You should tell them that the 24bit WAVEs are 99% identical to the 16bit files first...


...created with Photoshop CS3 Mac.


Thanks for pointing that out. I had already canceled and deleted the partial 24bit download. Now I'll fetch those two tracks anyway.

Thanks for the updates Alvaro.
I would not be as keen to get these if they were all 16/48 sources, but if all but two are 16/96 and the other two are 24/96, it's well worth it in my book; particularly since it's free.

I wonder if the Ghosts I-IV 24/96 versions released on Blu-ray as part of the deluxe package are true 24/96 files?

This is not the first case of tracks released as "24-bit" being actually 16-bit padded up to 24-bit. A CD+DVD pack of some Depeche Mode album had "24-bit LPCM" tracks on the DVD with exactly same symptoms.

To tell the truth, I don't know it they were lower sample rate before. By default I wouldn't think but I can't say it for sure after we all have seen that many of them were converted to 16 bits and back. What if they have also been resampled? I can't tell it for sure. I hope some more qualified person can

BTW I like the music better than Ghosts I-IV. I'll keep it. Still thinking about this "24/96" version. Compressed with TAK it may not be THAT much waste of space

once you are sure/have proof that the 24bit recording was made from 16bit recordings, you can tip off the admin of theninhotline.net whom operates the most popular nin site, he'll post the info on the front page and everyone will read it including trent. but i'd be careful that there is no mistake made and there is some tangible proof of the claim to show theninhotline admin.

A good question. I have the Blu-ray, but no Blu-ray drive (currently) and no time to really go through and check into it. If anyone in the L.A. area has the capacity and the interest to look into this, I'd be willing to donate the Blu-ray for a bit.

I should hope that they are true 24-bit files as opposed to padded 16-bit files, of course. I paid $75 for the package, and while I'm not disappointed in what I got for that price, padding 16-bit files and advertising them as 24-bit is unacceptable even if there are no audible differences. There's no potential excuse for that given that Pro Tools HD fully supports up to 24-bit/192kHz throughout the pipeline, and it's more than likely that Reznor's recording and mixing at 24/96 or beyond. Surely Tom Baker received real 24/96 bounces when he mastered it, one would think.

For a free release, is naturally makes no difference, and there's certainly nothing to complain about. I'd be happy with 16/22.05

That's probably because the master tapes were 1630s.

Surprisingly the 96/24 version appears to be more compressed and clipped. Something must have gone terribly wrong.

Here is an example - Amplitude statistics and waveform screenshots from Adobe Audition:

44.1 kHz 16-bit, Track: 10 Demon Seed, passage: 3 min 50 s - 4 min 00 s

Left Right
Min Sample Value: -32766 -32766
Max Sample Value: 32765 32765
Peak Amplitude: 0 dB 0 dB
Possibly Clipped: 0 0
DC Offset: -.037 .013
Minimum RMS Power: -8.42 dB -5.46 dB
Maximum RMS Power: -5.15 dB -2.08 dB
Average RMS Power: -6.41 dB -4.15 dB
Total RMS Power: -6.38 dB -4.06 dB
Actual Bit Depth: 16 Bits 16 Bits

Using RMS Window of 10000 ms




96 kHz 24-bit, Track: 09 The Four of Us are Dying, passage: 3 min 50 s - 4 min 00 s

Left Right
Min Sample Value: -32768 -32768
Max Sample Value: 32768 32768
Peak Amplitude: .01 dB 0 dB
Possibly Clipped: 75 439
DC Offset: -.099 .038
Minimum RMS Power: -7.64 dB -4.06 dB
Maximum RMS Power: -3.39 dB -.61 dB
Average RMS Power: -4.83 dB -2.76 dB
Total RMS Power: -4.8 dB -2.7 dB
Actual Bit Depth: 24 Bits 24 Bits

Using RMS Window of 10000 ms




(Both songs are the same. As said before, there appears to be a naming error.)

But wait - it gets better. The 24/96 version of 1,000,000 is 14.5ppm faster than the 16/44 version. Can you say "analog bounce in one master but not the other"?

Looking at those waveforms confirms what my ears have been telling me. Trent really needs to be told about this!

Cheers, Slipstreem.

At first I thought that too, but if you zoom in and compare individual peaks in each track they appear to be the same, at least to my eyes.

The waveform just looks more horrible in 96/24 because of higher sampling rate, I guess.

I agree with CoyoteSmith, we need to get our story straight first.

