Im just curious really i was looking through my itunes library and with all my music in apple lossless i have a range in bit rates of 1200 down to 473 at a sample rate of 44.1khz just wondered if this was roughly the norm? If anyone else looks that is! Thanks
Bit rate question
- Thread starter admin_exported
- Start date
The lossless bitrates will reflect the complexity of the particular tune - my iTunes library runs from 1108kbps down to around 600.
Simon Lucas:The lossless bitrates will reflect the complexity of the particular tune - my iTunes library runs from 1108kbps down to around 600.
Agreed.
Agreed.
Alec
Well-known member
Depends on the complexity of the music. A tune with many instruments will take more information (data rate) to encode than one with a single instrument and periods of silence. Frenetic dance or rock is usually around 1000-1100 kbps laid back jazz on the lower end.
eg invaders must die by the prodigy 1117 and Einaudi Le Onde 409kbps (both lossless)
Imagine the music written down in notation, Le onde is largely a single piano, the prodigy is a wall of noise.
eg invaders must die by the prodigy 1117 and Einaudi Le Onde 409kbps (both lossless)
Imagine the music written down in notation, Le onde is largely a single piano, the prodigy is a wall of noise.
professorhat
Well-known member
It's called Variable Bit Rate (or VBR) and ensures the best compression can be used and therefore smaller files.
Alec
Well-known member
I know of VBR, but i must be missing something as im not sure how a file that uses it can call itslef lossless...
a CD is about 1411 kbps so that is the maximum possible constant bitrate encoding possible.
lossless uses a VBR approach that uses 1411 as the top end (more or less) and re-encodes the music so that the same amount of information is encoded.
compression techniques like MP3 and AAC set a lower limit eg 320kbps and throw away some of the information, using a psychoacoustic model see http://en.wikipedia.org/wiki/Psychoacoustics
It is lossless in that nothing is lost from the 16bit 44.1khz stereo encoding from the CD.
the so called studio masters start with a higher quality 24bit 96khz signal.
However a quiet section with periods of silence will still result in a less than say 1411 encoding, hence the results for Einaudi
Similarly compressing an application is a lossless compression, you want the same thing back again
does that help?
lossless uses a VBR approach that uses 1411 as the top end (more or less) and re-encodes the music so that the same amount of information is encoded.
compression techniques like MP3 and AAC set a lower limit eg 320kbps and throw away some of the information, using a psychoacoustic model see http://en.wikipedia.org/wiki/Psychoacoustics
It is lossless in that nothing is lost from the 16bit 44.1khz stereo encoding from the CD.
the so called studio masters start with a higher quality 24bit 96khz signal.
However a quiet section with periods of silence will still result in a less than say 1411 encoding, hence the results for Einaudi
Similarly compressing an application is a lossless compression, you want the same thing back again
does that help?
Alec
Well-known member
zzgavin:
a CD is about 1411 kbps so that is the maximum possible constant bitrate encoding possible.
lossless uses a VBR approach that uses 1411 as the top end (more or less) and re-encodes the music so that the same amount of information is encoded.
compression techniques like MP3 and AAC set a lower limit eg 320kbps and throw away some of the information, using a psychoacoustic model see http://en.wikipedia.org/wiki/Psychoacoustics
It is lossless in that nothing is lost from the 16bit 44.1khz stereo encoding from the CD.
the so called studio masters start with a higher quality 24bit 96khz signal.
However a quiet section with periods of silence will still result in a less than say 1411 encoding, hence the results for Einaudi
Similarly compressing an application is a lossless compression, you want the same thing back again
does that help?
i think it does a little, thnaks, but i understand khz less well than kbps, so i need ot do some thinking and learning i think...
a CD is about 1411 kbps so that is the maximum possible constant bitrate encoding possible.
lossless uses a VBR approach that uses 1411 as the top end (more or less) and re-encodes the music so that the same amount of information is encoded.
compression techniques like MP3 and AAC set a lower limit eg 320kbps and throw away some of the information, using a psychoacoustic model see http://en.wikipedia.org/wiki/Psychoacoustics
It is lossless in that nothing is lost from the 16bit 44.1khz stereo encoding from the CD.
the so called studio masters start with a higher quality 24bit 96khz signal.
However a quiet section with periods of silence will still result in a less than say 1411 encoding, hence the results for Einaudi
Similarly compressing an application is a lossless compression, you want the same thing back again
does that help?
i think it does a little, thnaks, but i understand khz less well than kbps, so i need ot do some thinking and learning i think...
zzgavin:compression techniques like MP3 and AAC set a lower limit eg 320kbps and throw away some of the information, using a psychoacoustic model see http://en.wikipedia.org/wiki/Psychoacoustics
Interesting, particularly this bit:
"Given that the ear will not be at peak perceptive capacity when dealing with these limitations, a compression algorithm can assign a lower priority to sounds outside the range of human hearing. By carefully shifting bits away from the unimportant components and toward the important ones, the algorithm ensures that the sounds a listener is most likely to perceive are of the highest quality." (From the above-mentioned Wikipedia article.) Presumably this might account for some of the differences in SQ between different lossy formats.
