16/44.1 AIFF vs. 24/96 FLAC from HDTracks. not impressed, am I doing it wrong?

[HiFiInsider]

I recently signed up for HDTracks and downloaded both 16/44.1 & 24/96 HDTracks sampler. There the same music at different resolutions. Here's how I am playing them.

 

16/44.1 AIFF tracks imported to iTunes, no problems.

 

24/96 FLAC tracks reencoded with Max and sent to iTunes as AIFF files.

 

- Mac Mini set to 2 ch/24 bit/96 in Audio/Midi Setup

- opened iTunes AFTER the changed settings.

- listened to both tracks with Mac Mini connect to PS Audio DAC Link III set at 96khz via optical cable. PS DAC to McIntosh C2200 tube pre-amp to Sony MDR-7506 studio headphones.

The only thing I noticed was a bit more "air" with 24/96 tracks, that's it. Do I have it all wrong or is that what you get, more "air"?

 

 

 

[catastrofe]

I believe iTunes was upsampling the 16/44.1 track to 24/96 due to you audio/midi settings. I'm not saying that this should have sounded as good as 24/96 native, but I wanted to make sure knew that you weren't comparing 2 "native" files.

 

Best,

 

Bob

 

[flatmap]

Not sure if it is an issue for you, but I often get errors using Max to decode high res Flac to Aiff. Regular 44.1 Flacs not a problem but 96kHz never work for me.

 

I switched to xAct, and this has worked better for high res decoding.

 

[HiFiInsider]

"I believe iTunes was upsampling the 16/44.1 track to 24/96 due to you audio/midi settings".

 

Are you saying I have to change Audio/Midi settings to 16/44.1 when I play that sample and reopen iTunes? Then close iTunes and set Audio/Midi to 24/96 when I want to play that sample? That's no fun!

 

[markr]

YEP. That is what you need to do. OS X (Core Audio) has many advantages when used for audio, but auto-resolution switching is NOT one of them. Hence the uproar over Amarra.

 

- markr

 

[markr]

... I promise that I'm not following you around here, and I am not the 'truth' police around here either, but you again mention Max as having some conversion issues and you don't note which version that you are seeing issues with. I am using ver. 7.1 and have no issues with the program. I do know that an update after Max ver. 7.1 did have some issues, but that is supposed to be fixed with the current version: 8.1. Including the version of Max (or any program) that you are having issues with would be an important addition to your post and would help to clear any 'fog' that might be caused otherwise. It would be VERY helpful if you would include more specific information.

 

thanks in advance - markr

 

disclaimer: Markr is in no way connected to the Max project. He has even not managed to yet contribute any monies to the project. Shame on me.

 

 

 

 

[flatmap]

Ah, markr, good point.

 

Are you able to decode 24/96 Flac files successfully with version 8.1?

 

I wish I could bring up the version number I was using, but I deleted the program after I figured out what the problem is -- and haven't gone back to check for new versions.

 

[markr]

.. I don't use Max version 8.1. I'm still using version 7.1 and have no issues with converting any resolution up through 192 Khz. ... I'm purposefully slow to upgrade software. Long experience has shown that "If it ain't broke, don't fix it". Anywho, 8.1 is 'supposed' to be just fine. I cannot personally testify to that however.

 

- markr

 

[markr]

see above...

 

[John]

More air might be all you get, but you could try a few more tests:

 

- Firstly validate the source of the files. Can you be sure that the 24/96 file was not just upsampled from 16/44.1 in the first place? More likely, was the recording done at 24/88.2 and then both of your files were resampled from that master? It's hard to be sure you have really good files. Try the Linn and L2 samplers also.

 

- Worth playing with converters. I couldn't get Max to work and I also have now deleted the program. I'll try again. xAct works for me but I can't then add artwork to the aiff files in itunes. Audacity works but you must turn off dithering or it will not be bit-perfect.

 

- Play the flac files directly from a supportive player rather than converting.

 

- If you have a turntable you can get at the difference more easily. Rip the same track at 24/96, 24/44.1 and 16/44.1 I'll bet you have to listen really carefully to hear the difference between the 24/96 file and the 24/44.1 file (or even the 24/96 file being resampled to 24/48 on the fly).

