Skip to main content

Notice

Please note that most of the software linked on this forum is likely to be safe to use. If you are unsure, feel free to ask in the relevant topics, or send a private message to an administrator or moderator. To help curb the problems of false positives, or in the event that you do find actual malware, you can contribute through the article linked here.
Topic: Bit-perfect playback questions (Read 47105 times) previous topic - next topic
0 Members and 2 Guests are viewing this topic.

Bit-perfect playback questions

I'd say I've gotten into serious music listening the moment I learned that MP3's were lossy and that there IS a difference in SQ based on how the audio was ripped from the CD and how it was encoded. Since then, I've collected my music collection in lossless quality, invested in a pair of high quality headphones, and researched into digital audio in general.

For the most part, I've been able to learn pretty much the basics. However, I've recently stumbled upon a topic that I have yet to understand and find a straightforward answer to: bit-perfect playback. I've learned so far that Windows has an internal mixer that resamples audio played before it goes through the speakers. This internal mixer (generally speaking) decreases or alters the sound quality of an audio file before it reaches the speakers. Plugins such as WASAPI and ASIO bypass this mixer in order for the bits played to reach the headphones unaltered. I know what I'm saying may be incorrect but bear with me since I am fairly new to this.

Further research shows that in Windows Vista and 7, does not require WASAPI or ASIO plugins for bit-perfect playback because the sound settings are automatically sampled above 16bit/44.1KHz (mine are at 24bit/48KHz) and so if the audio is <= that sampling rate, the audio is essentially unaltered. However, other sources state otherwise.

In short, I've yet to find a simple answer to the following questions:
1) Does one NEED WASAPI or ASIO in order to enjoy bit-perfect playback?
2) Is bit-perfect playback even something worth pursuing? (I know it's left up to the listeners ears but let us assume I can pickup the slightest change in quality)
3) If neither is needed, then what sampling rate should I set for "Sound Settings" in Windows 7 to achieve bit-perfect playback?
4) Generally, I'm a bit confused about the whole topic. I've checked out certain sites like thewelltemperedcomputer.com which has shed some light but still hasn't explained everything.

I've read these forums for quite some time and I must thank the entire community because without this site, I wouldn't have learned what I know now about digital audio. For this reason, I've decided to become a member.

Stats:
OS: Windows 7 64-bit
Music Player: Winamp v5.61

Bit-perfect playback questions

Reply #1
I've learned so far that Windows has an internal mixer that resamples audio played before it goes through the speakers. This internal mixer (generally speaking) decreases or alters the sound quality of an audio file before it reaches the speakers.

Not necessarily - it is absolutely possible to get bit-perfect playback in WinXP, but it is also fairly easy not to.

As long as your playback software is the only application sending audio to your sound card, and all volumes are at max (unity gain), kmixer can pass bit-perfect audio to your sound card...which may or may not do resampling of its own.

Quote
Further research shows that in Windows Vista and 7, does not require WASAPI or ASIO plugins for bit-perfect playback because the sound settings are automatically sampled above 16bit/44.1KHz (mine are at 24bit/48KHz) and so if the audio is <= that sampling rate, the audio is essentially unaltered.

No, all audio that does not match the set output sampling rate and bit depth is resampled to match it.  So, for bit-perfect playback of 16/44.1 audio, you need to set your audio output device's bit depth and sample rate to 16/44.1.  FWIW, this is also how OS X handles it.
"Not sure what the question is, but the answer is probably no."

Bit-perfect playback questions

Reply #2
Thanks man. You cleared up the confusion with Windows' internal mixer. I was wondering how that worked. BTW: I've played around with the WASAPI plugin in Winamp and foobar2000 and I've experience a big boost in volume whenever I use the WASAPI plugin. When I get back to DirectSound, the volume boost drops. Just wondering if that's normal. I've also experienced some ever-so-slight clipping sounds when I use Winamp with the WASAPI plugin. Doesn't happen with foobar2000 though.

Bit-perfect playback questions

Reply #3
I'm not an expert on how to get bit-perfect audio, but I'll throw-in my 2-cents...

