You gave them a ONE and they gave you a ZERO

[Karl J]

O.K. So I get this artist's CD back from replication, and in the middle of listening to the first song (on my stereo), I notice a tonal difference between the pressed CD and the masters I was used to hearing. After getting the master used for replication back, we loaded a song off of each (yes, the same song) into WaveLab and there was a noticeable visual difference as well.

I did some searching in the archives, but didn't find a satisfying explanation as to why.

So, if it's a normal thing to have happen, then WTF? Is this the price we pay for a medium that doesn't degrade with playback like vinyl or cassette (remember those)?

Do mix/mastering engineers prepare their product to compensate for this? Is it even possible? If not, then why the hell am I sweating over my sounds I'm recording, or my mixes for that matter, if they're giving you back something that sounds noticeably different? Then, to make matters worse, the kids go and turn it into an Mp3. Does it really still sound like I was using my API pres? Or if I had a Neve? Why shouldn't I just use my board's pres? They capture the sound accurately.

I thought digital was supposed to be what goes in is what comes out. A one is a one and a zero is a zero.

[Andy Krehm]

Originally Posted by Karl J

O.K. So I get this artist's CD back from replication, and in the middle of listening to the first song (on my stereo), I notice a tonal difference between the pressed CD and the masters I was used to hearing. After getting the master used for replication back, we loaded a song off of each (yes, the same song) into WaveLab and there was a noticeable visual difference as well.

I did some searching in the archives, but didn't find a satisfying explanation as to why.

So, if it's a normal thing to have happen, then WTF? Is this the price we pay for a medium that doesn't degrade with playback like vinyl or cassette (remember those)?

Do mix/mastering engineers prepare their product to compensate for this? Is it even possible? If not, then why the hell am I sweating over my sounds I'm recording, or my mixes for that matter, if they're giving you back something that sounds noticeably different? Then, to make matters worse, the kids go and turn it into an Mp3. Does it really still sound like I was using my API pres? Or if I had a Neve? Why shouldn't I just use my board's pres? They capture the sound accurately.

I thought digital was supposed to be what goes in is what comes out. A one is a one and a zero is a zero.

First thing to do is do a null test, ie, line up a track from the production master and a production CD sample accurate on 2 stereo tracks and invert one of them.

Play them together and if you hear anything, it means there is a difference.

Manufactured CDs should be identical to the Production Master. They may sound a little different, due to one being a CD-R, but when subjected to the null test, they should be the same.

If it fails the null test, then someone at the plant did some processing to it. No reputable plant will do this so it will be interesting to hear the outcome of this story.

[Masterer]

Originally Posted by Karl J

After getting the master used for replication back, we loaded a song off of each (yes, the same song) into WaveLab and there was a noticeable visual difference as well.

What is the audible [and visual] difference? Can you describe it?

[glassmaster]

Originally Posted by Andy Krehm

They may sound a little different, due to one being a CD-R,

Huh? you'll have to explain that one. (No metaphysics allowed)

[chrisj]

Another thing- if it passes a null test but still sounds wrong, then if you play it on a proper DAC it will sound right again.

As far as I know the only thing that can go wrong and still pass a null test is bad timing on the disc mastering causing poorly designed players to jitter. In that case the data is still there and any sort of playback that includes reclocking- even ripping the CD and reburning it- will 'fix the problem'.

You're probably looking at jittery timing and accurate data, if it seems like a big obvious problem but the data still nulls. That means at least you know it'll sound good again on proper kit :D

[Timecode]

If there is a visual difference between the two (as the poster said in the first post) then they're not going to null. It's definately still worth the test though - there was a thread a while back where someone was complaining about the same thing and after a whole lot of posts he nulled them and got zeros - end of thread.

I've sent two tracks in recently for magazine compilation discs, and both of them came back different - even after I phoned the plant for the second one and told them not to mess with it. Both times the changes were subtle but verifiable - as far as I can tell for the second one all they did was drop it 0.2dB, which makes no sense to me whatsoever. But there are a lot of people out there who reckon they know what's best and like to stick their fingers where they don't below.

Cheers
Matthew

[Karl J]

First off, thanks for the replies to my inquiry/rant.

As far as the differences between the original and the replicated CD, the replicated sounds... not as open in the top and slightly... muddy in the low-mids. Visually, it's wierd. The transient peaks look different. There's peaks were there weren't and none where there were. It was quite noticeable.

So, I did the null test. As expected, I heard plenty.