Here's the information that should be compared, for each track on the album, between the 24/96 wav and 16/44 flac versions:

1. Whether or not the 24/96 version is truncated to 16 bits
2. An estimated count of clipped samples, along with specific examples of clipping
3. ReplayGain and RMS values - but only on the timebase-corrected versions of the 24/96 tracks
4. The estimated timebase difference

#4 is evidence of an additional mastering stage between the two releases that might be a good idea to know about, as it may result in an ABXable difference that could throw people off when comparing the two formats. #1 if found just means that the format quality has degraded to 16/96. #2 and #4 if found is a clear-cut quality degredation relative to the FLACs.

I'll volunteer to do #4, and by extension #3.

Once this is all hammered out I say we dump all the info on the wiki and scream our guts out at TR.




I have found some clear evidence of clipping on specific waveforms on a 24/96 track that does not exist in the 16/44 track. However, besides that, mastering differences appear pretty slight so far, at least from a visual comparison.

... that said, that a 24/96 release was ever made is of course a good thing, and the production quality of the record overall is excellent. The real problem here is the risk that people will use the 24/96 release as a judge of quality for 24/96 as a format, and the differences that appear to exist could cause a plausible ABX success.

Sure, 96 clips slightly which disappears in 44.1 downsampling, but the level appears to be the same.

Note that I don't have the 16/44.1 FLACs currently, so I'm not making any comparisons to that specific release, but...


All but "Lights in the Sky" and "Corona Radiata" are padded 16-bit files. Apparent bit depth is 16 bits per sample. Some tracks seem to have no musically-related content beyond 23kHz, and spectral meters commonly show intermittent narrowband, non-musically-related spikes at around 31.5kHz at up to -65dB. This would be uncharacteristic behavior of any dither algorithm I'm familiar with. Other tracks seem to have musically-related content well beyond 30kHz at high levels (around -70dBFS). It could be noise produced by some of the synths, I suppose (many of Reznor's synths are vintage modular synths like ARPs and Moogs, and noise is an expected byproduct of that deteriorating circuitry, but this is really narrowband stuff here).

"Head Down" has a few instances of three or more clipped samples, but many instances of what looks like a sort of soft clipping, in which peaks appear to exhibit a "cut off" continuous or near-continuous downward- or upward-sloping trend. The waveform is very unusual-looking at these sections.


Clipping appears to be frequent on all tracks except the two listed above. "Demon Seed" (actually "The Four of Us Are Dying"), for instance, has numerous instances of 25+ clipped samples at 0dbFS. Not exactly atypical of a Nine Inch Nails record.

EDIT: Removed unnecessary comment re: Tom Baker

This sounds like an unfortunate mistake (maybe an effect plugin in his sequencer operating at 16 bits per sample only).

And maybe he did not use that plugin on the two tracks that seem to be really 24 bits.

Well Hopefully when Neil Young's "Archive" 1963-1972 comes out this fall from Rhino/Warner with 10 discs (one for each year?) only available on Blu-ray with all files at 24/192 they'll have something else to judge the quality of high resolution PCM as a format with.

If anyone contacts NIN admin they should also mention that including ReplayGain data in the tags would be a really good way to support its benefits.

Hi there! I forgot I had an account here. The problem is being addressed as I type this. Thanks for the heads up

Thank you for making us aware of this issue.

The files provided to the public are the exact files provided by Brian Big Bass Gardener of Bernie Grundman Mastering Studios. As soon as the Mastering Studio has corrected this mistake we will repost the 24/96 version of the album.

I sincerely apologize for this mistake.

Oh wow! Thank you.

While you're here : Is the timebase difference (14ppm) between the two masters something that will be fixed? One would hope that the only difference between the two should be a downsampling, but that seems to imply a fundamental mastering difference between the two versions.

Now that's what I call customer service! Thankyou very much!

Cheers, Slipstreem.

I can only quess who you guys are, but welcome to the forum and thank you for the excellent news.

Please ask the mastering studio to create a considerably less compressed version and thus truly exploit the capabilities of the 24-bit format. It would also be good to keep the maximum peaks well below 0 dBfs because of the possible "intersample overload" problems that may be more severe when the sample rate is 96 kHz.

You could also consider releasing the fixed version in the FLAC format. It would be a lot smaller and make it easier to differentiate the two versions.



I noticed that each track has duration differences. In total the 24/96 version appears to be about 20 s longer.

However, I think the differences are mostly caused by the differing amounts of silence in the beginnings and ends of the tracks. For example, when I removed these silent parts as accurately as I could from the track 9/10 (which I tested earlier) I got almost exactly the same duration, 4 m 54s 95 ms.

Look closer. Take 1,000,000 and remove all the starting silence - get the tracks sample-synchronized. Then take a transient in one format and find it in the other. You'll find that the two transients are about 334/22627357 =14ppm off in time.

