I'm concerned that the drive might output jitter

I'm afraid you don't understand what jitter is.

Almost all albums for the last track have no audio in the part of the audio [...]

• Compact Disc Digital Audio#Data structure
• Disc ID Calculation#Reading the TOC
• Optical disc RAW format
• Spoon's Audio Guide: CD Ripping

• Actual end of audio data
• Then 24 samples of lead-out
• Then 6 samples of pure silence ("0x00" bytes) added by EAC

Almost all albums for the last track have no audio in the part of the audio (last half a second, way more than 30 samples)

And I thought I ...

Quote from: krafty on 2024-07-11 10:13:02

And I thought I ...

I do know that any of this is not relevant to audio quality but i guess its better to be careful when you are starting out.

However, a drive may be able to rip faster than another or may be able to extract better data from damaged CDs using C2 info for instance.

Concerning that the original poster is regarding the new +6 offset with the drive, against the +667 offset drive, is it correct to conclude that a +48 offset drive is also "worse" for that matter?

And... analyzing Spoon's offset graphic example here, can one say that the -2 offset drive is as bad as a +2 drive, meaning that -2 will read totally the lead-in part but miss the last samples in the lead-out?

Can anyone validate this graph I did?

[attach type=image]31937[/attach]

Right, but Spoon's graph is similar to this. He did it with the +6 and to me, that example shows that the drive can reach further lead-out and miss the lead-in, and the -2 example shows the drive could read before audio. But I will try to redo this graph in the "reversed" form. The goal of this graph is to visualize the 30 samples issue, so one can predict this by looking.

• A "perfect" drive would be a "-30" one (i.e. +30 offset correction in EAC) (i.e. actually a "0" drive, because of the "30 samples AccurateRip issue"), because EAC would also cut the 30 samples at the beginning (we can't avoid that, it's what AccurateRip asks for). And at the end, EAC would stop reading exactly at the end of audio data (i.e. before the lead-out part), then add 30 samples of silence.
• A -40 drive (i.e. +40 offset correction in EAC) would also cut 30 samples at the beginning (the drive being actually a -10 one, and we offset it by +40, resulting in +30 start position). And at the end, EAC would stop reading 10 samples before the end of audio data, replacing them with silence, then add the remaining 30 samples of silence as the offset correction is +40. Such a drive would lose 10 samples of audio data at the end.

(function () {

    // change this value:
    const eacOffsetSetting = +6;

    const erroneousOffset = -eacOffsetSetting; // -6
    const realOffset = erroneousOffset + 30; // +24



    let cutFromStart = realOffset; // 24
    let readFromLeadIn = -realOffset; // -24

    if (eacOffsetSetting > 0) {
        cutFromStart += eacOffsetSetting; // 24 + 6 = 30
        readFromLeadIn -= eacOffsetSetting;  // -24 - 6 = -30
    }

    if (cutFromStart < 0) { // never occurs
        cutFromStart = 0;
    }

    if (readFromLeadIn < 0) { // always occurs
        readFromLeadIn = 0; // -30 → 0
    }

    const eacSilentSamplesAtStart = (eacOffsetSetting < 0) ? -eacOffsetSetting : 0; // 0



    let cutFromEnd = -realOffset; // -24
    let readFromLeadOut = realOffset; // 24

    if (eacOffsetSetting < 0) {
        cutFromEnd -= eacOffsetSetting;
        readFromLeadOut += eacOffsetSetting;
    }

    if (cutFromEnd < 0) { // occurs when eacOffsetSetting < +30
        cutFromEnd = 0; // -24 → 0
    }

    if (readFromLeadOut < 0) { // occurs when eacOffsetSetting > +30
        readFromLeadOut = 0;
    }

    const eacSilentSamplesAtEnd = (eacOffsetSetting > 0) ? eacOffsetSetting : 0; // 6



    const report = `
        samples cut from start: ${cutFromStart}
        samples read from lead-in area: ${readFromLeadIn}
        silent samples added by EAC at start: ${eacSilentSamplesAtStart}

        samples cut from end: ${cutFromEnd}
        samples read from lead-out area: ${readFromLeadOut}
        silent samples added by EAC at end: ${eacSilentSamplesAtEnd}
    `;

    console.log(report);

})();

        samples cut from start: 30
        samples read from lead-in area: 0
        silent samples added by EAC at start: 0

        samples cut from end: 637
        samples read from lead-out area: 0
        silent samples added by EAC at end: 667

        samples cut from start: 30
        samples read from lead-in area: 0
        silent samples added by EAC at start: 0

        samples cut from end: 0
        samples read from lead-out area: 24
        silent samples added by EAC at end: 6

        samples cut from start: 32
        samples read from lead-in area: 0
        silent samples added by EAC at start: 2

        samples cut from end: 0
        samples read from lead-out area: 30
        silent samples added by EAC at end: 0

• In all cases, at least 30 samples are cut from start.
• The lead-in area is never read from.
• At most 30 samples may be read from lead-out area.

• The "worst" drives are those with EAC offset < 0, as they cut more than 30 samples from start (precisely, 30 + abs(EAC offset)).
• The "best" drives are those with EAC offset between 0 and 30 (inclusive): the former reads 30 samples from lead-out area, the latter doesn't read from lead-out area and adds 30 silent samples
• Drives with EAC offset > 30 cut samples from end (precisely, EAC offset - 30).

• EAC takes what the drive gives him. That audio is affected by the "real offset" of the drive.
• Then EAC shifts that audio, using the "erroneous offset" value, by cutting audio at one side and adding silence at the other side.