both files, so that the each start and end on the exact same sample, and compare the files.  EAC was a .wav compare feature which is nice and easy...

[Matt Quinn]

JAson you going to D'elf tonight?

[JasonSobel]

yes, I'll be at d'Elf

[timP]

OK, will see what I can do...
can get at my DAT, but will have to find a way to transfer......
good luck with your go at it.......

[JasonSobel]

this morning, I quickly threw the .wav files from the gmini recording of last night's Club d'Elf show onto my computer and took a look.  obviously, I haven't transferred my DATs yet to do a real comparison. however, it was clear that there were several glitches in the recording.  and they were definitely audible as "clicks"  These glitches seem to exhibit the same behavior as found earlier, just a few missplaced samples, sometimes just in one channel, sometimes in both.  I'll need to compare the .wavs transferred from the DATs in order to get the exact time intervals between glitches.

clearly, though, the firmware upgrade 1.7.0 did not fix the .wav recording issue (although, it never claimed to fix that issue).

[timP]

bummer  

[timP]

do you hear the clicks when you listen?
I have wavlabs... when I put a file in it, should the clicks and pops show up as jumps up or down?

[JasonSobel]

yes, these glitches are audible as very short "clicks"

and yes, I can see them in the waveform (I'm using Cool Edit 2000)...
back when I did my original tests, I posted some screen shots of the glitches.  those pictures are on this thread somewhere, so just go back a few pages to find 'em.  the glitches appear as a very short, tall, spike in the waveform.

they are more noticable during quieter passages (although they also occur during louder segments).

[timP]

thanks... damn this is huge thread to dig thru.......

I have found 2 very very fast pops is listening to my most recent recording, but they are more drop outs and don't show spikes on the wave form...
in fact I have no out of place spikes, just those 2 drop outs.... so far that is, I am gonna dig into it this eve.......

[JasonSobel]

I have found 2 very very fast pops is listening to my most recent recording, but they are more drop outs and don't show spikes on the wave form...

The issue is really misplaced samples.  but, in my experience, the gmini does not put the misplaced samples in the same place.  the majority of the time, the sample is way out near the 0dB point, and appears as a spike.  However, I've also seen when the sample is right on the -infinity dB line in the center of the .wav form (this may appear to be a drop-out)

either way, the unit is misplacing the sample.  let us know what your findings are once you look into it more thoroughly.

[timP]

anyone done any Bit Accuracy test with MP3...

maybe if there was the same porb there, Archos would be quicker to deal with the overall BA issue since most users prob will only use MP3's..
just a thought.............

[John Kelly]

Only problem is that mp3's aren't bit accurate.

[JasonSobel]

no, I haven't tested any .mp3 recording, although I see your point.

Only problem is that mp3's aren't bit accurate.

I think what he meant wasn't to test for "bit accuracy" with the .mp3 recording, but rather, if the unit put audible clicks and glitches in a recorded .mp3 file.  if it does put audible clicks and "glitches" in a .mp3 file, then they might be more interested in fixing the problem quicker (because most people probably want to use it for .mp3's), and then as a side benefit, the problem of glitches in .wav recording would be fixed as well.

please correct me if I'm wrong, but that is how I interpretted the idea...

[timP]

nope, you got it  right....
just a random thought while drivin' around...
damn taping on the brain

[Sharky]

am I gonna have to personally fly to france and kidnap this firmware programmer?!
Circuit City has the Gmini120 for $199. Great deal if the damn thing worked.