 

- Try listening through speakers. 24/96 vs 16/44.1 will be more about soundstage than tonal accuracy in my experience. You won't get that with phones.

 

Imo the difference really is there but you'll get it only through critical listening. For casual listening a 16/44.1 file is just fine imo though disk space being practically free I'l always get the hires materials if I can. At the end of the day, don't expect night and day. I think you summed it up rather well with "more air".

 

- John.

 

[Rob]

The optical link (3.5mm to Toslink) gets mixed reviews. If you are not using the Wireworld Supernova 6 glass cable or a similar product if there is one, then that might be the problem.

 

Rob

 

[ramagochi]

I did a small test with two files

 

Is a recording of a solo classical guitar in a church. This recording was a surround sound recording downmixing to a stereo The set was a Decca Tree of Earthworks TC30K (LCR) and a ORTF pair at 4m in the rear for the church ambience (LsRs), with two Earthworks SR30 , the ADC was an Apogee and the preamps Apogee as well.

 

The recording was made with Nuendo 4 at 24 bits 96kHz. The microphones have a frequency response of over 30kHz (This is important because a lot of HD recordings uses microphones with a short frequency response)

 

One file was the original mixing, and the other file was resampled to 48kHz, and resampled again to 44,1kHz with the Izotope SRC at max quality, dithered to 16 bit with the Izotope Mbit+ (Shape medium and amount normal) and cutted his wordlength to 16 bits. The resampled file was upsampled to 48kHz, and upsampled again to 96kHz (Izotope SRC), and was saved like 24bit file (The reason was that the DAC no have the need to change the filter, the two files are the same for the audio system)

 

TEST

 

[barrows]

I am impressed that you can hear any difference when running the toslink output from the Mac to the DL-III. I have a (highly modded) DL-III, and the performance via the optical input from a highly jittered source (the Mac) is pretty poor. Unfortunately I think you will need a better computer interface to really hear the performance the DL-III is capable of: something like a Weiss Firewire interface, or Wavelength Audios' forthcoming async USB to SPDIF interface will give much better performance with the DL-III.

As an example: in my system, "Raising Sand" sounds much better via CD transport into my DL-III, than the 24/96 download does via optical from my MacBook into the DL-III.

 

[peter73]

I have both the 16/44.1 and 24/96 version of Raising Sand by Plant & Krauss and was tempted to test the difference between the two. I have up-sampled the 16 bit version to 24/96 and 24/88.2 using iZotrope 64 SRC at best quality and compared using ABX compare on Foobar>M-Audio Transit (in 96kHz mode)>Benchmark DAC1>Sennheiser HD650.

 

The original 24/96 sounded a bit clearer than the up-sampled 24/96 but nearly identical to the up-sampled 24/88.2. At times I though I am hearing a difference, but was never able to achieve consistent result with ABX.

 

Having done the above experiment, I might reconsider paying the premium for hi-def files.

 

[smallp]

Before you make conclusion of hi-def files, please note that that may be performance bottleneck due to PC noise, PC background program interruption, foobar setting and quality of M-Audio Transit. After you optimize your computer audio system, you may be able to hear a big difference.

 

I can hear big difference between 24/96 and 16/44.1 files from Linn Records. I'm using Cplay + ASIO4ALL, Benchmark DAC1 PRE (USB connection). I have optimized the PC so that ASIO buffer can set to minimum 64 samples and it absolutely no click. I even stopped my windows desktop when I play music. I use notebook power by battery which is lower noise than desktop PC. I have also use the TOD2 USB cable to stop noise from PC entering my DAC.

 

So don't give up. Try investigate more and you will have fun getting each performance bottleneck removed and hear what you expected with hi resolution music.

 

[The Computer Audiophile]

Hi Peter - I have both the 44.1 and 96k versions of Raising Sand. This album is definitely not the best album to use when comparing the two sample rates.

 

In general it's impossible to conclude much when doing these comparisons. Consider some of the following issues.

