Skip to main content

Notice

If you are using a Hotmail or Outlook email address, please change it now, as Microsoft is rejecting all email from our service outright.
Topic: Burning audio CD's, comparing software (Read 6717 times) previous topic - next topic
0 Members and 1 Guest are viewing this topic.

Burning audio CD's, comparing software

I finally took on a task to find a better sounding (=no distortions) remix/remaster of Metallica's Death Magnetic album.. and I found one (I'm not sure if I can advertise it here) that finally allows me to _listen_ to this rather good album. Then I burned it to a CD with CDBurnerXP and found out that the CD-Text title for the 1st track was taken from the filename (the source WAV files had no metadata) though I hadn't set any CD-Text in the program, other burned tracks had no CD-Text. Then for some reason I can't remember anymore, I extracted the burned CD and compared them to the source WAV's. Here's the result:


Does anyone understand what CDBurnerXP did? I've hard time deciphering the results above. I burned 2nd one, with same settings/options, to a CD-RW and the result was identical. Both burns completed successfully according to the software.

This oddness prompted me to test other burning software. All CD's were burned with DAO (/SAO), no (2 second) gaps & finalizing the disc. Here's the same compare results from other burning apps (EAC, InfraRecorder & Burrrn):


I can explain A) the missing 6 samples = my burner has a -6 write offset, B) longer lengths = the source WAV lengths were not "correct lengths" for a CD (can't remember the proper jargon/technics).. only thing I can't explain is the "6 repeated samples" in the 9th track, in the source WAV side. I believe that the CD was properly burned, but does anyone know what that means?

Here's couple of screenshots (I'm gonna put them as thumbnails to keep the post size down) from IsoBuster showing the CD-Text failure in the CD burned with CDBurnerXP.. and the size differences of the tracks:

1) 1st CDBurnerXP burn with the CD-Text bug in 1st track


2) 2nd CDBurnerXP burn, this time no CD-Text bug, otherwise identical to the 1st one


3) CD burned with EAC


4) CD burned with Burrrn


5) CD burned with InfraRecorder



Ok, the difference of the file sizes (and TOC in EAC logs) were expected between CDBurnerXP and e.g. EAC rips. But why is the 1st track of the CD burned with InfraRecorder noticeably smaller when compared to CD's burned with EAC or Burrrn? When I compare EAC rip logs between the three burners (EAC, InfraRecorder & Burrrn), the ONLY difference is the date when the rip was made. What does this mean? Obviously the "audio parts" of the tracks are same according to EAC compare, so what is it then, what is the difference? EDIT: Wait a minute, the CD's burned with EAC and Burrrn have CD-Text's while the CD burned with InfraRecorder has not, is this the reason for the size diff?

CDBurnerXP is obviously out of the game (finally I might say, it has had long standing issues (the 3rd party burning library) and for past 1-2 years when I've used it very seldom, it has failed virtually every time in some way), it burns the audio CD wrongly (though I can't explain what it does exactly) and the CD-Text problem (can't reproduce).. but is InfraRecorder also not to be trusted because of the 1st track size difference?

Your thoughts?

Burning audio CD's, comparing software

Reply #1
Have you checked whether your source files end on frame boundaries?
Is 24-bit/192kHz good enough for your lo-fi vinyl, or do you need 32/384?

Burning audio CD's, comparing software

Reply #2
Have you checked whether your source files end on frame boundaries?
No, don't know how and don't know about the stuff. I quickly googled help for Audacity and found some but couldn't get the "snap on" (frames) to work. Isn't this exactly what I referred to in my "B) longer lengths = the source WAV lengths were not "correct lengths" for a CD" text? If yes, the answer is obviously that they do not end on frame boundaries. Then what? To which of my concerns is this relevant? 1st one?

Summary of my questions:

1) What did CDBurnerXP do differently (I believe erroneously) compared to the other burners when burning my WAV's?
2) What does it actually mean when EAC compare reports "6 repeated samples" in the 9th track, in the source WAV side for CD's burned with EAC, InfraRecorder & Burrrn? I guess it's nothing to worry about. Apart from this, I see that CD's burned with EAC, InfraRecorder & Burrrn were correct/expected (when comparing the audio parts).
3) Why is the 1st track of the CD burned with InfraRecorder noticeably smaller when compared to CD's burned with EAC or Burrrn? (IsoBuster screenshots) Is it because the CD's burned with EAC and Burrrn have CD-Text's while the CD burned with InfraRecorder has not? EDIT: Tried burning with CD-Text titles with InfraRecorder and the 1st track file size (in IsoBuster) grew a bit longer than compared to EAC or Burrrn.. so I guess that was it.. don't understand why it isn't the same size now.. but I guess it's nothing to worry about, at least the audio parts match/were burned correctly. Ok, it's Isobuster that's borked! If you right-click and show properties of a track in the left-hand side panel, then close the dialog and select another track in the same panel, the size of track changes! If you open the properties dialog again for the same track, the wrong changed size is there too. And a variation: If you right-click and show properties of a track in the RIGHT-hand side panel, the size changes in the right-hand side panel immediately, but the "old" value is shown in the just opened properties dialog. Refresh resets the values. Bugs bugs bugs, everywhere I turn, bugs bugs and more bugs.. I've been a computer "enthusiastic" for decades but for past few years I've been growing this dissatisfaction of using computers (also means mobile phones.. and other "computers".. digital TV (& digibox (=set-top box?))).. and my experience shows that it gets worse every year.  Ok, enough with that.. that concern/question is solved now.

Thanks for replying.

Burning audio CD's, comparing software

Reply #3
Ok, did more testing (thanks greynol for the nudge to the right path): CDBurnerXP failed (I still insist that it failed, feel free to refute ) because it burns source files that do NOT end on frame boundaries differently than EAC, InfraRecorder & Burrrn. I tested this by burning the already once burned (with EAC) & extracted WAV's with both EAC and CDBurnerXP. The extracted files from those CD's were identical.

So this leaves:

1) What did CDBurnerXP do? I'll let the author to figure that out.
2) What does it actually mean when EAC compare reports "6 repeated samples" in the 9th track, in the source WAV side?

Burning audio CD's, comparing software

Reply #4
2) What does it actually mean when EAC compare reports "6 repeated samples" in the 9th track, in the source WAV side?

No clue. I use Audition when I want to see how files like these differ.
Is 24-bit/192kHz good enough for your lo-fi vinyl, or do you need 32/384?

Burning audio CD's, comparing software

Reply #5
One set of the files are 6 samples longer. This can be reported 2 ways. As 6 samples are missing in the shorter version. OR... since the beginning is silence you could say that 6 samples (of silence) are repeated in the beginning of the longer version.

Burning audio CD's, comparing software

Reply #6
Thanks for the answers! I compared with Audacity and all I see is 6 samples "shifted" (-6) because of the 0 write offset correction used when burning.

Quote
As 6 samples are missing in the shorter version. OR... since the beginning is silence you could say that 6 samples (of silence) are repeated in the beginning of the longer version.

Thanks, this made me scratch my head for a while, especially when I looked at the compare shot, but then I finally understood. It's easy to get fooled (fool me once.. fool me twice, shame on you ) by the EAC compare messages. E.g. look at where the "missing samples" info is, then consider the "6 samples are missing in the shorter version".. you have to ignore the "longer" texts under the "missing samples" info, those are separate issues. Same thing with the "repeated samples". EDIT: this is a special case since my WAV's did not end on frame boundaries (burning them made them longer.. to the next frame boundary), otherwise the "longer" text would have appeared in the source WAV sides.

Though.. personally I find it odd that EAC would report silence as repeated samples, shouldn't there be lots more repeated samples in that case? Maybe the 9th track actually has some very very small audio waves right in the beginning, and that is what EAC is reporting, i.e. the start has some kind of very small "hiss"/noise (not "flat" silence). I noticed that in Audacity, very small "waves" of sound are hard to "get visible", even if you zoom in to the max, resizing the track vertically sometimes shows a "wave" (1 pixel flat line(s) over the "baseline"), and sometimes not. I zoomed in to the max and vertically resized the track so much that it took several "jump" scrolls (clicking the "track".. not the "slider") to reach the bottom of the track and still I didn't see any "waves" in the beginning of the track. So, I can't prove my "theory".

Burning audio CD's, comparing software

Reply #7
Perhaps someone can help you with Audacity.  I have no problems with this sort of thing using Audition.
Is 24-bit/192kHz good enough for your lo-fi vinyl, or do you need 32/384?

Burning audio CD's, comparing software

Reply #8
Quote
Perhaps someone can help you with Audacity. I have no problems with this sort of thing using Audition.

Or perhaps I'm teaching you something about Audacity.. I think that this is just a bit special case, you don't bump into this easily. Here's the visual proof:

1st compare picture set:




2nd compare picture set:



Burning audio CD's, comparing software

Reply #9
Consider me an old dog.

