Interesting, Pavel.
None of the programs I use can cope with it.
Audition, Nero, EAC, Foobar. I get different but consistent results every time I burn/rip.
I still have to find out which program is responsible for the inaccuracies.
First, I burned with Audition. I had to convert from mono
to stereo before burning.
Then, I ripped with EAC and compared both with Audition's Mix paste>invert. Files were different.
Next I burned with Nero, ripped with EAC, same result (different files)
Finally, I burned with Nero, ripped with Foobar. Again different AND different from EAC's rip.
What's the CRC of the original?
/Hugo
None of the programs I use can cope with it.
Audition, Nero, EAC, Foobar. I get different but consistent results every time I burn/rip.
I still have to find out which program is responsible for the inaccuracies.
First, I burned with Audition. I had to convert from mono
to stereo before burning.
Then, I ripped with EAC and compared both with Audition's Mix paste>invert. Files were different.
Next I burned with Nero, ripped with EAC, same result (different files)
Finally, I burned with Nero, ripped with Foobar. Again different AND different from EAC's rip.
What's the CRC of the original?
/Hugo
PMA said:
Sorry, am I being thick today? There's no ripping involved, I just open the WAV in Audition. It's a mono 44.1/16 WAV the spectrum looks like what you show. Do you want me to burn it and then rip it? This looks more like a test of pathological behavior. I don't know what's going on except some programs might have problems with mathematically generated files that violate the assumed sampling relationship.
I thought we were talking about dropped or interpolated bits.
john curl said:
VPI and Grado Platinum. The main thing is how the speed variations show up. Listening to the difference is even more fun. Another test I want to do, since I'm working on some mikes, is to play the same cut on repeat from a CD with a mike in front of the speakers and look at two consecutive plays, could be interesting.
John, I never asked you about that Wilson LP comparison with only a Soundstream in and out of the path. When I transfered both to my computer they didn't compare even closely. I meant the spectrum of short bits that I lined up carefully had 5 or 6dB differences in the short term spectral content and they didn't even get close to lining up.
WRT to ripping and burning, I have found to get the best results when using the bundled software that came with my Plextor (PlexTools and a OEM Nero 5). As I do also some CD authoring, I have a reference how the track boundaries should look like, and when I burn and rip back I get consistent and correct results, which are also stable when I rip back the actual production CD's from masters I did. So little chance that there is something going wrong here.
Pavel's JTEST result are strange, tracks offsets are to be expected, but manipulated data in the way you described isn't.
- Klaus
Pavel's JTEST result are strange, tracks offsets are to be expected, but manipulated data in the way you described isn't.
- Klaus
scott wurcer said:
It's possible to get dropped data in the burn/rip cycle under the following circumstances.
1) The original WAV file has no silence at the beginning or end
2) The read offset of the CD drive is not zero (all drives that I know of)
3) The drive cannot read into the lead-in or lead-out (all drives except a very small number)
OK, so burned it with Nero, ripped it with Nero, no problem. With a Plextor and an el-cheapo €20 DVD burner.
In the process I wrote a little tool to compare wav files, it skips the silence at the beginning of the files and tells you how much the ripper has eaten at the end of the file. Then it tries to find the offset between the files, and compares them.
I put it in attachment in case someone finds it useful. It's Python and you need numpy, though
No guarantees, lol.
In my case both drives used to rip ate about 600 samples at the end of the file. I should have used EAC instead of Nero, lol. Plextools crashes for some reason.
In the process I wrote a little tool to compare wav files, it skips the silence at the beginning of the files and tells you how much the ripper has eaten at the end of the file. Then it tries to find the offset between the files, and compares them.
I put it in attachment in case someone finds it useful. It's Python and you need numpy, though
No guarantees, lol.
In my case both drives used to rip ate about 600 samples at the end of the file. I should have used EAC instead of Nero, lol. Plextools crashes for some reason.
andy_c said:
peufeu said:OK, so burned it with Nero, ripped it with Nero, no problem. With a Plextor and an el-cheapo €20 DVD burner.
In the process I wrote a little tool to compare wav files, it skips the silence at the beginning of the files and tells you how much the ripper has eaten at the end of the file. Then it tries to find the offset between the files, and compares them.
