About a line 31 in cd-dvd.xml is sugested to use "hdparm" or "cdspeed" for slow down IDE-CDROMs and few lines after is something like there is no solution for SCSI drives. I've no experience with cdspeed but hdparm failing with IDE-CD if you use an IDE-SCSI emulation for cd writer. Best solution (after hard searching) is set a speed=<speed> option to mount command. This work better for me than hdparm even in native ATAPI with my writer, because this drive resets its speed setting with every cd exchange so it need run hdparm many times (boring and slow). Maybe it is a solution for SCSI drives too. I can't prove it myself, have no native SCSI. Next we should point reasonable speed settings because all drives do not support many speeds. For example my cd writer supports about 15 various speeds but my previous toshiba cd-rom could be set to 14x and 40x only, so if I set 24x it turn out as 40x (noisy) but after I set it <=14x , it slowed down. In my experience is 8x speed reasonable minimum and 24x reasonable maximum with noise under fans level. Any ideas? Jiri
On Wed, Feb 16, 2005 at 10:07:06AM +0100, Jiri Heryan wrote:
About a line 31 in cd-dvd.xml is sugested to use "hdparm" or "cdspeed" for slow down IDE-CDROMs and few lines after is something like there is no solution for SCSI drives.
I've no experience with cdspeed but hdparm failing with IDE-CD if you use an IDE-SCSI emulation for cd writer. Best solution (after hard
it works for me with both scsi (real scsi, not emulation) and ide cdrom drives.
searching) is set a speed=<speed> option to mount command. This work better for me than hdparm even in native ATAPI with my writer, because
this is good for when you're mounting the cd to play, but fails when you want to use dvd:// or vcd:// without mounting. it would be nice if mplayer could also set the speed when using dvd/vcd.
this drive resets its speed setting with every cd exchange so it need run hdparm many times (boring and slow).
imnsho this is an idiotic kernel bug. the kernel should remember your settings...
Next we should point reasonable speed settings because all drives do not support many speeds. For example my cd writer supports about 15 various speeds but my previous toshiba cd-rom could be set to 14x and 40x only, so if I set 24x it turn out as 40x (noisy) but after I set it <=14x , it slowed down.
In my experience is 8x speed reasonable minimum and 24x reasonable maximum with noise under fans level.
Any ideas?
yeah, set it barely above the speed you need to read a movie. keep in mind 1x is 1.2 megabit/sec, so 8x is probably enough for any sane movie file, and 12x is certainly safe. lower values would be ok with large -cache sizes. rich
On Wednesday 16 February 2005 21:07, D Richard Felker III wrote:
this drive resets its speed setting with every cd exchange so it need run hdparm many times (boring and slow).
imnsho this is an idiotic kernel bug. the kernel should remember your settings...
Are the settings remembered with -k?
On Wed, Feb 16, 2005 at 10:01:02PM +0200, Jan Knutar wrote:
On Wednesday 16 February 2005 21:07, D Richard Felker III wrote:
this drive resets its speed setting with every cd exchange so it need run hdparm many times (boring and slow).
imnsho this is an idiotic kernel bug. the kernel should remember your settings...
Are the settings remembered with -k?
nope. another kernel bug: even if you use -k, dma will get disabled whenever there are io errors, even though these errors are, of course, caused by scratched media, not by anything related to dma. it's very evil. whoever put this broken crap in the kernel should be run over with an 18-wheeler full of cola. rich
On Thursday 17 February 2005 02:57, D Richard Felker III wrote:
nope. another kernel bug: even if you use -k, dma will get disabled whenever there are io errors, even though these errors are, of course, caused by scratched media, not by anything related to dma. it's very evil. whoever put this broken crap in the kernel should be run over with an 18-wheeler full of cola.
Atleast it's somewhat better now than it used to be. Someone calculated that before, by the amount of times Linux retried a bad sector read, it would've taken a few hundred years to read a badly scratched CD. It was retrying each broken sector a few hundred times, and the drive firmware was trying it for 10 seconds itself too before giving up.... But that's supposedly fixed now ...
it works for me with both scsi (real scsi, not emulation) and ide cdrom drives.
I was no use for hdparm for about a year, so it probably get improved. My version of it ends with no support for SCSI (or soething like it). If it works for you wit SCSI, maybe we should remove lines with no solution for SCSI drives.
searching) is set a speed=<speed> option to mount command. This work better for me than hdparm even in native ATAPI with my writer, because
this is good for when you're mounting the cd to play, but fails when you want to use dvd:// or vcd:// without mounting. it would be nice if mplayer could also set the speed when using dvd/vcd.
Agree. I've no DVD and last I had a VCD years ago. I thing we should mention it as an option for such use, at least until mplayer can't set speed itself.
In my experience is 8x speed reasonable minimum and 24x reasonable maximum with noise under fans level.
Any ideas?
yeah, set it barely above the speed you need to read a movie. keep in mind 1x is 1.2 megabit/sec, so 8x is probably enough for any sane movie file, and 12x is certainly safe. lower values would be ok with large -cache sizes.
