Long playing times on CD

[Beef Oven!]

This morning I took delivery of Bruckner 3, Original 1873 version, Altomonte Orchester St. Florian, Rémy Ballot. Label: Gramola.

Single CD.

I knew the performance was Celi-like and expected it to be very long. It is 89 minutes, 6 seconds (btb, I also took deliver of Venzago’s B8 (1890 version Ed. Novak, 1955) and that’s 75.26).

I have a Shostakovich 10 on Collins, LSO with Maxim Shostakovich which is very long, but over 89 minutes on one CD!!??

[Eine Alpensinfonie]

That's quite amazing. I thought the Decca Phase 4 CD of Tchaik 6 and Bizet Carmen excerpts was unbeatable at 84:12 but it was seem that this one beats it by 5 minutes!

Presumably the pit spiral spacing cannot be varied, so I wonder what the theoretical limit is. Clearly not the 60 minute limit of the earliest CDs, or even the 74 minute maximum envisaged bit by the team that invented the medium. 80 minutes was then stated as the absolute maximum, yet still they got longer. But 89+? How do they do it?

[Dave2002]

Is your assumption on pit/spiral spacing actually correct? I have another assumption which is that the coding onto the spiral is essentially PCM with some digital markers/overhead, but maybe the CD spec includes some options for digital compression - though I very much doubt that because the techniques required were not available when CD standards were drawn up. This is a bit of a mystery - so maybe someone who's studied the specs/standards in depth will comment.

[pastoralguy]

Or the 39 minutes for the Pogorelich, Abbado, LSO Tchaikovsky first piano concerto on DG...

I wonder if modern CD players are better able to cope with these long discs. I think 83 minutes in the longest I have, (Janine Jansen and friends playing the Schubert Quintet and Schoenberg's 'Verklarte Nacht'. A wonderful disc!)

I might even buy the Bruckner just to see how my various machines copes with it! Of course, it does raise the question that if ONE discl can provide 90 minutes of music then why can't they all...?

[pastoralguy]

Is the Bruckner 3 any good, Beefy?

[Eine Alpensinfonie]

Is your assumption on pit/spiral spacing actually correct? I have another assumption which is that the coding onto the spiral is essentially PCM with some digital markers/overhead, but maybe the CD spec includes some options for digital compression - though I very much doubt that because the techniques required were not available when CD standards were drawn up. This is a bit of a mystery - so maybe someone who's studied the specs/standards in depth will comment.

It appears that my assumption is open to question, if Wiki is to be believed:

The program area is 86.05 cm2, and the length of the recordable spiral is (86.05 cm2 / 1.6 µm) = 5.38 km. With a scanning speed of 1.2 m/s, the playing time is 74 minutes, or 650 MiB of data on a CD-ROM. A disc with data packed slightly more densely is tolerated by most players (though some old ones fail). Using a linear velocity of 1.2 m/s and a narrower track pitch of 1.5 µm increases the playing time to 80 minutes, and data capacity to 700 MiB.

[ferneyhoughgeliebte]

maybe the CD spec includes some options for digital compression

Specsavers?

[Beef Oven!]

Is the Bruckner 3 any good, Beefy?

It is very good indeed. Sensational, even. In my opinion, it is superior to my beloved Georg Tintner recording. I think the Amazon reviews really explain it well..

I’ve only played it through once, but I am sure it’s going to become one of my most treasured Bruckner recordings.

[Jonathan]

Having read the Amazon reviews and based on what has been said here, I may well have to buy this to add to my other recordings of this work!

Edit: Just found it on Spotify....

[Gordon]

It appears that my assumption is open to question, if Wiki is to be believed:

The program area is 86.05 cm2, and the length of the recordable spiral is (86.05 cm2 / 1.6 µm) = 5.38 km. With a scanning speed of 1.2 m/s, the playing time is 74 minutes, or 650 MiB of data on a CD-ROM. A disc with data packed slightly more densely is tolerated by most players (though some old ones fail). Using a linear velocity of 1.2 m/s and a narrower track pitch of 1.5 µm increases the playing time to 80 minutes, and data capacity to 700 MiB.

The extension to 80 mins is done by changing nothing in the red book but using up the tolerances in the various parameters to eke out more time eg tightening the spiral slightly to lengthen the playing time. Older [and even not so old] players may baulk at plying such extended discs.

Getting another 9 minutes would seem to exhaust the tolerances and no compression can be used because it isn't specified and so the players wouldn't know what to do with a strange bit stream. So how they do it is a mystery to me at least. 9 minutes more is a lot, over 10%.

I suppose that because the disc rotational speed varies [about 2:1 from about 200 to 400 rpm across the disc] to keep the linear speed constant that the electronics/mechanics might also be tolerant of a range of linear speeds. That means that if the cutting machine runs slow by 10% then the pits are closer together along the spiral effectively adding 10% to its length which leads to extended time on playback at normal speed which is "slow" relative to the normal cutting speed. It challenges the injection moulding process a bit but perhasp nowadays this is OK? May be an explanation?

