Jump to content
IGNORED

AyreWave- A New OSX Audio Player Released AT RMAF


Lars

Recommended Posts

I'd suggest trying the AppleScript posted earlier for loading tracks into AyreWave from iTunes. It's slower than the built-in AyreWave import function, but uses a different approach to loading files that *might* work in this case.

 

........................

 

AyreWave requires a 64 bit processor (like the Intel Core 2 processors) for some of its operations. That should not mean any changes in device drivers or anything of the sort. You do not need to run a 64 bit OS X kernel, either.

 

As far as whether the data to the device is sent in integer format or floating point, I believe that is entirely dependent on what the device demands. That all should be transparent to the user. The only user Preference for this is whether you want to obtain exclusive access for your DAC or not.

 

Personally, I wouldn't get too wound up in the minutiae of the process. Try the application and see if it works for you and you like it. If it works and you like it, great! If it doesn't work, ask about the problem. If you don't like it, then you've lost some time, but perhaps learned something that might be valuable to you in some way.

 

Link to comment

My LP recording hardware capabilities are wav and dsd and just using Max to convert is not going to solve the issue with those files.

 

Additionally, a lot of the finest quality hi rez downloads are wav files, i.e iTrax and the 24/192 stuff Chesky sells on DVD's.

 

Seems to me that part of minimum functionality for a hi rez music play back program would be to accommodate these issues, either through importing the file tags or enabling edit within AyreWave the same way you can edit the tags in iTunes.

 

I also "think" is sounds a tad better than PM, though the difference is only of substance on the best source material and I think I still prefer PM's upsampling on many Red Book CD's.

 

Therefore, while sound is most important, imo AyreWave is going to have to become a little more feature rich to have an adequate market, especially with Mr. Robinson supposedly working on an Integer Mode version.

 

Link to comment

The problem with wav files and iTunes is that iTunes stores the tags and all in its own data base, since wav files does not embed metadata. At least that is my understanding.

 

There is a way to work with this in iTunes to solve the issue, pretty much painlessly, but if Pure Music is your choice then the point is moot.

 

Link to comment

about memory playback leveling the playing field of format choices, I decided to run a simple blind test last night. I took five different familiar demo cuts and removed their metadata and loded them into AW, in all three formats I would use....AIFF, FLAC and WAV. I had no idea which was which, just sorted by duration so all of the same song sorted together. In each example it was easy for me to pick out (and prefer) the cleanliness and overall higher perceived resolution of the wav files. Go figure.

 

Anyway, like some others who prefer wav, I have a majority of my Mac server converted to this format, and use Silverlight's somewhat painstaking-but-now-have-it-down-to-a-quick-process of using iTunify and coverartserver to get metadata and cover art into itunes for my wav stuff. However, as mentioned, when used outside of iTunes it is metadataless (new word). So any tips on how to take this metadata and store it with the actual song file rather than with itunes library file is helpful

 

CG, I may ping you about your script. I'm a relatively new MAC user. I understand terminal commands, but have never installed or run a script.

 

Link to comment

 

 

"From the CoreAudio HAL doc, it seems a user level program (that is a standard application, not a kernel extension) can only output PCM data in 32bit floating point (the virtual format)."

 

Yes, this means that so-called integer mode requires communicating directly with the output device drivers (rather than through CoreAudio).

 

"But it is the physical format that counts for the output stream to the sound card/DAC.

The trick is that the 32bit float is still composed of 24 "integer" bits (mantissa + sign bit, in addition to the exponent), so this enables bit perfect output for 24bit streams."

 

IMO, the purported advantages (if any) of integer mode are believed to be in the elimination of the unnecessary conversions (of integer data) to float. As CG says above, no matter if you're using CoreAudio, or speaking directly to the device drivers, the software can only deliver the data (to the DAC) in the format that the driver expects/accepts.

 

Perhaps I've misunderstood your point?

 

If so, please clarify,

thanks much

clay

 

 

Link to comment

You can save existing files in the iTunes library as aiff files if you choose to do so. (NOTE: I'm away from my Mac right, so I have not tried this specific use, but have with other file types.)