I put it in attachment in case someone finds it useful. It's Python and you need numpy, though
No guarantees, lol.
In my case both drives used to rip ate about 600 samples at the end of the file. I should have used EAC instead of Nero, lol. Plextools crashes for some reason.
Aw c'mon Andy the lead in stuff is a mis-direction, a detail. In UNIX you can just move around and chop files at will and then line them up.
scott wurcer said:
It's not a detail if the number of samples of silence at the beginning (or end) of the WAV file you're burning, then ripping, is less than the drive offset, and the drive cannot read into the lead-in (or lead-out as appropriate). If there is no silence at beginning or end of the WAV file, and no lead-in/lead-out reading capability in the reader/burner, you'll never get the missing non-zero data back on a burn/rip cycle. It will be permanently lost.
EAC automatically fills in zeros for these missing samples if the appropriate option is chosen. But if the samples aren't actually zero, they're gone and you cannot get an exact copy. The only reason it works with the majority of music CDs is that there is plenty of silence at the beginning and end of the CD - more than the drive offset. But if you've got a computer-created WAV file for example, and it starts and/or end with no silence, it's impossible to not miss data at the beginning or end, unless your drive can successfully read into the lead-in or lead-out (depending on if the drive offset is positive or negative).
Look, I'm as "anti-audiophile-BS" as anybody here - perhaps more than most. The stuff I'm talking about is verifiable with controlled experiments.
Apparently, at least one manufacturer recognized this:
"An innovative hard disk master recorder to automatically store and catalogue your CDs.
Hush Technologies Deutschland GmbH – Atrium 4, room E106
Hush Technologies intends to set new benchmarks in the area of hard disk recording with their new “Hush HDR6” recorder. Users
may store and catalogue their CDs quite easily and conveniently with this digital recording device. It is operated by remote control;
titles and covers are maintained over the internet. Cataloguing takes place automatically by album, title or genre. In addition, songs may be downloaded over the internet from all available music portals. Digital master copies may even be produced by means of an additional piece of software. The CD is read repeatedly until no more read errors are present and subsequently compared with a check sum from the original master. Users are therefore ensured that every bit is reproduced in master quality when playing back from the hard disk. "
I like the bit where they compare the checksum of the rip with the original master over the internet. presumably, that original master is available *somewhere* ?
Jan Didden
"An innovative hard disk master recorder to automatically store and catalogue your CDs.
Hush Technologies Deutschland GmbH – Atrium 4, room E106
Hush Technologies intends to set new benchmarks in the area of hard disk recording with their new “Hush HDR6” recorder. Users
may store and catalogue their CDs quite easily and conveniently with this digital recording device. It is operated by remote control;
titles and covers are maintained over the internet. Cataloguing takes place automatically by album, title or genre. In addition, songs may be downloaded over the internet from all available music portals. Digital master copies may even be produced by means of an additional piece of software. The CD is read repeatedly until no more read errors are present and subsequently compared with a check sum from the original master. Users are therefore ensured that every bit is reproduced in master quality when playing back from the hard disk. "
I like the bit where they compare the checksum of the rip with the original master over the internet. presumably, that original master is available *somewhere* ?
Jan Didden
A topic for the Lp to CD Wiki
Ok Andy I understand, it's an important detail but not related to the sound of one ripped file vs another.
So I remembered last night one of the important things I wanted to discuss. Has anyone had success at doing RIAA in software after transfer? I think needing only a generic flat low noise preamp would be very appealing. In fact a single JC-2 stage might be all you need.
I'm shooting from the hip on this, but pre-RIAA wouldn't you have more bits in the mid's and high's. A double precision LOOOONG convolutional filter is probably a piece of cake with all the computing horsepower these days.
Ok Andy I understand, it's an important detail but not related to the sound of one ripped file vs another.
So I remembered last night one of the important things I wanted to discuss. Has anyone had success at doing RIAA in software after transfer? I think needing only a generic flat low noise preamp would be very appealing. In fact a single JC-2 stage might be all you need.
I'm shooting from the hip on this, but pre-RIAA wouldn't you have more bits in the mid's and high's. A double precision LOOOONG convolutional filter is probably a piece of cake with all the computing horsepower these days.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.