Bit-perfect: Really that important?

[Alec]

Whether i get a mac or not (or get something like the Juli@ in a PC) pretty much comes down to the whole bit perfect thing. Am i really going to hear that much difference? Or is it one of those things that one might just hear in the right circumstances?

And this stuff about using your amps volume control. Much of the time im right at the pc, so would use the WMP control. Would i really be decreasing the sound quality that much? If i get the Juli@ in a PC, should i still be need, ideally, to keep the volume at 100% in wmp or itunes or whatever?

Would i hear the kind of difference, say, as i did when i addeda beresford to my previous setup (MA B2s and NAD C320 BEE)?

 

[Anonymous]

Although I think that it is sometimes exaggerated, there are differences. But there are different aspects to this, depending on what your system changes compared to the original source:

1. Resampling, eg from 44.1kHz to something else: this depends on the quality of the resampling algorithm implemented (as computer software or hardcoded in the soundcard) and the sample rate created Obviously, downsampling is to be avoided. If one can hear the difference between 44.1k upsampled by a good algorithm to 48k I doubt, but there is no good reason to do so. Upsampling to 96k or 192k is better left to a DAC I think,

2. Digital volume/balance control means reducing the range from 16bit=64k possible values (assuming CD quality) per sample to something lower. This introduces more "rounding errors" (google/wiki quantization noise), and a more jagged primary signal in the DA conversion that is more difficult to process/filter by the DAC and the analog amplification may pick up the increased noise/reduced faithfulness, Then again, if you reduce the volume digitally this usually means that you play at low volumes so you may not notice this.

3. Changes to the digital sound using sound effects, equalizer settings and what not. I suspect that when a certain configuration sounds really bad it is likely that this is happening, for instance by the settings of a program like iTunes. Then again, if you are happy with an equalizer setting that creates a better sound you should be a happy camper. Of course it is less Audiophile but who cares..

Often (also here) Windows kmixer is blamed for altering a signal, but I have found no hard evidence that this is necessarily the case if you pay attention to the settings (at 100& volume, no other sounds mixed, no digital processing etc).

Vista & 7's WASAPI is interesting as it allows exclusive mode, so no interference from system sounds or irritating webpages that start playing something. Note that in W7 (Vista?) you can select the sample rate/depth of the mixed signal if needed (I think it automatically selects the highest rates of the signals to be mixed) by clicking the properties.

 

[Alec]

Thanks Pete.

I have just been reading elsewhere about bit perfect ou put from the SB Duet. Aparently, it is bit perfect as long as you dont go below 70 with the digital volume (every 10 is .5 decibels aparently). I wonder if this is a comfortable level to listen at (i dont really know what people mean wehn they talk about high or low volumes of course, its subjective; im told my music is loud but it iosnt to me when theres only me here).

My reading also involved something about matching your amp and your digital volume somehow, but i didnt understand it and my brain was hurting. didnt understand whether they were saying that would allow you to use you digital volume or not.

Does any of this mean anything to you? Also, im a little unclear of if im going to hear a difference...

 

[Anonymous]

I doubt that a reduction to 70 is still bit perfect. More something like difficult to hear a difference in SQ rather than in volume.

In principle I would suggest 100% digital volume and hence low amp volume. I imagine that what you read is for an amp that does bot perform very well at low volume settings, so that a lowered digital volume let you pump up the amp's volume. This is a strange workaround, if this problem exists one should probably buy a neutral attenuator (Goldenjacks appear to be doing this - never tested it) or a better amp.

 

[Alec]

it was something to do with there being enough bits that til a certain point they bits in the music hadnt been interrupted. this is from one of the top logitech bods, tho at one point his maths was a bit shakey even to me.

its too complex for me as it is in the thread, and i appreciate if you dont want to wade in, but here it is.

 

[PJPro]

However clever the algorithm doing the upscaling, I would suggest that it is something to avoid.

There is a clear and rational explanation for SQ improvements around bit perfect playback. Not only that, in most instances, with a bit of fiddling and know how, it's free. So why wouldn't you want to do it?

As far as I am aware, the only blocker to bit perfect playback is iTunes on Windows. As long as you avoid that combination, it should be possible to provide bit-perfect playback.

 

[Alec]

There might be rational arguments but can you hear it?