I've used Audacity just enough to despise it and the issues you've raised certainly aren't helping.
Is 24-bit/192kHz good enough for your lo-fi vinyl, or do you need 32/384?

Burning audio CD's, comparing software

Reply #10
Oh, I haven't used it much.. despise is such a strong word that there must be much wrong with it then, thanks for the warning. I guess it's commercial software then if one wants to do something serious.. IIRC Audacity is the best (& only?) freeware audio editor.


I failed to create a "repeated samples" situation, I guess I don't understand the issue enough to pull the right ropes. I tried to find proof that EAC does not report silence as "repeated samples". However, I found evidence that speaks that the track in question with repeated samples conundrum does have some sounds at the start, here's the last screenshot I'll post, hopefully it's useful info to somebody:


Burning audio CD's, comparing software

Reply #11
despise is such a strong word that there must be much wrong with it then, thanks for the warning.

Let me apologize for my choice of words.  I've tried a couple of different audio editors in the past before settling on Cool Edit Pro (now Audition).  I've since tried Audacity on more than one occasion because it has been so highly praised here (perhaps it's just OSS fanatics singing quite loudly, IDK), but I've never been able to get it to do what I want and I've completely run out of patience trying.
Is 24-bit/192kHz good enough for your lo-fi vinyl, or do you need 32/384?

Burning audio CD's, comparing software

Reply #12
Ok, thanks for the clarification.

Burning audio CD's, comparing software

Reply #13
IIRC Audacity is the best (& only?) freeware audio editor.

No it isn't, there's Wavosaur for instance if wanting to make usable GUI comparisons to Audacity, plus it's fully portable.

I also wouldn't put too much faith into CD-TEXT, I've seen countless burning software write rubbish CD-TEXT info with strange characters that EAC and CDex for instance can't decipher.

Burning audio CD's, comparing software

Reply #14
Thanks for the tip, I'll try it later.

I actually didn't want any CD-Text, that was the problem, CDBurnerXP wrote CD-Text for the 1st track, and only for that track although it should have written all (though that would have been also incorrect behavior) or none in my case (I didn't enter/edit any CD-Text). All other burning apps I used during my tests didn't have any problems writing (/not writing) CD-Text.

Burning audio CD's, comparing software

Reply #15
Why are we using ImageShack? If they're in such a financial crisis as to use pop-ups, certainly their service must be of such a neglect and poor unusable quality that nobody in their right mind ought to be using it, right?

Anyway I've been looking for a solution to that samples-below-threshold-aren't-even-appearing-at-all problem for a while. It's not apparent in Audacity 1.2.x.
Mixing audio perfectly doesn't take more than onboard.

Burning audio CD's, comparing software

Reply #16
I haven't had any problems with ImageShack.. it's the only one I know/remember when I need to post images.

I tested Audacity 1.2.6 (latest stable) and the problem is not so bad as in the latest beta version. However, vertically resizing the track still hides/unhides some waves. Funny thing is that the latest stable (1.2.6) Audacity is from 2006-11-16. 4 years of beta testing the next version?!

Hopefully this doesn't go too much off-topic.

Burning audio CD's, comparing software

Reply #17
You can try to set 'write offset' in EAC. I use EAC 0.95 beta3 with read offset and write offset configured. There are two ways I tried to get a bit-identical backup CD:

1) a) Insert source CD
b) 'Detect Gaps' -> Select all tracks -> 'Copy Selected Tracks' -> 'Create CUE Sheet'
c) Insert blank CD-R
d) 'Write CD-R' -> 'File-Load CUE Sheet' -> 'Write CD'
e) 'Detect Gaps' -> Select all tracks -> 'Copy Selected Tracks' -> 'Create CUE Sheet' (for comparison)

Resulted 2nd set of files in ideal case (no read/write errors) was bit-identical to the 1st set of files in my experience except for case when 1st track pre-gap is more than 2 seconds (HTOA is present). In that case as I remember I got 1st track pre-gap equal to 2 seconds in created CD-R (.wav files were still bit-identical).

2) a) Insert source CD
b) 'Detect Gaps' -> 'Copy Image & Create CUE Sheet'
c) Insert blank CD-R
d) 'Write CD-R' -> 'File-Load CUE Sheet' -> 'Write CD'
e) 'Detect Gaps' -> 'Copy Image & Create CUE Sheet' (for comparison)

Resulted 2nd set of files in ideal case (no read/write errors) was bit-identical to the 1st set of files always in my experience (including HTOA case).