 

Rather than going through the explanation of what to do here, please look at

 

http://www.brighthub.com/computing/mac-platform/articles/65168.aspx

 

This article is complete with pictures. Of course, instead of setting iTunes to convert to mp3 format, you'd want aiff.

 

If you so choose, you could make aiff copies of every single wav file you have with a few clicks and waiting while iTunes does the job. This should embed your manually entered metadata into the new file, but I haven't tried that. I've always done the conversion first and added the metadata after. Note that you can do the conversion in batches by selecting multiple tracks or albums to convert.

 

I presume (oops!) that you already have wav file back-ups of all those albums you carefully made into digital files, so deleting the wav files from your iTunes library would not be a hardship.

 

wav files generally are the accepted lossless file format for the Windows world, because it is the native format supported by Microsoft. There's a description of the two formats in the FAQ for this web site

 

http://www.computeraudiophile.com/content/Whats-difference-between-WAV-AIFF-AIF-MP3-FLAC-WMA-ALAC-M4A-etc

 

Hope this helps. If not, an AppleScript could do the job automatically. There's some over a Doug's AppleScripts that might meet your needs. I do something like this to resample CD tracks to 88.2 KHz with a specific filter set-up.

 

Link to comment

"CG, I may ping you about your script. I'm a relatively new MAC user. I understand terminal commands, but have never installed or run a script."

 

The script will run just like a regular application. When you've selected the tracks in iTunes you want to play with AyreWave, just double click on the Play with AyreWave icon, or single click if you have it located in your Dock. No terminal anything needed; that would be no fun!

 

But, as I just wrote, the metadata may not transfer well. That's a problem.

 

BTW, I've tried the same experiments with wav and aiff files, and could determine no difference at all. Interesting.

 

Link to comment

I was astounded by the improvement in soundstage provided by AyreWave.

 

This program seems to be extremely accurate in reproduction during even the most complex of passages. Instead of the periodic smearing in iTunes, and consequent fatigue, everything is pellucid.

 

The combination of a firewire DAC to reduce jitter and AyreWave on the software side is very sweet indeed.

 

This program is a brilliant achievement.

 

Link to comment

 

ted_b: "..about memory playback leveling the playing field of format choices, I decided to run a simple blind test last night. I took five different familiar demo cuts and removed their metadata and loded them into AW, in all three formats I would use....AIFF, FLAC and WAV. I had no idea which was which, just sorted by duration so all of the same song sorted together. In each example it was easy for me to pick out (and prefer) the cleanliness and overall higher perceived resolution of the wav files. Go figure.."

 

 

 

A possible explanation for this is how AyreWave might read files from disk into memory. One way would be to read the file from disk and convert it to a common target format that the Mac OS X audio sub system can understand, and copy that to the computer's RAM. Another way might be to create a 'memory mapped file' of what was on disk, in the computer's RAM as an exact copy of what was on disk, and then convert that to the target format expected by Mac OS X at runtime while the track was playing.

 

 

System (i): Stack Audio Link > Denafrips Iris 12th/Ares 12th-1; Gyrodec/SME V/Hana SL/EAT E-Glo Petit/Magnum Dynalab FT101A) > PrimaLuna Evo 100 amp > Klipsch RP-600M/REL T5x subs

System (ii): Allo USB Signature > Bel Canto uLink+AQVOX psu > Chord Hugo > APPJ EL34 > Tandy LX5/REL Tzero v3 subs

System (iii) KEF LS50W/KEF R400b subs

System (iv) Technics 1210GR > Leak 230 > Tannoy Cheviot

Link to comment

Ted, after you sorted the WAV, FLAC & AIFF tracks, did you inspect the sort order to see whether the order of the 3 files was random? If WAV always appeared in the same position (1st, 2nd or 3rd), that could explain your preference.