You'll find this exact variation in at least tracks 2-5, which are the ones I've looked at so far. For those tracks the 24/96 version always runs about 0.0014% fast.

This will generally only come about if one of the two formats was subjected to an additional analog bounce (digital -> analog -> digital) that the other was not subjected to.

It's ultimately a somewhat trivial issue, but it could generate an audible difference between the two formats that could falsely be attributed to the quality of the formats themselves.

it would be cool if the 24bit version was in flac but idk how they would go about this, just in case anyone is listening can someone give instruction?

Another big thank you from here!



I second on both propositions. And feel a bit spoiled

While that would be neat, I doubt it's going to happen, for a couple reasons:

1. That would be a totally different master, so they'd have to pay the mastering studio again.
2. Anyone comparing the 16-bit and 24-bit versions would get the immediate reaction that the 24-bit version sounds worse, since it would be quieter. (You and I know it's an issue of having to turn the speakers up for better sound, but that's a pretty hard sell. Even to 24-bit connoisseurs who think they're audiophiles.)

Since the multitracks were released, a fan remaster is possible, and might be a cool project. They're only 16/44, though.

We've discussing lossless formats for the 24/96 files. During testing found a lot of issues when converting the files for burning. Ultimately we decided that uncompressed would be better because those who wanted to archive the files could use their preferred format. If anyone can recommend a method of lossless encoding that decodes and converts reliable on mac and pc please let me know.

No, thank you!


Wonderful news. Thank you Alan/Atticus/Trent/imposter


Unless it should happen to be a "pro bono" effort on the part of Gardner, which isn't exactly entirely out of the realm of possibility, I suppose.


I don't see this as being a particular problem.


Maybe at some point in time, higher-resolution multitracks will be released. Now that remix.nin.com is distributing new multitracks via torrents, I don't see any particular reason why it couldn't happen, though some small time investment must be made to create stems, bounce those stems down to a manageable number of tracks (for those of us with pitiful Pro Tools LE rigs) and actually package the multitrack releases.

Should you guys decide any format for the next release, please make it clear in the filenames and/or tags that it's 'another version' of The Slip to avoid confusion. And thanks again, I'm really impressed.

I don't think itunes supports apple lossless 24/96, so that pretty much leaves flac. there should be quite a few mac and pc tools that will handle flac 24/96, what did you try that didn't work?

Actually, iTunes does support apple lossless playback 24/96 within a computer (but not compatible with ipods). However, iTunes will not convert 24/96k files into an uncompressed files without downsampling (nor does it create 24/96 A.L.L. but I'm doing that else where).

Flac conversion is similarly hairy. Playback was usually fine with flac 24/96, the problem arose when trying to convert to other formats for burning. Pro-apps usually worked fine but freeware and shareware had more difficulties. Ultimately monkey's audio, wavpack and Ogg flac all seemed to work better for that resolution. After some trolling around, it seems other who choose to store high-def audio in a lossless compression format find flac to be unreliable. If someone knows of a flac converter/codec that works universally at 24/96 for playback and reconversion (mac and pc), please let me know what it is and I'll see about making those torrents available.

For maximum compatibility, use flac 1.2.0a. More recent FLAC versions do funky stuff with 24bit which is not yet supported by all FLAC decoding software. Ask jcoalson for more details about this.
FLAC 1.2.0a has been very reliable to me, I've worked with lots of DVD-A, True-HD (Like NIN's Beside You in Time HD-DVD), DTS-HD etc.

Regardless, thank you very much for this great release and support!

Just wanted to add another "Thanks!" for being so responsive.

(And for releasing all this free in the first place!)

Almost totally off topic, but i just got floor passes to the Calgary show thanks to the fantastic presale setup. Ningeneer - pass on my thanks to the man! (Or alternately; thanks Trent! You're a gentleman and a scholar. I'll be front and center!)

In truth, it's no huge deal that the 24/96 files are only available in WAV.

If they were FLAC only, then as ningeneer points out people might have problems converting them to a playable format.

If there were both WAV and FLAC versions, then it's a whole other set of files that there could be problems with and that need tracking (and as we've seen, problems are possible in anything!) and it's another choice which might confuse people who don't know a lot about audio.

So while there are benefits to putting up a lossless compressed version, I think they've done the sensible thing by sticking to WAV.

Also I'd like to add my thanks to everyone involved in this for taking the time to discuss the technicalities and putting the effort in to improve things. It's nice to know that the technical side of things is being looked after as well as the creative and distribution aspects.

I think the 16/44.1/FLAC and 24/96/wav format choices may give the wrong impression that wav somehow provides better hires quality.

IMO, using wav is simply waste of bandwidth. FLAC would also provide file tags.

