Which software player do you like?

[dave2010]

Perhaps all the software options for playback on a computer are equivalant, but I don't think so. I do think there are differences between iTunes and Real Audio, WMP etc., and more so if mp3 or aac files are concerned.

I'm currently using a Win XP based netbook, which recently I've taken to link directly to my Caiman DAC via USB, and I can play music via Napster, Spotify or from my network drive. I often actually prefer to use my Squeezebox, if only because it can sometimes be easier to manage, and doesn't get messed about if I start net surfing or sending e-mails etc. That said, the sound quality via players on the little PC seems to be rather good (via the exernal DAC), so I wonder what others think of different players.

Currently I'm using Winamp, with the MAD plugin for mp3s, and this is useful for listening to music from the BBC via radio. I have suspected that MAD is a better decoder than some others, though some people have suggested that recent Winamp codecs are just as good. Until recently I would definitely have thought the MAD plugin was better - it just sounded more musical to me.

I've noticed some round here like foobar - but I'm wondering whether there are real differences, and what sort of consensus there is. It's quite possible that different players have a particular forte, so maybe iTunes would be good for aac encoded material, and Winamp for mp3, and maybe foobar for lossless - e.g. WAV files.

Do others have a view on this?

1. Do the players really make a difference?

2. If so, which players are good for what?

[Rare Bird]

Well i've tried a few & Media Monkey with FLAC is the best sounding to me.

[Marc B]

I use foobar , I've never heard any difference in media players once all the dsp stuff is turned off.

[The Vinyl Adventure]

Foobar is best in my view, it works, it's insanely simple, it only does the stuff you ask it to do, get asio drivers sorted and it sounds just fine

[ZebuTheOxen]

Another vote for Foobar.

Turn off Replaygain and turn the Mono to Stereo DSP on (Caused me no end of headaches figuring out why I couldn't play mono files :lol: )

[Themis]

You have some data here : http://www.benchmarkmedia.com/wiki/index.php/Computer_Audio_Playback_-_Setup_Guide

And a comparison between players here : http://www.benchmarkmedia.com/wiki/index.php/Category:Setup_Guides

I can't comment, I don't use computers for audio playback. But there are differences.
At one point someone has to decide and choose between "bit-perfect" and "better sounding" playback. I always prefer bit-perfect, as far as I am concerned.

[slate]

At work I have worked my way through the various players as years has passed.
At the beginning I used WinAmp but at some point it used a lot of CPU and I then shifted to Sonique until it was killed in the combination of a Lycos takeover and the IT bubble.
Eventually Sonique were replaced by my beloved iRiver iMP 100 (RioVolt) CD/MP3/WMA player. Tottenham Court Road can be a dangerous place to visit. It still works here 9 years later.
But for 2 years ago it went into the drawer as... Modern times!!... well ever since I FLAC'ed my CD collection... I started out using WMP (bad library handling).
Then my employer forced Vista on us and for some reason I could not get WMP to play FLAC. Searching for alternatives I tried among others MediaMonkey, VLC, etc before ending up with SqueezeBox Server + SqueezeSlave.

[dave2010]

There is some information about dithering and the MAD codec here - http://ff123.net/dither.html

and

http://www.underbit.com/products/mad/

This doesn't help much with lossless compressed files, but does give a feeling that there can be better ways of handling compressed audio.

My own impression has been that the MAD decoder worked as well or better than any other mp3 decoder.

Another idea to try might be for WAV (or similar) files - to upsample them and see if the results on replay are better. This could either be done off-line i.e with a batch process to generate new files (e.g 24 bit/88.2 kHz or 96kHz) from CD quality (16 bit/44.1 kHz) originals, or could be done dynamically, on the fly during the playback process. Upsampling by a factor of 2 shouldn't introduce too many unwanted effects, though upsampling by 96/44.1 might give some side effects which would need to be controlled. Theoretically there might be no significant audible differences possible, but it'd be interesting to actually do the experiments and see (hear!) if the upsampled fiiles sound better.

Some players are frameworks which can be used with different components (plugins) so it may be possible to develop or use upsampling components to (possibly) give better quality sound.

[Vincent Kars]

Both OSX and Win can do SRC.
Benchmark Media published some measurements using OSX: http://www.thewelltemperedcomputer.com/SW/OSX/OSX.htm
dCS measured various setups on OSX and Win: http://www.thewelltemperedcomputer.com/Lib/OperatingSystemsHandlingOfSampleRates.pdf

Obvious, programming a good SRC is not trivial.
Using it will probably generate audible differences but not favorable ones.
If you do, you might have a look at pro-audio software. They often use a full 64 bit audio path to eliminated all audible artifacts.
http://src.infinitewave.ca/ compares the results of various SRC implementations

[dave2010]

Vincent

Thanks for this.
Both OSX and Win can do SRC.
Benchmark Media published some measurements using OSX: http://www.thewelltemperedcomputer.com/SW/OSX/OSX.htm
dCS measured various setups on OSX and Win: http://www.thewelltemperedcomputer.com/Lib/OperatingSystemsHandlingOfSampleRates.pdf

Obvious, programming a good SRC is not trivial.
Using it will probably generate audible differences but not favorable ones.
If you do, you might have a look at pro-audio software. They often use a full 64 bit audio path to eliminated all audible artifacts.
http://src.infinitewave.ca/ compares the results of various SRC implementationsThese articles are interesting, and do illustrate that there can definitely be problems with sample rate conversion. They also demonstrate quite a variation in the quality of sample rate conversion done by some systems, and a surprise is that WinXP based software does better than Vista. The Apple software generally seems to do quite well, though problems are pointed out with sample rates in iTunes.