The artist asked the dupe facility about it. They said they'd check it out. Said they analyzed it, and all was o.k.

The artist says he can live with it, and I guess it still sounds o.k. after listening to it some more. However, for me, it still comes back to "hey, what happened to my hard work?" I just want to know what happened. Did they screw it up? Did they do some additional processing (they said they didn't)? Is this to be expected with replication? If so, that's BS. I wanted to see what I could find out here before I called them to inquire.

Any other suggestions?

Thanks,
Karl

[Dave Peck]

Regardless of why the files are different and regardless if they 'fess up to doing some unauthorized processing (they almost certainly did), the point is that you have the evidence that the files do not match. Based on that alone, they should scrap the job and do it again with no argument. Don't let them pull that BS. Any reputable replicator would agree.

DP

[lucey]

Originally Posted by Dave Peck

you have the evidence that the files do not match. Based on that alone, they should scrap the job and do it again

Reject it - they messed with it. No question!

<L>

[foldback]

Originally Posted by Karl J

First off, thanks for the replies to my inquiry/rant.

As far as the differences between the original and the replicated CD, the replicated sounds... not as open in the top and slightly... muddy in the low-mids.

Any other suggestions?

Thanks,
Karl

Yea, tell us who did this to you so we can avoid them :-)

[stevep]

Originally Posted by foldback

Yea, tell us who did this to you so we can avoid them :-)

.

[jamesp]

Originally Posted by Timecode

I've sent two tracks in recently for magazine compilation discs, and both of them came back different - even after I phoned the plant for the second one and told them not to mess with it.

This might be veering off topic a bit but If I'm creating a compilation disc for a client then I am unlikely to be able to use a track if I wasn't allowed to change its level. How do you know how one single track is going to sit when put beside tracks from all sorts of other people? With one compilation that I worked on recently I found that some tracks had to be attenuated by up to 6dB compared to the quieter tracks.

I wouldn't change the sound of any of the tracks without consulting the artist/producer first but level changes are fair game for a compilation.

And back on topic again - the only difference between the master CD and pressed CD that I would accept would be a change in offset. Samples on a pressed CD are often offset by a different amount to the samples on a CD-R. This is very dependent on the drive used to create and extract the data and also depends on the mastering system used at the plant. That's why you have to manually line up the extracted audio with the original audio when doing a null test.

Cheers

James.

[pan]

Any chance that some donk at the label sent the mixes or "premastered" listening copy and not the masters???

[Jules]

I recomend reading the eq column piece entitled

CDs Give Me The Jitters!

over here

http://rogernichols.com/

[Karl J]

DOH! Thanks for the heads-up about the sample offstes, James. I was in a hurry when I first did the null test, and DIDN'T make sure they were sample-accurate lined up. Last night I had more time and did the test again. They actually DO null out! Damnit! How anti-climatic.

My bad. Sorry for misleading you all.

However, it sounds a bit different! Both the artist and myself noticed the difference listening to the cd on different players. I will follow some of the other suggestions, such as Chris J's

Thanks for helping me learn.

Karl J

[Karl J]

I just read the Nichols article. Thanks for the link Jules! That was a good read. And explained a lot. I'm still amazed that it can sound and look different, yet null out. At least I'm not the only one to have experienced this problem. Also reassuring to know that even though it nulls out, I'm not crazy and can trust that my ears are indeed hearing a difference.

Still not sure why it is acceptable.

For the record, the artist used Disk Makers for the replication. He's not with a label, but is shopping this CD with the hopes that someone might pick it up. It's not a big budget project by any means. In fact, I did the project for a very low rate. The artist is a good friend and great to work with, and I like his songs. I'm just trying to make sure he's shopping a good sounding product.

Can't say I'd recommend he use DM in the future.

Karl J

[jamesp]

Originally Posted by Karl J

Still not sure why it is acceptable.

A decent CD player will ignore the jitter. Pressing plants assume that everyone uses decent CD players but the problem is that very few CD players are totally immune to jitter on the disc. Oddly enough - pressed CD's sound better on my Philips player than CD-R's so things can go both ways sometimes.

Cheers

James.

[Lagerfeldt]

Originally Posted by Karl J

I'm still amazed that it can sound and look different, yet null out.

It certainly can't look different and still null, not unless your DAW is messing with the graphical representation.

[Darius van H]

It's possible that the CD's sound the same and you imagined the differences.

This is especially likely if you were expecting to hear a difference before hand - like if you read about something like this before on a forum.