Recarding to the FLAC problems on Mac I think the currect "Max" version should be able to handle conversions to/from 24/96 FLAC without any problems. I don't have a recent Mac nearby, but could someone verify this? http://sbooth.org/Max/

after reading this WAV seems to be the best choice for 96/24, why cant the latest flac support those rates well?

No one was implying that FLAC can't handle it...it was more of an issue with downsampling when converting.
JXL

I bought my copies from the NIN website to support their marketing model. I got a two CD set and downloaded flac files to listen to until the CDs were physically released. If everyone gets everything for free all the time, then this better source of music content will fade away.

If the giveaway of Ghosts I-IV is indeed sanctioned by NIN, then go ahead and grab your free copies. Otherwise do the right thing. The only free copies I found at the time were torrents and I wanted to bypass the RIAA but still support the artists.

This free latest album url seems to be legit and I'm going to grab it for free

NIN isn't my favorite but it's still listenable and worth the effort to aquire.

The audible utility of 96/24 delivery format is questionable at best, unless you factor in filtering concerns and downstream application of DSP. Either way, a better showcase of the format would be an all-digital release, not a transfer of ~40 year old analog tapes, where the 'resolution' is effectively limited by the tape noise.






This can be (and should have been) monitored at all of the digital recording and production stages right through mastering. Intersample overs have been known about for years now, and the tools exist to prevent them.

i paid for the Flac download, however it is perfectly legal to download GHOSTS since it is published under the C.C. non-commercial sharealike liscense. furthermore trent is NOT releasing these new albums under a record company, he is releasing them independently and there is no sign of reznor supporting the RIAA with the required funding to be a member.

since there are issues with the 24/96 version of the slip, i wonder if there are similar issues with the 24/96 version of ghosts? they made ghosts first and then the slip, if they followed ghosts as a guide for the slip, and the slip has issues, what does that mean for ghosts?

I also think that lossless compression of 24/96 material is, in many circumstances, a losing proposition. Assuming that the source is truly 24-bit, then the compression ratio is generally much poorer than with 16-bit sources because the additional lower bits are uncompressable (if they're not all zeros, that is ). Combine that with the possibility that the user will have trouble playing or converting the files, and I can see uncompressed WAV being a logical solution.

However, there may be a better alternative. WavPack lossy used at around 1024 kbps preserves the dynamic range and audio bandwidth of the 24/96 source material while giving the approximate data size of losslessly compressed 16/44! I have been meaning to put together a white paper describing how this works exactly, but in the meantime it is easy to demonstrate by doing an invert paste operation in something like Audition and see that the difference between the original 24/96 source and the WavPack lossy encoded version is very low in level and mostly in the inaudible band (it uses noise shaping somewhat like DSD).

I realize that there will be resistance to any “lossy” format, but the fact is that the conversion from 24/96 to 16/44 is a far more destructive operation than what I am suggesting, but nobody ever calls that “lossy”!

WavPack is not as easy to play or convert as WAV, but it's getting pretty close. On Windows there's the DirectShow filter, the Winamp plugin and native Foobar2000 support. On Linux it's part of the gstreamer framework and so plays out-of-the-box on many Linux distros. It's not quite as straightforward on the Mac, and not having a Mac I haven't tried all these, but in addition to Max mentioned above there is Play and Cog and this program out of Japan.

Just an idea...

If transcoding from such files doesn't add the kind of artifacts one would get from psychoacoustic files, your idea sounds like a fair deal to me (for high-res files, at least, as far as distribution goes). I'm gonna try it with WavPack 4.50.0 beta on my DVD-Audio rips.

It's worth noting that the same advantages also apply to lossyWAV, IIRC, with the additional advantage that it is compatible with FLAC. (albeit not Apple Lossless.) But I'm not sure if it's ready for prime time yet.

I ended up going with utorrent and that 24/96 wav wasn't too bad to snag in under an hour. Bittorrent kept erroring out so I gave up and have quit using it.

Is transcoding to lossless WavPack fairly strightforward? What's the best app? foobar2000?

I can't say I love the album but it's not bad either. The price was right

It should be very simple using foobar, yes. At least I used foobar to convert the album to TAK and could also copy the tags from the FLAC version with it (the WAV files have the tags too in some kind of data chunk, but foobar can't use that).

This is another full length album entitled, "The Slip". As of today, the only way to obtain this album is downloading it from NIN's website or grabbing someone's torrent. NIN has not released information to purchase a physical copy yet but it did imply it will be available for purchase soon.

It will be available on CD in about 2 months. This is what the NIN website says: "for those of you interested in physical products, fear not. we plan to make a version of this release available on CD and vinyl in july. details coming soon."