It is clear that the problems do arise with conversion between (say) 44.1kHz and 48 kHz, or 44.1kHz and 96 kHz, but I'd expect upsampling from 44.1 to 88.2kHz or from 48 to 96 kHz to be fairly benign. Not only that, but some might claim that they could detect an improvement if a good interpolation filter is used on the upsampled data.

I can't say for sure with audio data, but my feeling on image data is that upsampling with some appropriate filtering produces a very modest improvement when enlarging images. It's hardly ever as good as getting better images in the first place, but can be subjectively slightly better than cruder systems which don't upsample and interpolate "properly". For examples of good upsampling of images, see the work by Orchard and Li - http://neuron2.net/library/nedi.pdf & http://chiranjivi.tripod.com/EDITut.html

Whether the problems noted with some of the audio upsamplers are really severe it's not so easy to say, as though there are clearly very significant measured differences, even in the worst case they may not be audible. However, if audio is passed along a processing chain, the errors may build up and become obvious. For preference it would be better to start with a system which does not exhibit undue additional artifacts.

One final issue is the bit resolution of the output. Even if software produces 24/96 output, generally outputting over USB reduces this back to 16 bit. Using Macs can avoid the problem, as newer ones have SPDIF capability via their headphone mini-jack, but for PCs it may be necessary to use a USB -> SPDIF converter, such as the m2tech hiface, which delivers 24 bit resolution data using USB. It'd be interesting to know whether users of this think they can hear the difference. In the longer term, since there is in principle no reason why 24 or even 32 bit data should not be possible over USB, it would be a good idea if OS developers could raise the audio standards appropriately, even for USB. There should not be a need for additional hardware in order to get higher resolution using a USB interface, as fundamentally most of the problems are in software.

[Vincent Kars]

Currently all major OS support USB audio 1 standard. This is limited to 24/96.
A lot of implementations are limited to 16/48 simply because this can be done using a cheap drop in USB receiver.
Early 2009 the USB audio 2 specs where finalized. This supports 174/196.
As far as I know, there are no native mode audio drivers at the present supporting USB audio 2.
32 bits is not supported by this standard but I don't think there are audio files using more than 24 bits.
24 allows for a dynamic range of 144 so -144 dBFS, well below the noise floor of any gear.

[dave2010]

Currently all major OS support USB audio 1 standard. This is limited to 24/96.
A lot of implementations are limited to 16/48 simply because this can be done using a cheap drop in USB receiver.
Early 2009 the USB audio 2 specs where finalized. This supports 174/196.
As far as I know, there are no native mode audio drivers at the present supporting USB audio 2.
32 bits is not supported by this standard but I don't think there are audio files using more than 24 bits.
24 allows for a dynamic range of 144 so -144 dBFS, well below the noise floor of any gear.Interesting about the current 24/96 USB. Presumably that affects some manufacturers - e.g Stan - so one upgrade he could do is to fit 24/96 USB on the next generation Caiman. It also seems likely that bottom end computers would now have the 16 bit limit not for any particularly good technical reason, but simply to differentiate them from middle and high end kit. I shall look out for a 24 bit audio interface on the next computer I buy. [Does the iMac have 24 bit USB audio?]

24 bits should normally be more than enough for high quality audio - you are right about the noise floor etc. However, I think I could see a use for 32 bit floating point - though maybe it's a bit dubious. It'd still be 24 bit data, but level changes would be easy, could be done digitally. Obviously it should be possible to convert between 24 bit integer and 32 bit floating point without too much difficulty. It might be easier to make a digital mixing desk to work with floating point.

I found the USB audio 2 spec documents, but I couldn't find anything about the rates allowed. I take your word for it that they go up to 196 kHz. There could be some issues about sampling rate conversion - as we noted earlier. The spec does permit SRC to be done in some fairly flexible ways I think, at a quick glance. Most of the time end users should be happy enough with 96kHz or lower rates, but it's always possible that some will detect better quality at higher rates - possibly due to side effects, rather than theory.

[Vincent Kars]

It also seems likely that bottom end computers would now have the 16 bit limit not for any particularly good technical reason, but simply to differentiate them from middle and high end kit. I shall look out for a 24 bit audio interface on the next computer I buy. [Does the iMac have 24 bit USB audio?

If you talk USB audio, all PC/iMac support 24/96 simply because that is the USB 1 audio standard and their native mode USB audio drivers adhers to this standard. Most outboard DACs are not capable to use this simply because of the USB receiver chip they use as mentioned above.
So this is not a limitation of the PC but of the (outboard) DAC
Bit more detail: http://thewelltemperedcomputer.com/KB/USB.html

[dave2010]

vincent

Thanks again.
If you talk USB audio, all PC/iMac support 24/96 simply because that is the USB 1 audio standard and their native mode USB audio drivers adhers to this standard. Most outboard DACs are not capable to use this simply because of the USB receiver chip they use as mentioned above.
So this is not a limitation of the PC but of the (outboard) DAC
Bit more detail: http://thewelltemperedcomputer.com/KB/USB.htmlSo we really do need to talk to our friendly DAC manufacturers about this then. :) Not quite sure what the point of having the audio at 24/96 is if it can't drive an external DAC, and using another computer as an external soundcard doesn't make too much sense either.