Simply the task was to take existing cd and import the .wav to ardour for massaging.
Managed to create cd by:
a) setting location markers and cd checkboxes
b) fixing pregap by deleting 2s
c) exporting session
d) fixing .toc file START errors with emacs
e) burning playable cd-rom with cdrdao
The recording âsoundsâ great when played with ardour via soundcard but the burned cd-rom sounds âdistortedâ with all my players.
What to fiddle?
1.) dithering - used none and rectangular without effect I can hear. And is this needed when the session is 42/16 from the beginning to end?
2.) mixer window - does fiddling with the âslideâ fix this problem?
3.) edit window - gain automation - is this the way to go?
Or some other trick? I remember that once I have seen some âwhite-paperâ explaining how to re-record the tracks by routing via jack to gamix and then back. Is this the way?
Now I have made numerous trials with creating cd and the âdistortionâ persist still. What makes me curious is that when played via ardour or alsaplayer the session or the exported session sound fine. After they have been burned to cd the âdistortionâ occurs.
0-finding:
It is wise to have cd-rw and play it immediately e.g. with consumer audio
after (ac97 card) instead of testing with real cd-roms.
Also related finding is that not all cd-roms âget burnedâ from 50 pcs spindle
cdrecord blank=minimal -dev=ATAPI:1,1,0 ## to blank the cdrw
cdrdao disk-info --device ATA:1,1,0
2nd finding: âdistortionâ exists with cdrdao burned cdâs
2a) session export with and without dithering modes for original (copy from cd)
2b) session export Jamin -3,5 dropped input level for the whole âdiskâ.
4th finding:
Ardour and jack would be wise to closed when burning because of the xruns.
5th finding: smaller range exports burned with cdrecord doesnât distort
cdrecord dev=ATAPI:1,1,0 -v -pad -dao q*.wav
But they are significantly more silent than the cdrdao recorded ones and I think they are even too silent. The wavâs were as follows:
q1 - minus 1 â both limit and output
q21 - minus 21
qr - rectangular
qs - shaped
qt -triangular
Because of the success in 5th I decided to range export all the songs from the original âhotâ track piece by piece and then cdrecord as in 5th. And
6th finding: large session exported with ranges and burned with cdrecord distorts
So I am now recording again via ardour-jamin-ardour and by dropping the input about -1,5 with Jaminâs input (all songs are about equal level) and then I try both cdrdao and cdrecord methods.
Oh I forgot to mention that I havenât fiddled with:
a) big endian and little endia setups in the export
Arenât the cdâs actually motorola like?
b) envy24controls setups, but why should I?
c) change my cd-rom and cdrw brand - the distortion occurs in the files anyway
Uh ⌠how does this exported wav sound play on your soundcard? Tried to open the exported wav on rezound in order to see clipping? The session is 44.1 from the beggining, isnât it? ⌠Have you checked if you have latest libresample?
A-oops ⌠Is there resampling for the session/range in export if I have envy24control / qjacktl set to 44.1k / 16 and I export the file to 44.1k / 16
Is it some internal thing? If so should I consider dithering - what each dithering does?
B-oops ⌠I have the âlatestâ rezound but I donât know how to see the clipping. Or is it so that when the clip checkbox is active but there are no clips then there are no clips. btw. why to use rezound - could this feature be in ardour?
C-oops ⌠libresample, where to look?
The export wav-files both range and session exports sound ok with alsaplayer -ojack
but when I burn them to cd-rom or cdrw the file is somehow distorted.
I can also verify by extracting the wavs with
cdda2wav dev=ATAPI:1,1,0 -vall -B -Owav
and then listening the wavs with alsaplayer and the distortion is now in the file extracted from the newly burned cd.
Again numerous trials and I have narrowed so far the only way to have not distorted cd.
record the original (cd-copy) track with jamin -1,5dB to new track and have TAP limiter on 0dB for the master output post plugins.
range export all range-marked tracks simply by names T01.wav - T19.wav (in my case)
Note that in exports are named with .wav (could this be somehow the default behaviour?). Note also T01-T19 not T1-T19 because *.wav order might change with the intended one
I tried many ways with cdrecord and cdrdao and I got mostly distorted frisbees. The 4) was the only way to get it to work so far.
Now new problem is introduced the default pregap of 2s is now inserted between each track. As I write this I try to use the cdrecordâs defpregap=0 (TEAC drive option which might go in the future) which should place 2s pregap only for the first track. I let you know what happened for my trial to make âdaoâ via âTAOâ
The problem with 2s pregap between each track (song) is that break between tracks is now extended with 2s and with danger of having âecho tailâ of the previous track in the new track. Actually I like the idea to be able to control this pregap issue, but I personally think it should be in ardour rather than expecting user to do some strange magic for the exported master tracks on command-line.
(actually it was very handy to have old sony discman player which flashes -2 sec in beginning of each track as indication of pregap)
I have started to suspect the following:
cdrecord
cdrdao
kernel?
lightscribe gigabyte dvd-writer
also the imation spindle has apparently cd-râs which doesnât pass the power calib or similar step right in the beginning
Also I noted that when burning wav-tracks (range exported ones) the -swab option has no meaning, no diff between files burned with -swab and without it. Maybe because -audio toggles -swab. btw. -swab and -audio (apparently) changes the little endian to big endian which is preferrable for audio-cdâs ?!
btw. I found this in ardour 0.99.3 a little bit confusing - should I be able to fiddle with it.
PREGAP resistance is futile ⌠iow. Apparently on tao-mode the trial to get rid of 2s pregap between tracks with defpregap=0 doesnât work on my gigabyte drive (maybe some ancient teac could do it according to man-page).