Interesting, particularly this bit:
"Given that the ear will not be at peak perceptive capacity when dealing with these limitations, a compression algorithm can assign a lower priority to sounds outside the range of human hearing. By carefully shifting bits away from the unimportant components and toward the important ones, the algorithm ensures that the sounds a listener is most likely to perceive are of the highest quality." (From the above-mentioned Wikipedia article.) Presumably this might account for some of the differences in SQ between different lossy formats.
Just to add a few details:
With lossless the datastream is compressed in such a way that the original datastream can be recreated at playback. It is the result of the fact that almost all sequences of 0 and 1's can be stored more efficiently, without losing any information, just like zipping a computer file. As a consequence, indeed the compression rate varies, and random noise may even be more difficult to compress faithfully than real music. Different lossless formats differ in the way they try to find the optimal compression during encoding. The algorithim (eg for flac) may have a high setting so that it spends more time trying to look for the best possible compression and a low one that is fast but with less efficient compression. In practice the settings do not differ much, the low ones are already quite efficient. Most algorithms try the spend their computer time in the encoding phase (making the compressed file) and keep the decoding fast, so that playback is not limited by computing power on a (portable) system with a slow cpu.
Lossy compression is more interesting technically, as choices have to be made what information to keep and what can be ignored - not just informatics but also the human factor of real hearing sensations comes into play. Lots of different methods have been developed, but of course the higher the bitrate the higher the faithfullness. It is quite amazing that a reduction to 10-20 percent is possible without too much damage to the listener's appreciation. VBR helps because the algoritm can use more bits for complex music phrases and less for the simpler ones, resulting in a higher quality given an _average_ bit rate for a track. Downside is that at playback the cpu must be able to decode the complex sections without getting behind (which would result in a dropout). If storage space is not limiting there is no reason to resort to lossy compression.
Upsampling (by modern DACs) is in a way the reverse of lossy compression: an algothm is used to recalculate and reconstruct a hypothetical original higher quality datastream before it was reduced to the 44.1kHz/16bit. The latter is a lossy compression of the master recording that if I remember correctly usually uses a 96kHz/24 bits AD conversion. It involves intrapolation, but appears to make the postprocessing by analog filtering at playback easier/better.
Pete
With lossless the datastream is compressed in such a way that the original datastream can be recreated at playback. It is the result of the fact that almost all sequences of 0 and 1's can be stored more efficiently, without losing any information, just like zipping a computer file. As a consequence, indeed the compression rate varies, and random noise may even be more difficult to compress faithfully than real music. Different lossless formats differ in the way they try to find the optimal compression during encoding. The algorithim (eg for flac) may have a high setting so that it spends more time trying to look for the best possible compression and a low one that is fast but with less efficient compression. In practice the settings do not differ much, the low ones are already quite efficient. Most algorithms try the spend their computer time in the encoding phase (making the compressed file) and keep the decoding fast, so that playback is not limited by computing power on a (portable) system with a slow cpu.
Lossy compression is more interesting technically, as choices have to be made what information to keep and what can be ignored - not just informatics but also the human factor of real hearing sensations comes into play. Lots of different methods have been developed, but of course the higher the bitrate the higher the faithfullness. It is quite amazing that a reduction to 10-20 percent is possible without too much damage to the listener's appreciation. VBR helps because the algoritm can use more bits for complex music phrases and less for the simpler ones, resulting in a higher quality given an _average_ bit rate for a track. Downside is that at playback the cpu must be able to decode the complex sections without getting behind (which would result in a dropout). If storage space is not limiting there is no reason to resort to lossy compression.
Upsampling (by modern DACs) is in a way the reverse of lossy compression: an algothm is used to recalculate and reconstruct a hypothetical original higher quality datastream before it was reduced to the 44.1kHz/16bit. The latter is a lossy compression of the master recording that if I remember correctly usually uses a 96kHz/24 bits AD conversion. It involves intrapolation, but appears to make the postprocessing by analog filtering at playback easier/better.
Pete
One other interesting thing about compression.
On a DVD the video is compressed about 40:1, the audio about 6:1
We notice audio artefacts much more readily than video ones.
On a DVD the video is compressed about 40:1, the audio about 6:1
We notice audio artefacts much more readily than video ones.
zzgavin:
One other interesting thing about compression.
On a DVD the video is compressed about 40:1, the audio about 6:1
We notice audio artefacts much more readily than video ones.
