Buffer underrun reports in ImgBurn log

[Bosanek]

Hello,

 

does ImgBurn report the occurences of buffer underruns in its log? I am asking this because I have a Plextor PX-755A, and I wrote a MCC 03RG20 (Verbatim DVD-R) using PlexTools 2.35 at 8x. There were no other running processes on the PC (not even the anti-virus). I was not doing anything during the writing process, in fact I was having a breakfast. It wrote the DVD with 12 buffer underruns. And its final writing speed was 6x, probably a speed reducion measure because of those buffer underruns.

 

 

Shocked by that outcome, I wrote another identical DVD, but using ImgBurn, also at 8x. This time I was cleaning the table & the dishes from the breakfast. ImgBurn also "finished" at 6x, but it did not report any buffer underruns.

 

By the way, PowerRec is ON and AutoStrategy is AUTO both in PlexTools and in ImgBurn.

 

----------

 

Now, the question which I am asking is - does ImgBurn report occurences of buffer underruns? If it does, does it also report a LBA sector at which every buffer underrun occured? And a final question - why there were no buffer underruns reported in ImgBurn, while the writing process was performed similiarly (started at 8x, finished at 6x) as in PlexTools which had 12 of them?

 

P.S.: The DVD writer is on a separate ATA channel than the drive from which the source data was read. The content which I was writing was an image file of 4.2 GB (not thousands of small files, which in general case can be the cause of buffer underruns, since the HDD might not be able to "collect" all of them in time).

[blutach]

It does show waiting for buffers to recover. Most drives nowdays have BURP. Do a scan - the discs are probably just fine.

 

In case 2, the drive may have slowed down without losing buffer - just a spike in CPU or IO might have done it, so ImgBurn wouldn't report waiting for buffers to recover.

 

Regards

[Bosanek]

As I mostly write discs for archival purposes, I want top-quality writes, so buffer underrun points are very undesirable, although they are not fatal. That is why I must know of every occurence of a buffer underrun, with an exact LBA where it happened.

 

So, are you saying that the message "Waiting for the buffers to recover" (which I have never seen) in fact represents an ocurrence of a buffer underrun? In other words, is this message a reliable indicator of a buffer underrun?

 

And does it display a LBA at which the buffers were emptied? If not, I highly suggest adding that info, if it is possible for ImgBurn to know that.

[LIGHTNING UK!]

Plextor probably get the info via a special command sent to the drive.

 

It's impossible for a normal burning tool to be 100% accurate because some drives empty the buffer during the burn anyway - like when doing OPC type stuff.

 

If you really need the buffer underrun info, stick to plextools and plextor drives.

[blutach]

Again, do a scan and see how the discs look.

 

Regards

[Bosanek]

OK, thank you both for your suggestions.

 

A side note:

It's just a pity that nowadays drives come with miniscule buffer sizes, so they rely on BURP as a measure of solving buffer underruns. In my opinion BURP is the worst "cure" for buffer underruns, because a BURP point is equivelant to a steel welding point - it is a patch in essence, which can never be as sturdy or as reliable as a solid piece of work.

 

The only good side of BURP is that it completely eliminates the chance of failed writes which are caused by data feeding insufficiency, so that Joe won't get mad and call optical producer's tech support saying that his drive is not working (because it failed to write a disc because of BU). BURP ensures that ordinary Joe probably won't ever know that his drive is patching his PC's inefficiency very often.

 

Large buffer sizes, on the other hand, do not completely eliminate a chance of buffer underruns, but they are a far better solution in general, because with a buffer size which would give you for example "grace time" of 4 seconds, a buffer underrun would be a well deserved one indeed . Such a drive would "endure" system overloads in most cases, and the burns would not have welding points.

 

My Yamaha CRW8824s (a 8x8x24x SCSI CD-R/W writer from 2000) has a 4 MB buffer, giving me over 3 seconds of grace time (4 s / 1.2 MB/s = 3.33 s). I have never had a buffer underrun with it, and it is still working flawlessly on my Pentium Pro. On the other hand, my Plextor PX-755A has a 2 MB buffer, giving me only ~90 miliseconds of grace time on 16x DVD writing speed. Since I write at 8x, that time is 180 ms, but that is still a hair thin and fragile. If the drive had a 32 MB buffer (and AFAIK, cache RAM on optical drives is dirt cheap these days, as any RAM is), that would provide almost 3 seconds of grace time on 8x DVD writing speed, what would be very good.

 

But as I said, it is commercialy more desirable to have BURP instead, because Joe won't notice the problem at all and won't bug the company.

 

 

=====

 

Al this text has nothing to do with the subject or ImgBurn, this was just mine own opinion on the general matter, and excuse me for taking ~3 KB of your storage space for this message .

[blutach]

With systems getting faster, how big a data buffer would you have? It seems 2Mb is good enough for the vast majority of burning.

 

Regards