Help - Search - Members - Calendar
Full Version: Drive caches audio data
Hydrogenaudio Forums > Hydrogenaudio Forum > General Audio
burnett_s
Drive caches audio data !!:

QUOTE
EAC reads every audio sector twice and compares them. If the two sectors are different EAC knows that a read error occurred and will reread the suspicious sector over and over again until a satisfying result is achieved (or an error reported to the user). If a drive caches audio data EAC will read the audio data from the cache on the second read attempt instead of from the CD. Of course the buffer always gives the same sector to EAC so the error detection no longer works. If the first sector is read incorrect EAC will not detect this as the buffer will give EAC the faulty audio sector twice. Since no difference between the two sectors is detected EAC will assume that the sector is read correct what is obviously not true. In this case the "Drive caches audio data" option has to be enabled , so that EAC will clear the cache by overreading it.     EAC Tutorial by SatCP


I think there's something wrong with this explanation that confuse me a lot.
For me, the explanation should be... the "Drive caches audio data" option has to be disabled , so that EAC will clear the cache by overreading it.

So, what should I do ? sad.gif

Thank you
Greetings
ty1er
im 99% sure that it should be disabled like you stated.

that i i have my EAC setup.

-ty1er
Doctor
What's wrong with saying:

Dear EAC,

This drive caches audio data.
Please clear the cache each time you read.

Sincerely,
burnett_s

?
ancl
By enabling "Drive caches audio data" you inform eac that your drive is caching. Because EAC now knows that, it can use a different (but slower) way of extracting the data to avoid the cache being used (ie make sure the cache is overwritten before reading same data again).

Conclusion: The correct way is to enable "Drive caches audio data", like SatCP said.
ty1er
QUOTE
By enabling "Drive caches audio data" you inform eac that your drive is caching. Because EAC now knows that, it can use a different (but slower) way of extracting the data to avoid the cache being used (ie make sure the cache is overwritten before reading same data again).

Conclusion: The correct way is to enable "Drive caches audio data", like SatCP said.



i just ran the EAC "Dectect Read Features" and it said
Caching: No

So i guess that i do not have to check "Drive caches audio data"... obvoiusly....
i was confusing myself..... when i first read this post.


?? Do most new CD/DVD roms come with this caching feature? or is it more of an older CD/rom feature?
AstralStorm
Some cache audio data, other don't.
It doesn't depend on the generation of the CD drive.
E.g. I still have 4x NEC which doesn't cache audio data,
but my quite new LiteOn LTR-4825S burner does this.
My now broken Hitachi GD-2500 didn't cache.

If your drive caches audio, then to have a secure extraction you should check the box.
If your drive doesn't, then leaving it checked will only halve extraction speed.
Doctor
On a tangent:

Why not read in bigger chunks? This would rewrite the cache yet not hurt speed, it seems.
lucpes
QUOTE(Doctor @ Jun 26 2003 - 12:26 AM)
On a tangent:

Why not read in bigger chunks? This would rewrite the cache yet not hurt speed, it seems.

The smallest cache I've seen is 128KBit. The number of samples read once is smaller than 100, thus less bits overall than the cache's size.
AstralStorm
Most drivers read samples in blocks of 24-32 sectors...

You'd need to read many blocks (like ~64 for 4MB cache) once,
store them, read again, compare. When correcting errors just use older cache bypassing
method without buffering. An additional config option would be needed - size of read buffer.
(You'd have to know it or set it to high value)
Unfortunately EAC's sourcecode isn't open.
Andavari
QUOTE(ty1er @ Jun 25 2003 - 05:00 PM)
QUOTE
By enabling "Drive caches audio data" you inform eac that your drive is caching. Because EAC now knows that, it can use a different (but slower) way of extracting the data to avoid the cache being used (ie make sure the cache is overwritten before reading same data again).

Conclusion: The correct way is to enable "Drive caches audio data", like SatCP said.



i just ran the EAC "Dectect Read Features" and it said
Caching: No

So i guess that i do not have to check "Drive caches audio data"... obvoiusly....
i was confusing myself..... when i first read this post.

EAC during detection set my Memorex drive as Caching: No, but when I read the features on the box was written on it that it did cache audio. Now when manually detecting the drive features EAC then stated that the drive did cache audio.

What's more troubling to me is that this damn caching drive of mine is fooling EAC, by that I mean errors are going completely undetected as I hear problems in the ripped wav's, even with CRC's matching. Yet another reason I'm glad my old drive lasted long enough to archive my audio error-free.
AstralStorm
EAC's first-time feature 'detection' uses database of known drives' features.
It might as well be wrong.

Just turn the switch 'Drive caches audio data' on.
Pio2001
You might also have a different firmware. Different firmwares are known to sometimes change the caching behaviour (there is a famous Plextor example, that switched from no cache to cache).

For safety, run a drive test with Feurio, it sometimes detects audio cache while EAC doesn't.

It is probable that EAC actually reads bigger chunks of data in order to "clear the cache". Anyway this is going to slow down error correction at least.
AstralStorm
It doesn't have to... It might just as well use the smallest block size
while correcting errors.

I don't think that EAC uses a method of bypassing cache similar to the one proposed by me.
It seems to read a tested block, then a random block, then tested block again and then compares the result.
/EDIT\ Added missing "to". \EDIT/

But only EAC developer can confirm that...
burnett_s
Thank you for all the answers !!!
rohangc
For perfect rips, always enable this option-whether the drive actually caches data or not is irrelevant. If you were paranoid about quality of rips (like me), you should enable this option at all cost.
This is a "lo-fi" version of our main content. To view the full version with more information, formatting and images, please click here.
Invision Power Board © 2001-2008 Invision Power Services, Inc.