Mass CD conversion

[kwesloh]

OK Slutters. I've been tasked with ripping a very large CD collection into a broadcast playback system and need to accomplish the following with every disk with the priority indicated by number:

1. Fastest possible rip.
2. No (audiible) loss of sound quality.
3. Need to normaiize all selections to a fixed in-house level standard.
4. Ability to Trim head or tail of any piece.
5. Fade in or out after trimming
6. Sample-Rate conversion to 48K

Cost is not necessarily an object if the performance gains of the given system outway the cost.

Oh yes...It needs to be basic enough that interns and librarians can do the work from an SOP.

I've benchmarked the ripping times for all of our in-house systems: Wavelab, Protools, DSP Quattro and Bias Peak with a couple real surprises
but would really like input from anybody with experience ripping lots of data.

Thanks in advance for any advise you can muster.

-Kyle

[Karl Jackson]

If cost is no object, why not outsource?

If you still want to do this, put a system together that will rip multiple copies simultaneously. The state of the art solution is Quadriga, but I use Exact Audio Copy, with two instances running, with two Plextor Premium drives.

Once ripped, there are various other programs that can batch the trimming and SRC. Goldwave comes to mind at the cheap end, but I would use Wavelab since you already have it.

kj

[kwesloh]

These are great ideas, Karl. I will be looking at Exact Audio Copy as a software to benchmark. I have looked at the Quadriga system in the past for reel-to-reel archiving and it is very impressive. It may be worth requesting a demo unit if they allow for that.

Outsourcing is something we have thought about. Unfortunately, our needs are more custom than outsourcing would allow for. Tracks need to trimmed in a way that makes sense musically. If you allowed a piece of software to strip quiet sections (below a certain threshold for instance) it would potentially be taking out quiet passages within the music. Also, when you rip a Mozart symphony from CD it would typically be ripped as four separate tracks which represent the four unique movements. software that tops and tails files would do so for each file. This would remove all of the carefully created space between movements and the files would be rendered unusable.

The normalizing function also needs to be intelligent. We don't want the software to change the gain on a piece of music that was originally mastered too low for instance. We need to be able to identify peak passages exactly, change to a specific peak level setting according to our building and network standards and then listen to the results to make sure it all still sounds good. This is more "gain-change" than "normaiizing" function

Another key part of this transfer that I didn't mention in my original post is meta-data entry. Each file needs to be assigned a name that correlates to an identifier field in a pre-existing data base. If this is not done during the rip on a track-by-track basis we are going to be in a deep hole when the tracks arrive back in-house.

We have looked buying a classical library off the shelf but the cost is, let's say, highly prohibitive.

-Kyle

[Karl Jackson]

Originally Posted by kwesloh

Tracks need to trimmed in a way that makes sense musically. If you allowed a piece of software to strip quiet sections (below a certain threshold for instance) it would potentially be taking out quiet passages within the music. Also, when you rip a Mozart symphony from CD it would typically be ripped as four separate tracks which represent the four unique movements. software that tops and tails files would do so for each file. This would remove all of the carefully created space between movements and the files would be rendered unusable.

Oh, I see. It sounds like you almost have to have a manual top/tail task as part of the process then. Just to play devil's advocate, if you would have otherwise played the tracks straight from CD, can you not simply play the ripped file w/o editing?

Originally Posted by kwesloh

The normalizing function also needs to be intelligent. We don't want the software to change the gain on a piece of music that was originally mastered too low for instance. We need to be able to identify peak passages exactly, change to a specific peak level setting according to our building and network standards and then listen to the results to make sure it all still sounds good. This is more "gain-change" than "normaiizing" function

If you're careful to avoid clipping, you may be able to batch this using RMS normalization, but if you're going to have the file open in an editor anyway, and you have a controlled, predictable monitoring environment, I suppose it's just as easy to do manually. Another option is to integrate ReplayGain scanning into the ripping process, and then using a ReplayGain compatible player to level match for you in the playback process. This has the added benefit of preserving the original levels.

Originally Posted by kwesloh

Another key part of this transfer that I didn't mention in my original post is meta-data entry. Each file needs to be assigned a name that correlates to an identifier field in a pre-existing data base. If this is not done during the rip on a track-by-track basis we are going to be in a deep hole when the tracks arrive back in-house.

Are you talking MODS/METS/PREMIS style meta-data, or just a single unique ID? If the former, best of luck. Otherwise, I enter an accession serial ID into EAC, instructing it what kind of folder structure to create (ARTIST/ALBUM/##-TRACK) and how to tag the file. You're probably way ahead of me on this, but there are some very powerful apps out there for tagging and renaming. I use foobar2k and Tag&Rename regularly for this.

As an aside, you may thank yourself in the far distant future for maintaining a lossless rip of the CDs as a preservation format. EAC can rip to WAV and MP3/FLAC/etc... in one step.

kj

[roonsbane]

Hey Kyle,

There is a lot of good info thus far here. I was wondering if the mastering guys might also have some good info on this subject. I know posting in multiple forums is a no no, but I wonder if Steve might make an exception here. The remote sluts are usually my first resource, but this would seem to be more of a mastering type question.

Cameron

[kwesloh]

Originally Posted by Karl Jackson

Just to play devil's advocate, if you would have otherwise played the tracks straight from CD, can you not simply play the ripped file w/o editing? kj

Our hosts spend a certain amount of time identifying the actual start point of a piece of music before it is played every time they cue up a CD. Classical mastering houses /producers love to have ambience roll ahead of the music. This can be as short as a fraction of a second to as many as 3 full seconds. While we all appreciate the ambi when we're home listening to CDs on our nice stereo systems, more than a second of ambi sounds more like dead air than a "sense of space" when a host cues in a a cut. The same thing applies to the end of certain pieces. Some are cut tight and some have a vast expanse of room tone at the end of the file. This would result in dead air if the host were to simply throw the pieces together in a playlist. Right now it's up the host to make sure this doesn't happen. It puts a lot of pressure on the hosts and eliminates the possibility of using an automated playlist which will be part of our standard.

Originally Posted by Karl Jackson

Are you talking MODS/METS/PREMIS style meta-data, kj

No. As I said, we have a pre-existing database and will be simply renaming the title field to link it directly to the existing catalog database used to to track and program the CDs for air.

Originally Posted by Karl Jackson

If you're careful to avoid clipping, you may be able to batch this using RMS normalization, but if you're going to have the file open in an editor anyway, and you have a controlled, predictable monitoring environment, I suppose it's just as easy to do manually. Another option is to integrate ReplayGain scanning into the ripping process, and then using a ReplayGain compatible player to level match for you in the playback process. This has the added benefit of preserving the original levels. kj

Again, We would have to listen to and adjust each ReplayGain level setting after the rip and update the meta accordingly. My experience with Itune's version of automated gain-normalization has been less than positive.

[boojum]

foobar2000 has a lot of ReplayGain options and applies RG in decimal increments to two decimal places, i.e., 8.97+. It also has a library function but it may be too simple for your uses. On the up side it is free with plug-ins which can further enhance its usefulness. I would check it out.

foobar2000