kentech

For sharing evaluations of “bit-perfect” player software, I would like to propose a couple of high-resolution sound files than can be downloaded for free. This way we experimenters can be sure we are listening to the same material. This list should be expanded, too: other suggestions are welcome. Here are two files I can highly recommend. The piano recording, in particular, has a midrange clarity and definition that is quite stunning. The overlapping harmonics of this well-tuned piano seem perfectly represented in the file, and your ear/brain can easily follow certain sound-colors buried in the larger tapestry. This fine quality is easily damaged, however, by digital and/or analog imperfections. You can hear the result during comparisons — and this is the point: I believe that this is an example of a source that can be trusted as Really Good. If the piano sounds clouded or furry or granular, it’s NOT the file’s fault! (And it can be played very LOUD without pain.) If the playback is right, you will think someone has dumped your speakers and moved in a piano! The second is a cut from a Chesky release of Columbian vocalist Marta Gomez. Of her several albums, this one has especially fine production values, and the whole album can be downloaded, track by track. But the song “Sophia” is available free. It contains sounds of great delicacy (breath, throat, lips, fingers on strings) that can be muffled by poor processing or falsified if harmonic content is perverted digitally or by parasitic junk in your amp. So the files are: (1) A superbly recorded first movement of Beethoven’s Sonata #32, from Norway. Download it from here, third item from the top (139MB): http://www.2l.no/hires/index.html Check out what else is available, too. The piano is phase-correct as-is. Open the FLAC file in any editor that can then save it as 32-bit floating-point, do the Save As, and label that file appropriately to distinguish it from the original. (2) The song “Lucia” by Marta Gomez can be downloaded from here: https://www.hdtracks.com/index.php?file=login&redirectto=samplealbumdownload&ialbum_id=6446 (Get the Rimsky-Korsakov track, too; the dynamics are stunning!) The whole album of Marta Gomez, Chesky Records No. JD301, Entre Cada Palabra, is absolutely spectacular for d/a and other sound-system evaluation. Other tracks offer more of the same fine quality as “Lucia”: more swoon-worthy songs, plus Spanish-style hand-clapping and multilayered, complex, multi-instrument Big Sounds that are good stressors for sound systems. Track #4, for example. Any single track can be purchased for $2.50 here: https://www.hdtracks.com/index.php?file=catalogdetail&valbum_code=HD090368030122 The Gomez cuts are *inverted* absolute phase, IMO. Open them, invert the phase in the computer, then save as both 24-bit and 32-bit FP for testing. I hope to contribute some commentary on sound-file player software in this forum using these tracks. I hope some of you folks find them useful and enjoyable.

In response to Bob Stern: The digital part of the system: (1) Mac Pro, mid-2008, 8 processor cores, 2.8GHz, 8GB RAM. (2) System 10.7.3 “Lion.” 10.6.8 optional on reboot, for testing. (3) 25’ Firewire-400 cable from either the native FW port or the same on an auxiliary PC card; leading to . . . (4) A PreSonus FireStudio Mobile. With the output level control all the way up, the main outputs are at normal line level or slightly above. (5) 2’ Litz interconnects to analog amplification. The $300 FireStudio Mobile, a versatile “recording interface,” is outstanding for transparency and clarity, with built-in JetPLL 60dB jitter suppression and a special high-efficiency Firewire interface. It tops out at 96kHz and 24 bits, but also syncs at other common sampling frequencies. It will run on FW-bus power, but that’s usually quite dirty; I use a separate 12V power converter that was chosen for extremely low noise, with a manual switch in the 12V power cord. The Firewire cable has its power line cut, to force the use of the clean power. The software driver is un-idiosyncratic; everything just works transparently with all MacOS audio software I’ve tried. I also use the FireStudio’s microphone input with a moving-coil cartridge to digitize LPs. All RIAA EQ and level adjustments are done digitally at 32bit floating-point precision, sometimes 64bit. There are 4-5 “equivalent” ways to perform RIAA — all of which sound different. The performance standard for me at present is the “RIAA” Audio Units plugin from Sound Consultants. Firewire is the high-performance standard for studio recording. The sophisticated error-correction built-in to the usual interface chips make it far more data-reliable than USB, and I use it and eSATA exclusively for external hard drives for my Mac. I’m surprised that Firewire is not more common in “audiophile” d/a converters. With this arrangement, I can easily hear tiny differences in software behavior, crossover capacitors and bypassing, cabling, speaker-wire termination, amp-circuit mods, etc. The music I love is reproduced with the beauty and realism I'm addicted to. I can't help but wonder, given my long experience with this hobby, what another $10,000 invested in a converter would buy me in incremental improvement (if any).

agentsim, I'm grateful for your insights and experimentation with making BitPerfect work in this mysterious "Integer Mode." But . . . I use a Firewire-interfaced 24/96 d/a converter with great pleasure, and I'm appalled (as are others) at how different the various "bit perfect" players sound, from Audirvana and Fidelia, to Decibel and the ordinary sound-editing programs (Audition, Sound Studio, etc.). So far, I'm avoiding the iTunes' interface and its libraries. I wonder what "integer mode" might buy me. Is "Integer Mode" something that is relevant to a Firewire driver, or only to USB-connected devices? I wonder what I can expect. So far, using impeccably recorded 24/96 files from Chesky (I'll list in another post), the best and most authentic quality is coming from humble players, such as Fission, not from the hifalutin "audiophile" players. And, most paradoxical of all, I get uniformly better quality if I open a 24-bit FLAC file and re-save it as 32-bit floating point. (Much bigger file, but it's the same data.) Then play *that* file instead of the original 24 bit one. (I'm absolute-phase-sensitive and have corrected for it, too.) So far, I don't yet believe a word of any "bit perfect" claims because of the elephant in the room: the players sound different and shouldn't! Is the user simply not in control of these buried parameters, no matter the desire?