 

01. What is the native sample rate of the recording?

02. Is the Master from analog tape or digital?

03. What A to D converter was used during the recording process?

04. Was hardware or software used to create an up/downsample for the non-native resolutions?

05. What hardware or software was used for the conversion?

06. Who did the conversion?

07. Is the high resolution version a remaster?

08. How does the DAC in use handle various sample rates, asynchronous sample rate converter etc...?

09. Does the DAC have a sweet spot like 88.2 or 44.1?

10. Does the type of music lend itself well to such comparisons?

 

 

 

 

All of these effect the results of listening tests, plus many more issues not mentioned. I'm not sure if you are a Stone Temple Pilots or Metallica fan but both have DVD-A disc that can be ripped to high resolution files. Both high resolution albums are clearly better than the CD versions and it's very obvious from the first few seconds of playback. I often recommend these two albums to people looking to hear a difference between standard CD resolution and high resolution. It's popular music and it sounds great.

 

That said, no conclusions should be drawn, after listening to both, about high resolution as a whole. There are too many issues involved. Some CDs sound better than 24/192 versions and vice versa.

 

[peter73]

Dear Smallp,

 

I am not such a newbie.

 

Kind regards,

Peter

 

[peter73]

Dear Chris,

 

There is a very interesting thread at HydrogenAudio concerning the difference between the 16/44 and 24/96 files sold by Naim on-line. It seems that these are not only different resolution but also different masters. Could this be the case with the Stone Temple Pilots and Metallica's DVD-A? I had similar experience with Doors DVD-A being clearly superior to the CD-A of the same series, but it was clearly s different master (less dynamically compressed).

 

I fully agree with everything you listed as affecting objecting comparison, but it only proves that the difference introduced by resolution alone is not all that important - there are at least 10 more important factors (your list). In fact I have also compared the 16/44 and 24/48 versions of Peter Gabriel's "Scratch My Back" and the conclusion was the same - insignificant to no difference.

 

Kind regards,

Peter

 

PS The Naim story might sound absurd at first, but I recently sent them an e-mail about an issue I had with a file I bought from them and their answer was even more absurd. So, to me at least, crippling the lower resolution files by means of lesser mastering in order to make you pay more for the higher resolution, is quite believable.

 

[debt_collector]

Dear Chris

I am aware that DVD-A disc that can be ripped to high resolution files in Foobar2000. Is this the best way of doing it? My DVD-A is 24bit/96khz so will Foobar extract the audio files at its native resolution? Or will it have to convert and potentially degrade quality?

Thanks

Debt Collector

 

[The Computer Audiophile]

I've never tried to rip DVD-A in Foobar. I use DVDAExplorer http://download3.videohelp.com/DVDAExplorer2008.07.21.zip

 

[sturgl]

Dj, One other thing to think about is that your D-Link automatically upsamples to 96 or 192, depending on the setting you choose. You cannot switch off upsampling on the D-Link. Unlike with many DACs (imho!), the D-Link provides noticeable improvement to standard redbook via upsampling. So on the one hand, your DAC is already making your 16/44.1 files sound "better" than they "should". On the other hand, and as mentioned by others, you're using TOSLINK and a high-jitter "source". It's pretty widely documented that optical is audibly somewhat inferior to coax - particularly where dynamics are concerned. Dynamics, pacing, and that "air" are where 24/96 really shines imho. More tapping of the toe, shaking of the ass, bobbing of the head. Anyway, that lovely DAC of yours just isn't going sound as good as it could using TOSLINK off the Mini. All told, I'm afraid I'm a bit long of wind and short of advice here. Maybe try the Bel Canto USB Link? That's probably like $500 you don't want to spend. Then maybe add Amarra to your iTunes setup? ($300).

 

[firedog]

Just want to comment (again) that I've personally corresponded with HDTracks and they categorically state that they do not "upsample" to get hi-res files.

 

Acc'd to them, anything labelled "hi-res" is either:from a hi-res source (88k or 96K) or an analogue source that has been digitized at hi-res.

 

Again, acc'd to them, any file that they suspect of being "upsampled" will be removed from their site pending review. They have done this to a few files that came under suspicion.

 

[Archimago]

Here's something I wrote about this back in February on AudioAsylum:

http://www.audioasylum.com/cgi/t.mpl?f=pcaudio&m=70002

 

Rapid A-B suggests "air" but I couldn't tell difference once ABX'ed.