I dont want to avoid it as such, i just think its a faff and im too used to using the digital volume control.

my latest idea (well ive come full circle) is teh squeezebox, which, as i say it seems is bit perfect to a point, i just wonder if its at reasonable listening levels.

ive been driving myself (and probably others, sorry) a bit mad of late and im close to saying "hang it all, ill do whats most convenient as long as it doesnt decrease my sound quality.

as ive said (somewhere), i found the beresford an improvement, and certain players, but i just wonder if bit perction is the same.

sorry im typing in a hurry and waffling ill shut it now.

 

[Craig M.]

al, you could quite easily check if you can hear any degradation yourself. just make sure you are currently getting bit perfect playback, and then compare using your speakers remote to the media player volume control. you say you find media player volume adjustment more convenient, but it's not exactly difficult to lift a remote and point it at your speakers.

 

[Alec]

Craig M.:al, you could quite easily check if you can hear any degradation yourself. just make sure you are currently getting bit perfect playback, and then compare using your speakers remote to the media player volume control. you say you find media player volume adjustment more convenient, but it's not exactly difficult to lift a remote and point it at your speakers.

Oh, why you cheeky swine!

But i spose you might have a point. I'll try asio again. tried it once but just got silence. think i must have done something wrong.

EDIT - Hang on. if i use asio it disables WMPs volume doesnt it?

 

[Craig M.]

i think so, oh god! - that means using the remote....

 

[Alec]

Yes, oh sarcy one, but it also means not being able to test for an audible difference.

 

[Anonymous]

Kinda grasped what you've got going on and I guess you want FLAC with no digital upbringings, I'd especially avoid the Beresfords variable system which isn't very good.

 

[Alec]

I just want to know if going bit perfect is worth it, really, simple as that. Obviously it would be with lossless files, tho things are in dissarray just now, but the format doesnt really matter. I'd go with WMA in windows and ALAC in OSX i think, keep everything playing nicely But if i start forcing bit perfect playback, even i might not play nicely!.

the beresford was in my last system, my current one is in my sig. i just wondered if the increase in sound quality in going bit perfect would be comaprable to going from a 3.5mm from onboard sound, to a dedicated DAC, where i experimented and definitely noticed a difference, switching from one to the other. someone else was doing the switching but it wasnt blind, but thats by the by.

 

[Anonymous]

To put it as simply as I can: no.

 

[Alec]

Octopo:To put it as simply as I can: no.

I'm assuming that means it would not be comparable at all in which case whoohoo! thats exactly the question i was looking to answer.

Thanks.

 

[Anonymous]

I didn't mean to make that sound so blunt but that was my thoughts on the subject. You won't find a difference, if chappies with scientific paranormal monitoring equipment what to jump out of their stools so be the case but I don't think it will be. Waste of HD space unless you work at Dreamworks.

 

[Craig M.]

al7478:Yes, oh sarcy one

, but it also means not being able to test for an audible difference.

could you setup foobar or whatever, as bit perfect and use your speakers vol control. then compare that to wmp as you have it at the mo? if i had to guess though, i think i'd agree with octopo, you won't really notice an obvious difference.

 

[pete321]

No doubt about in my mind (and ears!) that bit perfect is noticeably better. With PJPro's help, I've set up Foobar2000 with the WASAPI component and the difference is very noticeable when playing my lossless PC based music. For a start, my DacMagic automatically changes to the sample rate of the music file playing, but more importantly, the sound is far more spacious and detailed, lossless music played via WMP sounds a bit flat and compressed by comparison.

 

[Alec]

Thanks for the spanner, Pete!

I'll give it a bash, Craig. Foobar and asio4all?

 

[Alec]

Well ive got foobar and asio, but the instructions foxed me a bit with installing the plugin to asio, then there were, aparently, things i had to find in foobar that just arent there.

if i ever get round to it the comparison wil be with itunes, as i believe wmp sounds too different to both foobar and itunes in the first place.

 

[Anonymous]

My approach to playing back sound via the PC is just based on common audio sense principals, keep signals untouched and as clean as possible whether digital or analogue.

