Topic: bit accuracy testing help

[JNT]

I'm trying to test my sound card for bit accurate prior to transfering DAT tapes.
From the tests I've done so far the card does not appear to be accurate.  I would like to make sure I'm testing this correctly before I start looking for a solution to the problem.


Does this test procedure look correct?

1) play a 44.1k DAT tape through the sound cards digital input and record to the HD (I'm using CD Wave)
2) Repeat step 1
3) Use the compare wav function in EAC
4) If they are correct shouldn't I see an error that one file is slightly longer than the other but otherwise no message?

Using the cmp -l feature in SHNTOOL as a second test I get a long list of differences between the 2 files.



Thanks,
Joe

[Brian Skalinder]

Sounds okay to me.  I know it's obvious, but you've checked your soundcard config settings to ensure it isn't configured to resample on the fly, right?

Are you trimming the WAV files on your PC to ensure you're working with files of the exact same length, starting at the same bit and ending at the same bit?

[JNT]

Sounds okay to me.  I know it's obvious, but you've checked your soundcard config settings to ensure it isn't configured to resample on the fly, right?

Are you trimming the WAV files on your PC to ensure you're working with files of the exact same length, starting at the same bit and ending at the same bit?

Thanks Brian. Yes the soundcard settings should all be correct.  No I have not trimmed the wav file.  I did a search earlier today and found a few posts about that but it wasn't clear to me.  If I'm using EAC for the compare shouldn't it just compare the middle part of the file and show the extra part and the beginning or end as one file being longer than the other?  What about one if the SHNTOOL options for cmp?

If this isn't possible do you have any suggestions for preparing the file to compare.  I have Cool Edit 2000 and N Track installed.

Joe     

[Brian Skalinder]

Not certain, but I think EAC and SHNTool's WAV compare features are simple comparisons - does the first bit in each file math?  The second?  Third?  And so on.  I don't think they're capable of matching up bits 1-20 in WAV 1 and comparing against the corresponding bits in positions 31-50 in WAV 2.  So...I believe you'll have to trim the WAV files.

Using CE2K, zoom into the wavform and find a spike like a drum-hit or some such.  Drill in and trim all the samples in the wavform before the top most sample in the spike.  Repeat for the exact same spike/sample in the second file.  Do it again near the end of the file.  FWIW, I found it easier than I expected to line up the wavforms and find the exact same sample in each file.

[JNT]

That did it    I trimmed the file and  ran it through wav compare.  No errors using the Emagic A62.  Now to check the Echo Mia.

Both of these devices are know to be bit accurate but I wanted to make sure there wasn't anything it my system causing less than accurate results.

Thanks +T


Joe