Quote
I'd say I've gotten into serious music listening the moment I learned that MP3's were lossy and that there IS a difference in SQ based on how the audio was ripped from the CD and how it was encoded.
If you want to avoid lossy encoding, that's OK...  Lossless is the "safe" way to go.    But with most music, you won't hear any difference between a high-quality MP3/AAC and the uncompressed original (in a proper blind listening test).  There can be an audible difference with some music that's "hard" to compress, or if you use low-bitrate encoding.

I guess everybody's heard bad MP3s, and I was a bit of a snob about lossy compression until I realized that Dolby AC3 is lossy...  And I realized how incredible 5.1 channel AC3 can sound on a really a good home theater system!  DTS is lossy too.  Whenever there's a choice between uncompressed stereo LPCM, and 5.1 channel AC3 (or DTS) on a DVD...  I always choose the (lossy) surround sound!

Quote
Generally, I'm a bit confused about the whole topic.
The main thing is, digital audio is very good!  ...Generally much better than human hearing (assuming lossless or high-quality encoding).  As long as you don't resample to low sample rates (which will eliminate high frequencies) or to low bit-depths (at 8-bits you'll hear quantization noise), everything will be OK.  (Be very skeptical of "audiophile nonsense.") 


And if you are not hearing noise from your soundcard, you're probably not hearing any other defects either...  Some soundcards do pick-up noise from the circuitry inside the computer.  If you are not hearing noise and you're looking for "better sound",  put your effort (and money) into the analog side...  i.e. Try some better headphones/speakers, or a bigger amp, etc.  Or you can intententionally manipulate the sound with EQ, or Dolby Pro Logic (or Dolby Headphone), etc.


Quote
2) Is bit-perfect playback even something worth pursuing? (I know it's left up to the listeners ears but let us assume I can pickup the slightest change in quality)
It isn't someting I worry about.  I've done quite a bit of re-sampling between 44.1kHz (CD) and 48kHz (DVD), and I never hear any difference (in either direction).  I'm not doing ABX tests, but I'm not hearing a difference in "casual listening", and I don't have a choice anyway...

Quote
I've played around with the WASAPI plugin in Winamp and foobar2000 and I've experience a big boost in volume whenever I use the WASAPI plugin.
I would guess the louder signal is bit-perfect.  As mixminus1 said, if you don't want to alter the bits you need unity gain.  The computer's volume control works in the digital domain, so you can't reduce the volume without altering the bits/bytes.    But, there's no harm in that... It's really no different than reducing the analog volume.    In both cases, you are reducing the "resolution" or dynamic range, and in both cases, anything "bad" happens at reduced-volume levels, so you never hear the loss of dynamic range unless you re-boost the levels later.  If you don't want to control the volume digitally,  computer-speakers usually have an analog volume control, or you can use the analog volume control on an amp or headphone amp.

And of course, the analog output from your soundcard is never "bit perfect" since it's not digital... There are no bits or bytes in the analog output!  If you've got a soundcard with an S/PDIF output, it's possible to send bit-perfect data from your hard drive (or DVD) into your reciever or S/PDIF computer speakers, or to a USB DAC or USB headphone amp.

Bit-perfect playback questions

Reply #4
DVDdoug: Your response pretty much answered most (if not all of my questions). Thanks again for your response. And yes, I try not to buy into the whole audiophile nonsense, I try to find a middle-way between getting the best out of what I got and not pushing it too far.

Bit-perfect playback questions

Reply #5
There are two things that the PC sound system can do to your audio that will disrupt bit perfection.
  • Level changes
  • Sample rate conversion

In answer to your question (2) about whether bit perfection matters. Some chances are more audible than others. The first is not a serious concern for most applications. The second is of greater concern especially on older operating systems.

Bit-perfect playback questions

Reply #6
I try not to buy into the whole audiophile nonsense

That's good.

Can you explain mechanism you used to arrive at the conclusion that you need lossless because of sound quality differences?

Bit-perfect playback questions

Reply #7
Can you explain mechanism you used to arrive at the conclusion that you need lossless because of sound quality differences?