Did the artist hear the differences too independantly from you, or did you first tell him to listen for differences?

[narcoman]

Originally Posted by Lagerfeldt

It certainly can't look different and still null, not unless your DAW is messing with the graphical representation.

They certainly CAN ! Not many of the DAWs represent waveforms with graphical accuracy! Its a sampling! Only when you zoom right in are the waveforms more accurate. Nuendo, Logic (up to 7, not used 8) and Cubase are aweful at this - very difficult for zoomed out editing - you can't trust the graphics! not a problem when you zoom in - even just a bit. PT, Sadie and RADAR are much better at this.

As for sounding different - technically they wont (the CD-R's and the pressed CD) HOWEVER CD-R's often produce more errors on playback since they are harder to read. When you do a waveform import on a DAW the errors shouldnt get in the way and the sounds will null.

NOT the same when you playback from CD player. I'd argue that your CDs will sound better or at least more like your original DDP or bitfile than your CD-Rs. The CD-R's, on most players, will not sound like your original mixes - only a subtle difference but they will contain many more playback errors.

[Thomas W. Bethe]

One of my clients used a local company for their replication. The CD that came back sounded nothing like the CD I mastered. We did a null test and it was not even close. I told the client to call the company and ask what they had done to the master after it left here. They told him that their "mastering engineer" had auditioned the disk before they sent it out and had adjusted the EQ and levels to be more "competitive". This was AFTER they were told the disk was mastered and that the client wanted it sent "as is" They apologized but said what was done was done and they could not alter the outcome and that if my client wanted it redone he would have to pay for it. Needless to say I don't send any business to them anymore and have warned many client about using them but they offer really really cheap prices and that attracts a lot of people. I just did a mastering for a harp client and she is going use them for the replication and I hope and pray that they don't screw this one up. But given their track record they probably will. The harpist main reason for going with them was.....they had the cheapest prices for 4 color printing......Oh well!

[twwalsh]

from my experience the zoomed out waveform previews in DAWs can be REALLY inaccurate. if you import the file from two different sources, the previews may look different from "far away", but if you zoom in and it gets re-drawn, it will look the same.

[Greg Reierson]

Was there a brokerage company between you and the replication plant? I've seen lots of strange things happen in the clumsy hands of some of these guys.

And, no, there is no reason data identical CDs should sound different. If they do you have a playback issue, not a content issue. Load them into a DAW and the difference will magically dissapear....


Greg Reierson
Rare Form Mastering

[Lagerfeldt]

Originally Posted by narcoman

They certainly CAN ! Not many of the DAWs represent waveforms with graphical accuracy!

You should probably read the last part of my sentence again. ;-) That's exactly my point.

[Karl J]

Yippeeeee!!!!! My old thread got revived!!!!!!

Well, the DAW was Wavelab. I can't recall how much I zoomed in on the waveform. But I had three different ones to look at. The original file (dumped from 1/4" and hit with L2), the imported track of the master, and the imported track of the duplicated disc. The original mastered file and the imported master looked the same.

As far as the sound: It was a noticable difference on multiple systems, with the original CDR sounding better. It was noticable to me within seconds of first playing the duplicated CD. I asked the artist if he noticed a difference and he replyed "yeah, I thought it didn't quite sound the same". My wife could hear the difference, although she couldn't tell which was which.

Does this mean I just love the sound of disk errors and error correction?

[Marty J]

Originally Posted by Karl J

Yippeeeee!!!!! My old thread got revived!!!!!!

Well, the DAW was Wavelab. I can't recall how much I zoomed in on the waveform. But I had three different ones to look at. The original file (dumped from 1/4" and hit with L2), the imported track of the master, and the imported track of the duplicated disc. The original mastered file and the imported master looked the same.

As far as the sound: It was a noticable difference on multiple systems, with the original CDR sounding better. It was noticable to me within seconds of first playing the duplicated CD. I asked the artist if he noticed a difference and he replyed "yeah, I thought it didn't quite sound the same". My wife could hear the difference, although she couldn't tell which was which.

Does this mean I just love the sound of disk errors and error correction?

Asking people if they can hear a difference between two things is a really bad way to find out if there actually is a difference.

[Karl J]

The artist had listened to it and reached his own conclusion prior to me discussing it with him.

Regardless, I could tell almost right away.

[Rubito]

I have heard that some CD player DACs get finicky and can overload in some circumstances if the level goes above -0.2 but that could be an old wives tale, but it would explain why somebody would want to drop the level 0.2db (but not why they think it's their problem).