I am happy to announce that I managed to cook cd for myself.
couple of advice:
a) keep .wav in the export name so there will be 2 files:
file.wav
file.wav.toc
The burning software can act on the extension.
b) check the settings per export and check the levels with sndfile-info
with master outs
with post plugin
[root@piano FULL]# sndfile-info FULL*.wav | grep Max
Signal Max : 2081 (-23.94 dB)
Signal Max : 20133 (-4.23 dB)
Signal Max : 0 (-inf dB)
From the above choices the winner was -4.23 dB
c) specify speed for cdrecord or cdrdao - I used:
cdrdao write --speed 2 --device ATA:1,1,0 FULL3masterin0TAP0.wav.toc
cdrecord -v speed=2 dev=ATA:1,1,0 -audio -pad *.wav
d) maybe its wise to close ardour and other tools before burning
The best result which wasnât distorted was cooked:
1.) ripped the live-recorded cd (canât recall what I used for that, shame on me).
2.) edited it with ardours edit
3.) placed range markers for pleasant places
4.) recorded the edited track by routing the outputs to Jamin and dropping input with -1,5dB and Jamins outputs to new track recording in ardour.
5.) placed plugin TAP limiter with limits in 0dB to master outs.
6.) exported several ranges (much faster than session) and burned tests, listened
fiddled with tons cdrecord options
7.) and finally exported session with best âsoundâ setup experienced 4.-6. steps.
check master output levels if monitoring with them they affect on export
8.) fix errors in .toc and burn with cdrdao
9.) testing of the fresh cd on different players
⌠mmm ⌠still curious about the problem you had. The fact that the export wavs sounded ok with alsaplay confirms that distortion didnât came from the ardour session nor the limiter use. Sometimes Iâve had problems with disks burned too fast, regular players do not readâem or at least do not readâem well. The option â-speed 2â in cdrdao doesnât always work, you should try ââspeed 1â when you have problems. As the default behaviour depends on the cd writer driver in use, you should try to use always something like cdrdao write --speed 1 --buffer-under-run-protection 1 --write-speed-control 1 mydisc.toc
in order to be sure that it uses all needed options.
Which cdrdao driver were you using? which cdrdao version?? which kernel, distro,ecc?
To get rid of 2s gaps anyway you must use SAO or DAO mode, which means usually cdrdao :^)
Thanx for very good advice -
AGREE I
you came to same conclusion that it is in the burning phase where things get broken. Now when I read this âsagaâ I pinpoint myself maybe one problem in the burning commands.
AGREE II
2s pregaps canât be avoided with TAO. Gotta use DAO or SAO.
I have used in the earlier trials device ATAPI:1,1,0 instead of device ATA:1,1,0
It takes ages to read the man-pages and to learn the hardware inside-out the hard-way especially when you are tired. Happy to use linux because of the logic in âabstractionsâ in windoze I would have been giving up already. But I will explain to myself someday their difference.
I figured out that as root life is easier with these commandline tools
Hmmh⌠I feel that I gotta mail you to ask if this gentoo makes my life easier
I gotta try these drivers as well. The cd is now 78:53 long and plays well in most of the players, two anomalies so far:
one dvd player refuses to play it
one cd player starts playing the disc again after end
So think this is due to:
cd-r media + burn quality issue
78 min long disk is oversize
Thanx for clarifying this speed 1 and speed 2 - I was wondering why those arenât actually *1 and *2 speeds.
And joke of the day: one of the numerous âdistortedâ cd-roms had very funny feature with my sony discman portable player. The player started to spin the cd-r but after a while the cd started to spin to OPPOSITE direction
As time of writing I am struggling with some minor amendments on the âfinalâ cd version and then I will jump to the âlightscribeâ world. I definetely would like to have neat lightscribed logos on top of the cd-surfaces, esp. now due linux support.
Iâm using kernel 2.6.17-beyond3 (kolivas + gentto patches, quite a nice kernel if you can try it) and cdrdao version 1.2.1. I guess you should try with the newer version but changing driver might worth a try. Try âdriver generic-mmc and --driver generic-mmc-raw
Besides, a normal user wonât be able to schedule the burning task in realtime mode (unless youâve done all the realtime module issue) so you end up burning disks in low priority which is no good at all for quality mastering ⌠Have you tried cdrdao as root?
As for speed, when I burn disks for myself (have a decent yamaha cd player) I go at max speed, when I burn disks for other people instead, I use to burn them with âspeed 2 or even with âspeed 1. These donât mean speeds 1x or 2x but âfirst and second speeds availableâ, in my case something like 4x and 8x.
Oh, Iâm using the ATA syntax for the devices!
⌠if need any more help, or perhaps doubts at 4am feel free to reach me at heuristic.records [#AT#] gmail.com
Canât resist to tell - that I managed to do very cute cd-label with lightscribe. I used the lacieâs 4L-gui to burn gimp-edited jpg.
So pure linux approach:
recorded with ardour
edited with ardour
mastered with ardour-jamin
burned with cdrdao
labeled with lightscribe in linux
and of course:
asked help with firefox+linux from this forum between the steps.
After creating several lightscribed cd-râs I detected that old pioneer PD-M40 player refused to play those.With more recent players they work.
After reading about yamahas audiomaster q.r. options for cdrecord (max.68min cd with it) I have started to think about the quality of the burned cd-râs - are they handicapped when compared to normal cdâs because of the âerror correction done behind the curtains by the cd-playerâ. Some claim that they can hear the difference of having the 'audiomaster option onâburned cd-râs.
So what to learn from this thread ? Donât blame ardour for âdistortionâ on your fresh cd-râs