Most codecs defaults to 1500kbits for movie and 128 kbits for audio, so 1x is not enough at all, plus you must compensate random slowdowns which can appear, so 8x is minimum for me. 24x is maximum you should set when you use it for common work too. It is reasonably fast and has a minimum noise. Jiri
On Thu, Feb 17, 2005 at 08:25:02AM +0100, Jiri Heryan wrote:
yeah, set it barely above the speed you need to read a movie. keep in mind 1x is 1.2 megabit/sec, so 8x is probably enough for any sane movie file, and 12x is certainly safe. lower values would be ok with large -cache sizes.
Most codecs defaults to 1500kbits for movie and 128 kbits for audio, so 1x is not enough at all, plus you must compensate random slowdowns which can appear, so 8x is minimum for me. 24x is maximum you should set when you use it for common work too. It is reasonably fast and has a minimum noise.
i didn't say 1x was enough, i just gave it as a base rate. in fact your 1500 is insanely low; a proper 2pass mpeg4 file will peak at 8 mbps or higher with most content! so you really do need 8x, for these rare peaks. there's little point in going over 8x, though. rich
Looking through old mails.. On Wed, Feb 16, 2005 at 02:07:58PM -0500, D Richard Felker III wrote:
On Wed, Feb 16, 2005 at 10:07:06AM +0100, Jiri Heryan wrote:
About a line 31 in cd-dvd.xml is sugested to use "hdparm" or "cdspeed" for slow down IDE-CDROMs and few lines after is something like there is no solution for SCSI drives.
I've no experience with cdspeed but hdparm failing with IDE-CD if you use an IDE-SCSI emulation for cd writer. Best solution (after hard
it works for me with both scsi (real scsi, not emulation) and ide cdrom drives.
Which tool works for you? I have a Plextor CD-ROM and a Yamaha CD-RW, none of the tools I tried are able to set the speed on them... Diego
On Tue, Jun 07, 2005 at 12:56:22PM +0200, Diego Biurrun wrote:
Looking through old mails..
On Wed, Feb 16, 2005 at 02:07:58PM -0500, D Richard Felker III wrote:
On Wed, Feb 16, 2005 at 10:07:06AM +0100, Jiri Heryan wrote:
About a line 31 in cd-dvd.xml is sugested to use "hdparm" or "cdspeed" for slow down IDE-CDROMs and few lines after is something like there is no solution for SCSI drives.
I've no experience with cdspeed but hdparm failing with IDE-CD if you use an IDE-SCSI emulation for cd writer. Best solution (after hard
it works for me with both scsi (real scsi, not emulation) and ide cdrom drives.
Which tool works for you? I have a Plextor CD-ROM and a Yamaha CD-RW, none of the tools I tried are able to set the speed on them...
Rich: ping.. Diego
On Thu, Jun 30, 2005 at 11:00:03AM +0200, Diego Biurrun wrote:
On Tue, Jun 07, 2005 at 12:56:22PM +0200, Diego Biurrun wrote:
Looking through old mails..
On Wed, Feb 16, 2005 at 02:07:58PM -0500, D Richard Felker III wrote:
On Wed, Feb 16, 2005 at 10:07:06AM +0100, Jiri Heryan wrote:
About a line 31 in cd-dvd.xml is sugested to use "hdparm" or "cdspeed" for slow down IDE-CDROMs and few lines after is something like there is no solution for SCSI drives.
I've no experience with cdspeed but hdparm failing with IDE-CD if you use an IDE-SCSI emulation for cd writer. Best solution (after hard
it works for me with both scsi (real scsi, not emulation) and ide cdrom drives.
Which tool works for you? I have a Plextor CD-ROM and a Yamaha CD-RW, none of the tools I tried are able to set the speed on them...
Rich: ping..
hdparm and a program I found called cdctl will both set the speed. However if you're using scsi emulation, you have to set the mode on the ide device node, not the scsi one, even tho the ide device node does not work for reading the drive. :) At least this is how I think it works... I stopped setting the speed a while ago because I was having problems with video getting choppy after the cache dried up from read errors. I'll try to figure it out better if you still have trouble.. Rich
On Thu, Jun 30, 2005 at 09:55:27PM -0400, Rich Felker wrote:
On Thu, Jun 30, 2005 at 11:00:03AM +0200, Diego Biurrun wrote:
On Tue, Jun 07, 2005 at 12:56:22PM +0200, Diego Biurrun wrote:
Which tool works for you? I have a Plextor CD-ROM and a Yamaha CD-RW, none of the tools I tried are able to set the speed on them...
Rich: ping..
hdparm and a program I found called cdctl will both set the speed. However if you're using scsi emulation, you have to set the mode on the ide device node, not the scsi one, even tho the ide device node does not work for reading the drive. :)
At least this is how I think it works... I stopped setting the speed a while ago because I was having problems with video getting choppy after the cache dried up from read errors. I'll try to figure it out better if you still have trouble..
Nothing works for me, oh well. Anyway, I've updated the docs slightly. Diego
participants (5)
-
D Richard Felker III -
Diego Biurrun -
Jan Knutar -
Jiri Heryan -
Rich Felker