Well, I had this one album encoded in 192CBR MP3 (Transdimensional by Dimension 5). I obtained the lossless version of the album and IMMEDIATELY noticed the difference in SQ. The bass was much deeper and the sound was far more spacious. Overall, it just sounded a helluva lot better. However, I have a 2GB Sansa Fuze and there was no way I could put most of my music on it in lossless format. So, doing a bit more research I learned about the Ogg Vorbis codec. I've read articles and conducted my own listening tests as well as compared spectrographs (V0 MP3 vs. Q6.5 OGG vs. FLAC) and found that OGG sounds way better than MP3 and could barely notice the difference between the lossless version and the OGG version.

For the record, I understand that the type of music that one listens to does make a difference on how hard/easy it is to tell the difference between lossless and lossy. The genres I listen to are almost always electronica; specifically Goa Trance, Psytrance, Psychedelic, Progressive Trance, Tribal, etc. Of course, there is this one group that's essentially genre-less called "Shpongle". Those of you that know this group know that this group can only be described as psychedelic and that the vast array of sounds and effects they use in their music is unmatched. Any loss of quality is easily noticeable in most, if not all of their tracks.

To further prove my point, here's two sets of spectrographs comparing FLAC (top), Q6.5 OGG (middle), V0 MP3 (bottom).

Shpongle - Dorset Perception
dBV: http://i53.tinypic.com/4u86lu.png
dBV^2: http://i56.tinypic.com/35ib24i.png

Solar Fields - Circles of Motion
dBV: http://i56.tinypic.com/23ihr9l.png
dBV^2: http://i55.tinypic.com/2llzj2p.png

These spectrographs further cemented my choice in choosing lossless over lossy. Even more so, it was the last thing I needed to make the conversion from MP3's to OGG as my choice for lossy compression. Of course, I didn't rely on just the graphs, I did blind sound tests and there was an audible difference. Not much but they were still there. It's for these reasons that I do any quality tests with lossless audio. In any case, I hope this reply doesn't give the wrong impression that I'm not willing to learn anything new. Please, if there is something wrong with my logic, I'd love to know. I'd rather know I'm wrong than to not know at all. Thanks again!

Bit-perfect playback questions

Reply #8
Properly controlled, level matched and time synchronized double blind tests are the only useful means to judge the audible sound quality of lossy codecs on this forum.

This is stated in our Terms of Service:
"All members that put forth a statement concerning subjective sound quality, must -- to the best of their ability -- provide objective support for their claims. Acceptable means of support are double blind listening tests (ABX or ABC/HR) demonstrating that the member can discern a difference perceptually, together with a test sample to allow others to reproduce their findings. Graphs, non-blind listening tests, waveform difference comparisons, and so on, are not acceptable means of providing support."

It goes further:
-> 8.

Hydrogenaudio is supposed to be an objectively minded community that relies on double-blind testing and relevant methods of comparison in discussion about sound quality. The usual "audiophile" speak of non-audio related terms which are completely subjective and open to redefinition on a whim, are useless for any sort of progression in discussion.

This rule is the very core of Hydrogenaudio, so it is very important that you follow it.

Here is a discussion explaining why
http://www.hydrogenaudio.org/forums/index....showtopic=11442

You can read how to easily perform double blind listening tests here :
http://www.hydrogenaudio.org/forums/index....howtopic=16295

I hope you take this constructively.  I would not have bothered if I didn't think you might have been misled through improper methods of testing.  It was your presentation of support in the way of graphs that tipped me off.  One of the links above talks about how to conduct proper double-blind tests.  Please read it and then consider whether the "blind" tests you conducted were sufficient.

Thanks for reading!

Bit-perfect playback questions

Reply #9
I hope you take this constructively.  I would not have bothered if I didn't think you might have been misled through improper methods of testing.  It was your presentation of support in the way of graphs that tipped me off.  One of the links above talks about how to conduct proper double-blind tests.  Please read it and then consider whether the "blind" tests you conducted were sufficient.

Thanks for reading!

Thanks man. I appreciate the effort you took in correcting me. I honestly didn't know graphs were insufficient given that they show what the ears may not be able to hear. Then again, that may not be the case and I wouldn't know. I'll definitely take a look at the links you mentioned.

Bit-perfect playback questions

Reply #10
Can you explain mechanism you used to arrive at the conclusion that you need lossless because of sound quality differences?

Well, I had this one album encoded in 192CBR MP3 (Transdimensional by Dimension 5). I obtained the lossless version of the album and IMMEDIATELY noticed the difference in SQ. The bass was much deeper and the sound was far more spacious. Overall, it just sounded a helluva lot better.