I am currently using XP and if playing back music of movies, I always try and bypass the kmixer using Kernel streaming, purely and simply because it is an unnecessary step of digital filtering. Whilst it is possible to pass a 44.1khz stream through the kmixer without SRC (sample rate conversion) and sometimes you have to e.g when listening to Spotify, because of the way volume control mixer algorithm works...quote:-

Kmixer will mix all date. nfortunately, kMixer does just that -- it "mixes" all data
even when it doesn't need to. The mixing algorithm as implemented by
Microsoft has been shown in many cases to only have a signal to noise
ratio of about 92dB. The problem is that CDs have 96dB (or 97.5dB for
the hardcore signal processing guys) of dynamic range. Effectively,
kMixer overwrites the last couple of bits of the signal with garbage,
reducing CD data from 16 bit data to effectively 14 bit data

Whilst this is probably immaterial for listening to lossy audio (mp3/spotify), it is hardly ideal for 16-24bit lossless audio. The Vista mixer whilst being superior and allowing a fixed shared mode sampling rate still involves PCM integer to 32bit float conversion for its mixer stage, so it is best to avoid this using the WASAPI exclusive mode wherever possible for serious audio applications.

A bit perfect soundcard should then be able to output the original pcm wave in its native format either via the SPDIF output or through its own on-board DACs. If you have a good quality external DAC, then use it and feed it the purest digital signal possible (SPDIF/USB), if not purchase a soundcard based on quality of the onboard DACs and output stage and set the volume level to the maximum.

I have a Quad 99CDP2 which has an excellent 24/96 digital input section, so i need to feed bit perfect PCM via the computer's SPDIF. Anything other than bit perfect would defeat the whole object of using a quality DAC. Incidentally I don't use the variable outputs of the Quad, as i find the fixed outputs fed through a vishay dale stepped attenuator onto a Net Audio Quad 405 mk3 produces a more natural realistic sound.

 

[Anonymous]

bendrummond:
...quote:-

Kmixer will mix all date. nfortunately, kMixer does just that -- it "mixes" all dataeven when it doesn't need to. The mixing algorithm as implemented byMicrosoft has been shown in many cases to only have a signal to noiseratio of about 92dB. The problem is that CDs have 96dB (or 97.5dB forthe hardcore signal processing guys) of dynamic range. Effectively,kMixer overwrites the last couple of bits of the signal with garbage,reducing CD data from 16 bit data to effectively 14 bit data

Don't know where you got this from, this is highly unlikely and contrary to the Microsoft documentation: Kmixer only mixes data if there is something to mix. If there is only one source (as there should be!) there will be no resampling (or mixing), unless.the audio driver used requires a different sampling rate - kmixer then delivers that rate. Also, that kmixer by default resamples to 48k (as sometimes stated) is incorrect so it seems. It could be that the resampling algorithms are poor, and in general it should be avoided.

For bitperfect and undisturbed audio the _design_ of the newer implementation in Vista/7 is better Tried Wasapi - it sounded great when I listened the first time and I thought there was an improvement, but after doing some more extensive A/B comparisons with windows streaming plugin I concluded there was no difference to my ears.

 

[Anonymous]

Pete,

The quote above was taken from a forum, so I would take it with a pinch of salt as I was not 100% happy with it and was acually looking for something on Microsoft's site where one of the people involved with the kxmixer admitted that it was mathematically impossible for it to output at 100% as the 16th bit needs to be used for something else and could never be bit perfect because of this.

The article here (http://msdn.microsoft.com/en-us/library/ms790027.aspx) details how the kmixer should work and i can confirm with all other inputs muted, Spotifty will pass 44.1khz to the soundcard (so no SRC) but I would still prefer to bypass if possible.

W7/Vista has improved audio but WASPI works the same as XP Kernel streaming and I use both depending on what operating system I am using.

I will try and find the article on the MSDN site relating to the volume issue under XP.

 

[Anonymous]

This article may interest Vista users who playback lossless audio via direct sound (WMA lossless on Windows Media Player), it is written by the guy who wrote the Open Source bit perfect audio drivers for the CMedia 8738/8768 chipsets

http://code.google.com/p/cmediadrivers/wiki/Bitperfect

....Unfortunately though, Vista's reimplementation apparently doesn't
disable the sound processing when it isn't needed to the effect that
the audio service is not bitperfect anymore

This suggests ASIO/WASAPI is the only option in Vista to playback lossless 2 channel PCM audio.