My drive doesn't support HTOA fully though (it's NEC-4550A and it returns silence instead), so obviously created CD-Rs still lack original sound data of HTOA when it's present.

Burning audio CD's, comparing software

Reply #18
rednyrg721, it seems that you misunderstood what this topic is about, it's not about burning an identical backup of a CD. This topic is about how CDBurnerXP writes audio CD's differently, and IMO wrongly, when using source (WAV) files that do NOT end on CDDA frame boundaries, than rest of the tested burning programs.

Also, the source files that I used do not come from any CD, it is a custom remix/re-master of audio that was sourced from a game.. and I had to personally edit the 9th track to include the in and out fades to mimic how the track actually plays in the original album.

Burning audio CD's, comparing software

Reply #19
Actually I think CDBurnerXP handles it better than the others.  I remember this being discussed a while back but didn't remember which program did what.

Consider a gapless album that is not cut on sector boundaries (eg: digitized vinyl).  Would you prefer your burning application pad the end of each track with null samples to fill the last frame or have it fill the last frame of each track with samples taken from the beginning of the next track?  One retains gapless playback while the other does not.
Is 24-bit/192kHz good enough for your lo-fi vinyl, or do you need 32/384?

Burning audio CD's, comparing software

Reply #20
it's not about burning an identical backup of a CD. This topic is about how CDBurnerXP writes audio CD's differently, and IMO wrongly, when using source (WAV) files that do NOT end on CDDA frame boundaries, than rest of the tested burning programs.

Sorry, I thought you wanted to make a CD which when ripped would be bit-identical to original files. I missed the 'files do NOT end on CDDA frame boundaries' problem. Maybe it would be helpful then to look at some file pair (end of track 1 and start of track 2 for example) in audio editor to find out which program behaves in a best way?

Burning audio CD's, comparing software

Reply #21
Thanks greynol for remembering/figuring it out! Good stuff. Though the "have it fill the last frame of each track with samples taken from the beginning of the next track?" is not correct; CDBurnerXP trims the track(s) to the previous boundary and inserts the trimmed part to the beginning of the next track. This way the compare "missing samples" make sense in the original WAV side (damn that EAC compare window is hard to read ). I tested/verified this with numerous comparisons of the different WAV's.. thanks rednyrg721 for the nudge.

This is a hypothetical case (though possible.. I vaguely remember having at least few tracks that would fit this): a track where the audible sound ends right in the end of the track would produce a (small) noise in the beginning of the next track when burning with CDBurnerXP, if the next track originally opens with silence, then playing the track separately from the "album flow" would impair the listening enjoyment. E.g. the start of the track 8 in my CDBurnerXP compare has 2082 samples from the previous track, which is about 0.05 seconds and audible (tested this with listening few ~0.05 second (audible) ends from this same album.. even low sound fades are audible). And if there are more tracks, let's say 20, transferred samples might double -> ~0.1 seconds in the last tracks. Hypothetical case.

I tried to figure out the EAC compare messages for the 10th track (it's hard when comparing 1 pixel high "waves") and it seems that CDBurnerXP happily threw away the last ~500 samples (= ~0.011 seconds) of the song.. still it is 0.038 seconds longer than the original (like the EAC compare says). I can't make head or tails of the "different samples" messages, for example, the extracted WAV which was burned with CDBurnerXP is 4:59.093 long according to Audacity & Wavosaur, then how come EAC compare says "different samples" in the 4:59.093 - 4:59.097 range of that file?!?! Weird. Maybe the EAC compare is borked, or CDBurnerXP messed the samples.. or both.. I don't know, it doesn't look good.

Hmm, I think I'd rather have EAC et al. style of burning audio than CDBurnerXP. One can always try to fix the boundaries manually if such gapless album situations rise.. or use CDBurnerXP if lazy.

EDIT: Wow, almost everything is solved now! Only the "different samples" messages in the CDBurnerXP compare aren't clear. Excellent, thanks for the input everybody!

Burning audio CD's, comparing software

Reply #22
Thanks for the correction about CDBurnerXP.  Good work!
Is 24-bit/192kHz good enough for your lo-fi vinyl, or do you need 32/384?

Burning audio CD's, comparing software

Reply #23
CDBurnerXP author just forwarded what the 3rd party StarBurn library authors said about this issue: "He's correct here with this post*. We'll take a look at DAO engine after V13 release.". This might implicate that they are not happy about the way the library handles this kind of situation. *Doesn't mean this post/topic here in HA but in CDBurnerXP forum.

 
SimplePortal 1.0.0 RC1 © 2008-2020