Have you tried encoding the lossless version at 192CBR, and comparing it to the original MP3 you had? Assuming you heard correctly, the symptoms you describe seem typical of a better/different master; not of an inherent problem with MP3/192.

As always, use an ABX method to compare the songs.

Bit-perfect playback questions

Reply #11
If you use HDMI audio there are practical benefits of using WASAPI exclusive (which is needed for bit-perfect in Vista/7), which is that it will only send the number of channels of the source, so your receiver can apply processing appropriately. There's a way to change the Windows mixer's output according to # of channels from the source, but since that was broken when the protected audio path drivers from ATI and Realtek were introduced, I think it's possible DRM prevents it. I would be happy if I didn't depend on exclusive mode to do this to be honest.

Bit-perfect playback questions

Reply #12
Please, if there is something wrong with my logic, I'd love to know. I'd rather know I'm wrong than to not know at all. Thanks again!

Well, one does not listen with their eyes so I'm going to ignore the spectrograms. You should too. Also, if you ever run across someone who wants to demonstrate the evil of lossy encoding by letting you hear the mathematical difference between the lossy coding and the original, ignore that too.

On the other hand, electronically generated sounds are among the most challenging for lossy coders to handle correctly and extra storage space in many contexts is easy to come by. So, if you think you can hear a difference and it makes you happy, FLAC away.

Bit-perfect playback questions

Reply #13
Have you tried encoding the lossless version at 192CBR, and comparing it to the original MP3 you had? Assuming you heard correctly, the symptoms you describe seem typical of a better/different master; not of an inherent problem with MP3/192.

As always, use an ABX method to compare the songs.

No I haven't. I've just recently started using foobar2000 and I've installed the ABX Comparator component. I've tried out a couple of ABX tests and it definitely eliminates the possibility of a placebo effect. Definitely will be doing more of those.

Well, one does not listen with their eyes so I'm going to ignore the spectrograms. You should too. Also, if you ever run across someone who wants to demonstrate the evil of lossy encoding by letting you hear the mathematical difference between the lossy coding and the original, ignore that too.

On the other hand, electronically generated sounds are among the most challenging for lossy coders to handle correctly and extra storage space in many contexts is easy to come by. So, if you think you can hear a difference and it makes you happy, FLAC away.

Of course, at the end of the day, if all my albums in FLAC don't even amount to 10GB and I've got an MP3 player with 32GB of disc space, then by all means there's no need for lossy encoding. I guess the fault is my logic that all lossy codecs decrease sound quality without me properly testing for that. I must thank you all for telling about ABX testing because I honestly didn't fully understand what it was for.

Bit-perfect playback questions

Reply #14
I've been doing a lot of reading about this lately as well.  I have an X-FI card based on the EMU20K2 chip I got for gaming.  I don't have SPDIF capable speakers so I have them connected to the analog out of the sound card.  I generally use foobar2000 with the ASIO or WASAPI with most of my collection in FLAC.  They sound the same to me and when possible I config games to use the sound hardware instead of software.

So, my query:  Is the "bitperfect" playback even much to consider when using analog outputs on a puter?  I do notice there is some minor differences between directsound and ASIO/WASAPI but I only really notice it because I'm intending to listen intently for any differences.  I'm coming at this from a user perception, not based on graphs and what not.  The biggest difference is volume output with ASIO/WASAPI sounding considerably louder and my having to compensate for it but once leveled down, I'm hard pressed to really notice.  I mostly play music and while gaming and I hear that using anything other than directsound in that circumstances isn't always a good idea.  It was never made clear why.

Bit-perfect playback questions

Reply #15
I do notice there is some minor differences between directsound and ASIO/WASAPI but I only really notice it because I'm intending to listen intently for any differences.  I'm coming at this from a user perception, not based on graphs and what not.

If your perception was not ascertained through double-blind testing then we have no way to rule out that you weren't simply imagining a difference.

TOS #8 is very clear on the matter: if you can't substantiate your claim with objective testing then aren't allowed to present it.

http://www.hydrogenaudio.org/forums/index.php?showtopic=3974

So, was your comparison between DirectSound and ASIO/WASAPI double-blind?