True, but perhaps it is also that the complexity/information contents of most video signals is in general much less than that of an audio signal. Sampling up to 20kHz mixed signals (not that I can hear these any more..) with a 44.1Khz sampler is not really overkill. Or, in other words, if there is a red dot at position x,y in a picture is it quite likely that the surrounding pixels are also red(dish) and also to the next frames will contain a red spot at more or less the same position. I suspect that with audio this predictability is much less so, and hence more difficult to have a high compression rate for music.
One other interesting thing about compression.
On a DVD the video is compressed about 40:1, the audio about 6:1
We notice audio artefacts much more readily than video ones.
True, but perhaps it is also that the complexity/information contents of most video signals is in general much less than that of an audio signal. Sampling up to 20kHz mixed signals (not that I can hear these any more..) with a 44.1Khz sampler is not really overkill. Or, in other words, if there is a red dot at position x,y in a picture is it quite likely that the surrounding pixels are also red(dish) and also to the next frames will contain a red spot at more or less the same position. I suspect that with audio this predictability is much less so, and hence more difficult to have a high compression rate for music.
manicm
Well-known member
There seems to be more than just 1s and 0s when it comes to CDs and other digital formats. Jitter plays a big part as well.
Here's some food for thought. I read an interesting article in a recent issue of HiFi+ where they state there may be more jitter coming out of a PC than a good CDP. In other words they don't rate a hard drive highly in this regard, but definitely rate SSDs.
Here's some food for thought. I read an interesting article in a recent issue of HiFi+ where they state there may be more jitter coming out of a PC than a good CDP. In other words they don't rate a hard drive highly in this regard, but definitely rate SSDs.
Alec
Well-known member
manicm:
There seems to be more than just 1s and 0s when it comes to CDs and other digital formats. Jitter plays a big part as well.
Here's some food for thought. I read an interesting article in a recent issue of HiFi+ where they state there may be more jitter coming out of a PC than a good CDP. In other words they don't rate a hard drive highly in this regard, but definitely rate SSDs.
But can they hear it?
There seems to be more than just 1s and 0s when it comes to CDs and other digital formats. Jitter plays a big part as well.
Here's some food for thought. I read an interesting article in a recent issue of HiFi+ where they state there may be more jitter coming out of a PC than a good CDP. In other words they don't rate a hard drive highly in this regard, but definitely rate SSDs.
But can they hear it?
I think this is really complete nonsense. Pc's buses (PCI, USB) are not designed for real time operations and jitter free delivery of data, and they never can do this. They should provide a datastream without dropouts to an external dac or internal soundcard with dac, and these components are responsible for the rest. This is not a big problem for a well configured pc as stereo involves modest data rates. If the data comes from a hd or ssd or usb stick or whatever storage medium is totally irrelevant. Of course, it is entirely possible that an internal soundcard or external dac is vastly inferior to a good cdp but this depends on the design and quality of the components used (where jitter can play a role).
P.
P.
manicm
Well-known member
Pete10:
I think this is really complete nonsense. Pc's buses (PCI, USB) are not designed for real time operations and jitter free delivery of data, and they never can do this. They should provide a datastream without dropouts to an external dac or internal soundcard with dac, and these components are responsible for the rest. This is not a big problem for a well configured pc as stereo involves modest data rates. If the data comes from a hd or ssd or usb stick or whatever storage medium is totally irrelevant. Of course, it is entirely possible that an internal soundcard or external dac is vastly inferior to a good cdp but this depends on the design and quality of the components used (where jitter can play a role).
P.
I don't think it's that far-fetched, It's in the latest issue where they test iQ50 and the new Benchmark DAC1 USB.
Remember a hard drive is a mechanical device with moving parts. And that is why they would recommend a SSD in ultimate terms. They did however state that a good DAC like the DAC1 can do wonders as well, regardless of the transport.
So maybe they're contradicting themselves a little.
I think this is really complete nonsense. Pc's buses (PCI, USB) are not designed for real time operations and jitter free delivery of data, and they never can do this. They should provide a datastream without dropouts to an external dac or internal soundcard with dac, and these components are responsible for the rest. This is not a big problem for a well configured pc as stereo involves modest data rates. If the data comes from a hd or ssd or usb stick or whatever storage medium is totally irrelevant. Of course, it is entirely possible that an internal soundcard or external dac is vastly inferior to a good cdp but this depends on the design and quality of the components used (where jitter can play a role).
P.
I don't think it's that far-fetched, It's in the latest issue where they test iQ50 and the new Benchmark DAC1 USB.
Remember a hard drive is a mechanical device with moving parts. And that is why they would recommend a SSD in ultimate terms. They did however state that a good DAC like the DAC1 can do wonders as well, regardless of the transport.