 

HQPlayer (on 3.8 GHz 8-core i7 iMac 2020) > NAA (on 2012 Mac Mini i7) > RME ADI-2 v2 > Benchmark AHB-2 > Thiel 3.7

Link to comment

five song groupings). And no, my ears are really not that amazing or anything, but I've become a pretty good trained ear I guess..in fact, I have slight tinnutis that affects 8k+ in my left ear; some days worse than others, some days almost nonexistant. But the results weren't that hard to hear...more of a veil lifted...more air yet a slightly more gentle top end..as if the gain was reduced by a few tenths db and resolution increased (less shouting, more clarity). SPL meter says no, though.

 

Link to comment

The HAL is the lowest level API you can normally use on OSX from an application that runs as a normal user.

For a reason Apple has decided that this HAL takes only 32bit float samples (the so-called "virtual format"), and it is up to the driver to internally convert to the DAC format (integer).

Fortunately, the float precision is of 24bits, preserving the bit perfect precision.

But yes, this means the CPU needs to perform this conversion (multiplication by 2^15 or 2^23, and copy from float to int register) to get a signed integer value.

 

The other setting exposed by the HAL, called Physical format, is to tell the driver which format it will convert the 32bit float to. It is roughly setting the DAC mode (e.g. 16/24bit).

 

Most applications use the higher level Audio Units API (that also enables to use the effects plugins), but this means going through much more processing of the audio signal. More CPU load, external busses load, thus more power supply load, and possible electromagnetic interferences... that have an effect on the sound.

 

From my understanding (but maybe I'm wrong), AW uses this direct HAL API instead of the Audio Units used by nearly all the others.

 

Damien

 

MBP 15"/Mac Mini, Audirvana Plus, Audioquest Diamond USB, AMR DP-777, exD DSD DAC (for DSD), Pioneer N-70AE, Audioquest Niagara balanced/Viard Audio Design Silver HD, Accuphase E-560, Cabasse Sumatra MT420

Link to comment

 

Damien,

 

so you're saying that there is NO way for an OS X program, even using Physical format, to avoid the initial conversion of the bits (as read from an AIFF file) into Float-32?

 

You seem to be saying that this option will reduce some processing, but not all of the conversions to Float?

 

clay

 

 

 

 

 

Link to comment

From the CoreAudio HAL doc, it seems a user level program (that is a standard application, not a kernel extension) can only output PCM data in 32bit floating point (the virtual format).

But it is the physical format that counts for the output stream to the sound card/DAC.

The trick is that the 32bit float is still composed of 24 "integer" bits (mantissa + sign bit, in addition to the exponent), so this enables bit perfect output for 24bit streams.

By setting the physical format, you can tell you card/DAC to switch to 16/24 bit mode (if it supports it).

 

What you've posted above is mostly correct. A normal application that does not use hog mode must perform IO as 32-bit floating point. This is to allow mixing of audio from different applications before output is sent to the device. However, when hog mode is used the application gains exclusive access to the device, and can perform IO transactions in any virtual format supported by the device. This is how AyreWave is able to send, for example, 24-bit signed integer output directly to an audio device when hog mode is enabled.

 

Link to comment

So as not to be (possibly) grouped with “nearly all the others,” :-) :-) just for the record, Pure Music / Pure Vinyl do not use the high level AudioUnit APIs. In agreement with Damien’s statement, we have found the AudioUnit programming model to offer reduced performance (though it is a lot simpler to use then our much lower level, “roll your own” approach). And certainly, the AudioUnit API cannot be used with integer format data; the AudioUnit specified format is floating point.

 

(AudioUnit processing plug-ins in Pure Music or Pure Vinyl are implemented via a separate, “push” side chain mechanism that isn’t active at all, unless plug-ins are specifically chosen to be loaded by the user.)

 

(And let me also state that integer to float and back again is computationally very inexpensive, and does not affect audio quality... Direct integer support is, however, on our road map for release this quarter, to support true 32 bit DACs like the ESS SABRE.)

 

Rob Robinson

 

 

Link to comment

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now



×
×
  • Create New...