Bit-perfect playback questions

Reply #16
Quote
So, if you think you can hear a difference and it makes you happy, FLAC away


At the end of the day, being happy is probably the most important thing!

Even if one can't hear a difference, surely it makes sense to use the highest quality encoding available for any given situation?

Ali.

Bit-perfect playback questions

Reply #17
Even if one can't hear a difference, surely it makes sense to use the highest quality encoding available for any given situation?


Generally, the given situation is "not needlessly consuming storage space" and "not needlessly spending money".

 

Bit-perfect playback questions

Reply #18
Yes, especially considering the present HDD shortage and crazy prices.  . . .

Bit-perfect playback questions

Reply #19
Yes, especially considering the present HDD shortage and crazy prices.  . . . :cry:


Waste is a little bit more important when you're considering streaming, which is how many applications are using media today.  An FLAC of a given file can still be five to six times larger than a 256kbps AAC, and I'd daresay that few would be able to tell the difference.
FLAC -2 w/ lossyWAV 1.3.0i -q X -i

Bit-perfect playback questions

Reply #20
1) Does one NEED WASAPI or ASIO in order to enjoy bit-perfect playback?
2) Is bit-perfect playback even something worth pursuing? (I know it's left up to the listeners ears but let us assume I can pickup the slightest change in quality)
3) If neither is needed, then what sampling rate should I set for "Sound Settings" in Windows 7 to achieve bit-perfect playback?


1 – Windows default is DS (Direct Sound).
All audio is send to the mixer. The audio is converted to 32 bit float prior to processing and converted back to 16 or 24 bit depending on the audio device.
All output is dithered.
By design DS is not bit perfect (the dither).
ASIO/WASAPI bypass the mixer. The audio path between the media player and the audio device is bit transparent. This does not guarantee bit perfect playback as the media player might apply DSP e.g. volume control.

2 – Do a listening test. Set Win volume and media player volume to 100%. Disable all DSP (enhancements, etc). Compare DS with e.g. WASAPI.

3 – See 1.
Resampling might create audible artifacts depending on its implementation.
See http://src.infinitewave.ca/ for a couple of examples.
Do a listening test.
Play CD quality (44.1 kHz)  and set the sample rate in the audio panel to 44.1
Do the same with a different sample rate e.g. 48


TheWellTemperedComputer.com

Bit-perfect playback questions

Reply #21
I've played around with the WASAPI plugin in Winamp and foobar2000 .... I've also experienced some ever-so-slight clipping sounds when I use Winamp with the WASAPI plugin. Doesn't happen with foobar2000 though.
If I want to use WASAPI in Winamp - What plugin should I use?  I thought bit-perfect was only possible if WASAPI is using exclusive mode?

Bit-perfect playback questions

Reply #22
I've played around with the WASAPI plugin in Winamp and foobar2000 .... I've also experienced some ever-so-slight clipping sounds when I use Winamp with the WASAPI plugin. Doesn't happen with foobar2000 though.
If I want to use WASAPI in Winamp - What plugin should I use?  I thought bit-perfect was only possible if WASAPI is using exclusive mode?

I don't know which plug-in would do it, but you get unaltered (bitperfect) output with both kinds of WASAPI, ASIO, and Kernel Streaming.  They all bypass the Windows mixer.

Bit-perfect playback questions

Reply #23
You can get bitperfect output with any plain old media player over DirectSound if you leave your device's sample rate at the same setting as your files (e. g. 44.1 kHz) and set the volume to 100% since Vista.

This thread is hysterical and BS marketing driven.

16 bit INT -> 32 bit FLOAT -> 16 bit INT conversion, et ceteris paribus, is lossless, before you start the "but, but, but ... float pipeline"-whining.

Bit-perfect playback questions

Reply #24
I don't know which plug-in would do it, but you get unaltered (bitperfect) output with both kinds of WASAPI, ASIO, and Kernel Streaming.  They all bypass the Windows mixer.

No, WASAPI shared uses the Win mixer so dithered output

WASAPI exclusive bypasses the Win mixer

http://msdn.microsoft.com/en-us/library/dd...0(v=VS.85).aspx
TheWellTemperedComputer.com