24/96 - what equipment/tracks/ears do you need?

[jmtd]

I recently put an M-Audio "Audophile" 24/96 capable PCI sound card into my desktop PC. To try it out, I downloaded a 5.1 mix of 1 Ghosts I and 18 Ghosts II, as well as FLACs of the tracks in standard 16/44 for comparison.

For output, I used a Sony Minisystem and a pair of Sennheiser something-or-others closed-back headphones (crap! I've misplaced their precise model number. I'll edit it in later). The cans are pretty good, mid-level consumer ones.

I could not audibly detect a difference between the two bit/sample rates. I tried playback in VLC and foobar2000, both on Windows 7, and I double-checked that the M-Audio control panel was configured to output at 24/96.

I was under the impression that the higher sample/bitrate tracks were supposed to be easily noticeably different, as oppose to e.g. high-bitrate MP3 vs FLAC and such, which is much harder to discern.

Which is more likely: my ears aren't good enough, my output devices aren't good enough, or my machine is misconfigured? Or is Ghosts I-IV not a good listen test?

[botley]

Sorry, you compared 5.1 mixes (fan-made) with the stereo 16bit? The fan mixes are sourced from 16bit multitracks.

[blassster]

There may be another layer to this, you may want to look into stuff about Windows 7's tendency to resample everything to its own setting. Look into something like this for Foobar playback, and make sure the "exclusive" mode is checkmarked in your sound device's advanced settings as shown in the image here.

My post is open to corrections if this isn't the "proper" way, coincidentally read about this stuff myself the other week out of curiosity.

[Andrew]

Try with the 96/24 Slip files vs the standard ones. If possible, plug headphones straight into the card instead of having an external system in the path. If you must have something in between the card and your headphones, make sure to use analog out from the card and not an optical/digital connection - there's a chance that is downsampling to meet the destination.

[seasonsinthesky]

indeed, it seems it was just the lineage of the Ghosts surround tracks that proved the exercise... well, futile. Andrew's suggestion is great — but as an addendum, take a song you know like the back of your hand, a recording you've listened to endlessly. if this applies better to TDS rather than The Slip, it's fairly easy to get the 24/48 stereo mix from the TDS DVD-A (NOT the DVD-V layer!), which would also significantly improve over the remastered CD (in my experience). if it applies to With Teeth better, you can also get the stereo mix in 24/48 from that DVD-A. even the 48kHz sample rate makes a massive difference if you have good monitoring/headphones/bit-perfect (or as close as possible) soundcard in play. (even if you don't, 24/96 piped out to a not-so-good monitoring system capable of doing that sample rate will still sound better than piping 16/44.1!)

[jmtd]

Hi folks, thanks for all of your replies. Sorry Botley, I wasn't listening to a fan 5.1 mix, I've no idea why I put 5.1 in there. it was a rip from the blu rays. The versions I compared are http://archive.org/details/nineinchnails_ghosts_I_IV/ and http://archive.org/details/NineInchN...-ray24bit96khz

The cans are Sennheiser HD 438s.

I'll give the Windows/exclusive-audio tips a go and try to compare some Slip tracks too.

Thanks!

[sa_nick]

I think 438's are pretty low on the scale when it comes to pro audio gear. I mean, I can pick them up for $100 in Australia!

I'm no audio pro though, so I guess they might be good enough to give you some kinda noticeable difference.

[BenAkenobi]

before the listening test make sure you're approaching it without being tired, annoyed or frustrated.

[jmtd]

I think 438's are pretty low on the scale when it comes to pro audio gear

I guess that's why I described them as mid-level consumer ones, and not as pro.

[sa_nick]

Yeah, yeah, I just meant you might notice more of a difference with better cans. Even if the setup is configured properly, I'm saying that could definitely be the reason. I'm sure your ears and Ghosts are both adequate.

And what seasoninthesky said is good advice. Find a song you've heard a billion times and give that a listen. If it's all setup right and the cans are good enough you'll be able to hear so many subtleties you've never heard before.

[jmtd]

OK! I performed some more listening tests, this time throwing Demon Seed into the mix. (I'm pretty intimately familiar with Ghosts and The Slip, in fact perhaps more so than TDS in recent years). I also upgraded my audio drivers and tried Both WASAPI and ASIO output modes. Once the M-Audio panel was configured for 24/96, with WASAPI foobar refused to play 16/44.1 files (wouldn't automatically upsample) so I am pretty confident the settings were all right. Despite that I could detect no discernable difference in my tests. The output chain is all analogue but routes through an LCD monitor and the minisystem before it hits my cans so I can cut some of those out of the mix (not all easily, though, since there's no straight headphone jack on the M-Audio 2496 card.)

May I ask, for those who *most definitely can discern a difference*, what equipment do you use?

[ Edit: not a lot to add, but discussing on IRC revealed the following: "SNR on M-Audio 24/96 is 104dB. Which is about 17.5 bits of resolution." ]