[Beef Oven!]

The extension to 80 mins is done by changing nothing in the red book but using up the tolerances in the various parameters to eke out more time eg tightening the spiral slightly to lengthen the playing time. Older [and even not so old] players may baulk at plying such extended discs.

Getting another 9 minutes would seem to exhaust the tolerances and no compression can be used because it isn't specified and so the players wouldn't know what to do with a strange bit stream. So how they do it is a mystery to me at least. 9 minutes more is a lot, over 10%.

I suppose that because the disc rotational speed varies [about 2:1 from about 200 to 400 rpm across the disc] to keep the linear speed constant that the electronics/mechanics might also be tolerant of a range of linear speeds. That means that if the cutting machine runs fast by 10% then the pits are closer together along the spiral effectively adding 10% to its length which leads to extended time on playback at normal speed which is "slow" relative to the cutting speed. It challenges the injection moulding process a bit but perhasp nowadays this is OK? May be an explanation?

Brilliant post

I understood 14% of it

[Dave2002]

I suppose that because the disc rotational speed varies [about 2:1 from about 200 to 400 rpm across the disc] to keep the linear speed constant that the electronics/mechanics might also be tolerant of a range of linear speeds.

I heard that some CD players (particularly battery powered ones) run at much higher speeds, and buffer the data. This suggests that the readout speed is a possible variable. How is the data encoded on the discs? I think it's simply 0s and 1s - pits/no pits, but really not sure. What about the clock rate? Could the physical size of each bit be reduced, which would require a slower read out - but only if no/little buffering is used? How is the clocking of the bit stream synchronised? Perhaps if it's all buffered anyway, the only requirement is to buffer up the bit patterns, since the actual replay will be done by sampling at (one of) the standard sampling frequencies i.e. 44.1 kHz for a normal CD. Smaller bits along the spiral track would allow more of them, and a longer playback time. Just a suggestion.

[Gordon]

I heard that some CD players (particularly battery powered ones) run at much higher speeds, and buffer the data. This suggests that the readout speed is a possible variable. How is the data encoded on the discs? I think it's simply 0s and 1s - pits/no pits, but really not sure. What about the clock rate? Could the physical size of each bit be reduced, which would require a slower read out - but only if no/little buffering is used? How is the clocking of the bit stream synchronised? Perhaps if it's all buffered anyway, the only requirement is to buffer up the bit patterns, since the actual replay will be done by sampling at (one of) the standard sampling frequencies i.e. 44.1 kHz for a normal CD. Smaller bits along the spiral track would allow more of them, and a longer playback time. Just a suggestion.

Portable CD players do read the disc differently as you describe. The bits are read and stored for several seconds and so any missing bits can be corrected by re-reading the disc whne the FEC fails to deal with them. The DAC MUST play at 44.1 and from a stable jitter-free clock. The input clock is nominally the same on average as the output clock and is derived from the stream off the disc which will be jittered and so the player must store the bits and then read them out under control of a good clock. Depending on the amount of input jitter a short memory can do the necessary buffering [remember that the FEC system needs lots of memory too] usually using the First In First Out [FIFO] method.

Yes shorter pits will mean more of them in a given length of spiral which is what I suggested above. The "bit" on the disc is derived from the audio bits. For each audio half byte [8 bits] there are 17 bit slots on the spiral where this bit period is a minimum length pit. The "bit" rate on the spiral is about 4.32 MBits/sec but the stereo PCM audio is only 44,100 x 2 x 16 bits/sec or 1.411200 Mbit/sec the increase being caused by FEC and other system overheads. Each block of 192 audio bits becomes 588 time slots/bits on the spiral so the spiral "bit" rate is 588/192 x 1.411200 "bits"/sec

In the stream there is a 75Hz timing pattern which the player uses to get the rate correctly set for reading data and also is used to count down to 1 second pulses to give the playing time readout. The player will spin the disc so that its speed is synchronised with the locally generated 75Hz derived from the local stable DAC clock. Because the "new" 75 Hz markers on the "slow" disc are closer together than normal the disc will run slow on playback, matching the slow cut, thus lengthening the programme content. Some players may object of course if they are built rigidly to the red book.

When ripping a disc the linear speed is quite high but the data integrity is maintained so a small amount of variation like + or - 10% shouldn't be a problem.

[pastoralguy]

[QUOTE=Beef Oven!;560574]This morning I took delivery of Bruckner 3


I've ordered this cd out of sheer curiosity!

Let's see how it performs on my various machines. :smiley

[Beef Oven!]

I've ordered this cd out of sheer curiosity!

Let's see how it performs on my various machines. :smiley

There are better reasons!