So maybe they're contradicting themselves a little.
manicm
Well-known member
manicm: don't think it's that far-fetched, It's in the latest issue where they test iQ50 and the new Benchmark DAC1 USB.
Remember a hard drive is a mechanical device with moving parts. And that is why they would recommend a SSD in ultimate terms. They did however state that a good DAC like the DAC1 can do wonders as well, regardless of the transport.
So maybe they're contradicting themselves a little.
Sorry, but it makes no sense to me at all, hd and ssd have the same interface and deliver the data in the same way, usually in big chunks, and can maintain a datastream much much higher than required. Moreover, they are both designed to deliver bitperfect streams, unlike cdp's due to the original ancient redbook definition. The only reason to prefer a ssd is that it is totally silent, unlike a hd (although there are many cdp's that are noisier than a modern hd).
Remember a hard drive is a mechanical device with moving parts. And that is why they would recommend a SSD in ultimate terms. They did however state that a good DAC like the DAC1 can do wonders as well, regardless of the transport.
So maybe they're contradicting themselves a little.
Sorry, but it makes no sense to me at all, hd and ssd have the same interface and deliver the data in the same way, usually in big chunks, and can maintain a datastream much much higher than required. Moreover, they are both designed to deliver bitperfect streams, unlike cdp's due to the original ancient redbook definition. The only reason to prefer a ssd is that it is totally silent, unlike a hd (although there are many cdp's that are noisier than a modern hd).
Though your example is not about jitter. audio and video require in order presentation and jitter can affect this.
Computer interfaces are about delivering accurate data, but the order might not be perfect, the operating system makes up for this.
Editing video is a good example, having a clean drive as opposed to a half full drive makes a big difference to preformance, not quite the same, but there are jitter measurement tools and hard drives are affected. I can't find the appropriate article now and have work to finish, but computer delivery of data is not jitter free, it is just generally not an issue with non-av computer usage.
Computer interfaces are about delivering accurate data, but the order might not be perfect, the operating system makes up for this.
Editing video is a good example, having a clean drive as opposed to a half full drive makes a big difference to preformance, not quite the same, but there are jitter measurement tools and hard drives are affected. I can't find the appropriate article now and have work to finish, but computer delivery of data is not jitter free, it is just generally not an issue with non-av computer usage.
To me, they are just muddying the waters with this statement. Many people feel now that set up properly, audio delivered from a computer is superior to a decent cdp. It is true that music played back from a computer (no matter what type of drive) does have to contend with jitter but there are ways around that by getting a DAC that controls the flow of data rather than leaving it to the computer which has other things to worry about - they do tend to be expensive DAC's though. All DAC's try to minimize jitter regardless of if they are operating in asynchronous mode or not (asynchronous mode is when the DAC controls the audio data from the computer).
I have a Cyrus cdp that is excellent, but compared to my Cyrus int amp/Benchmark DAC/apple tv/computer audio (and video) set up, I am convinced the audio is better streaming it from the computer. As for the original question, you computer is doing the correct thing. There is a huge difference is the bit rate for lossless audio files. In fact, I have the same piece of music - Beethoven's Piano Sonata No. 14 in lossless 16/44 at 331kbps (Immortal Beloved soundtrack) and a performance by Artur Pizarro that is 2547kbps in 24/96 - now its a different performance obviously but of the same piece of music and it can vary widely. Most of my lossless audio is in the 700-900kbps range when using 16/44.
If anybody is wondering, iTunes can't natively record tracks in 24/96 lossless but if you buy them from Linn for example in 24/96 Flac and use MAX to convert them to Apple Lossless, iTunes will import 24/96 lossless audio files.
I have a Cyrus cdp that is excellent, but compared to my Cyrus int amp/Benchmark DAC/apple tv/computer audio (and video) set up, I am convinced the audio is better streaming it from the computer. As for the original question, you computer is doing the correct thing. There is a huge difference is the bit rate for lossless audio files. In fact, I have the same piece of music - Beethoven's Piano Sonata No. 14 in lossless 16/44 at 331kbps (Immortal Beloved soundtrack) and a performance by Artur Pizarro that is 2547kbps in 24/96 - now its a different performance obviously but of the same piece of music and it can vary widely. Most of my lossless audio is in the 700-900kbps range when using 16/44.
If anybody is wondering, iTunes can't natively record tracks in 24/96 lossless but if you buy them from Linn for example in 24/96 Flac and use MAX to convert them to Apple Lossless, iTunes will import 24/96 lossless audio files.
Similar threads
TRENDING THREADS
-
-
I test OLED TVs for a living – and trust me, different sizes perform differently- Started by What Hi-Fi?
- Replies: 2
-
-
-
Biggest Upgrade Improvement? Rega Aphelion 2 or Aura?
- Started by hybridauth_Google_112692162581130744674
- Replies: 5
-
-
Facebook
Twitter
Instagram