Topic: ratshack digital signal converter testing :)

[goose]

i'm on the phone right now with paul, and sony does indeed still have the 7-pin / mini-toslink cable.  i've got one ordered. (snip)

These mini-toslink 7 pin cables are still available, I ordered one from B&H two weeks ago, they have them in stock as well.  I think the part number is POCDA12SP or something like that - the "SP" at the end indicates the mini-toslink in and out.

[dr.ph0b]

i'm on the phone right now with paul, and sony does indeed still have the 7-pin / mini-toslink cable.  i've got one ordered. (snip)

These mini-toslink 7 pin cables are still available, I ordered one from B&H two weeks ago, they have them in stock as well.  I think the part number is POCDA12SP or something like that - the "SP" at the end indicates the mini-toslink in and out.

u can order them direct from sony too:

http://servicesales.sel.sony.com/web/viewItemDetail.do?operation=getItemDetail&itemID=24212&category=8&categoryName=Cable

[dr.ph0b]

I haven't seen any mention by Dr. Phob that he's listened to the complete recorded WAVs (20+ hrs) to identify if they contain audible dropped samples or artifacts, nor have I seen indication he's checked them for bit accuracy against a control file.  IME, it's much easier - and more authoritative - to verify bit-accuracy than casually listen to the WAVs played back. 

sorry sometimes work keeps me busy from checking back on all these forums   damn work

anyway, how would i _test_ for bit-accuracy?  does wavlab have a feature?

[F.O.Bean]

youd have to compare wavs in EAC or something

[Brian]

youd have to compare wavs in EAC or something

+T

[F.O.Bean]

+T backatcha bud

[dnsacks]

Testing for bit-accuracy --

Actually, open both wavs up in soundforge, find and mark a common sample point a few seconds after the very beginning of the wavs.  Select and copy one of the wavs from the common sample point forward.

then, go to the other wave and instruct soundforge to mix the inverse (invert the wav? -- don't have soundforge here) of one wav with the other.

The result, if both are identical will be a ruler-flat line (zero db) for the lenght of the entire mixed wave (since music plus the inverse of the music would be silence).  Any dropped sample, etc. would cause a spike in the mixed wave, or "music" to appear along the wave (albeit strange sounding distored music)

hope this helps

[sleepypedro]

alternately you could just md5sum both wavs, and provided you started and ended your test clips on the same samples AND the device in question is bitperfect AND both grabs were read with no errors, you'll get identical checksums

[Lil Kim Jong-Il]

Instead of md5 (you better be exact on start and stop sample) and wave inversion sum (same problem), I'd go with the EAC suggestion if your comparing 16bit samples.  The EAC results actually tell you the problem - missing sample, duplicate block, etc.  Both the other will work, but it's up to you to perfectly time align the areas of interest.

[cwfen]

Updates?

[dr.ph0b]

sorry.. been getting killed with work.. ill try to do something this weekend

[Brian Skalinder]

However, virtually no manufacturer follows the S/PDIF standard. 

I assume you meant the electrical interface?  Can you elaborate?

Yeah, I should've clarified - electrical interface.  I should also note that variances in the S/PDIF electrical interface causing DFC problems is merely conjecture on my part as I haven't done the testing to determine the root cause of the problems with some DFCs.

[cwfen]

...any results on the bit-for-bit testing?

[johnw]

What was the final word on this?

[hoobash]

I got one. I am not sure how to test. Maybe I should go back and read the thread again?