advertisement


WAV better than FLAC due to increased processing load on the CPU of the latter?

vln

vln

Shuns mooks. And MQA.
This guy

http://www.audiostream.com/content/cut-flac

claims that WAV sounds better than FLAC due to the fact that the latter is compressed, and thus first needs to be decompressed before playing. The decompressing puts a load on the CPU and this supposedly affects sound quality in a negative way.

Sounds a bit far fetched to me - what do you think?


Samuel.
 
Unless your system's CPU under load is giving horrific amounts of EMI... very far stretched.
 
Sam, think about the logic, computers do 100's of tasks in milliseconds. Decompression is a single, very simple one, even a P3 or P4 with 1GB of RAM will be fine.
 
spacey said:
Sam, think about the logic, computers do 100's of tasks in milliseconds. Decompression is a single, very simple one, even a P3 or P4 with 1GB of RAM will be fine.
Click to expand...

FLAC codecs are designed to be light on the CPU on the decode cycle, but naturally far more harsh on encoding. Some real **** on the net.
 
Thats right, its also simple to see via task manager. I use a core2 Duo T8400 Toshiba Tecra M10 with a 1TB HDD and 4GB of ram. When decoding i see about 2% CPU usage.
When transcoding from FLAC to ALAC, i see about 30-40% dual CPU usage. It's understandable, as it's creating a copy file, changing codecs and re-tagging via dbpoweramp.
 
It's common to encounter this viewpoint in serious discussion of computer transports, and it can't lightly be dismissed: any transport is sensitive to fluctuations in rail noise.

The objections raised against FLAC v WAV are framed in terms of what the PC is 'capable' of in percentage terms - a less than helpful mindset based on IT practice rather than audio system design.

In general terms, more processor activity is bad; more disk access is bad; more heat is bad; more vibration is bad; more current variation is bad; more interrupts are bad - all induce jitter and add to the interplay of galvanic noise. Quiet is good.

The bottom line is this: it's probably not possible that WAV can sound worse than FLAC, but on-the-fly decoding of a compressed lossless format like Apple Lossless or FLAC could: particularly in low-powered, clean-running computers. In our tests, more acute listeners were consistently able to detect subtle differences between the two formats, without knowing in advance what they were listening to. Similar results can be found in widely diverse circumstances.

Naim weighed into the debate recently after making some detailed measurements of their streamers. They concluded that there were measurable differences in terms of processor activity and rail noise, but that these were not significant enough to translate to audible effects on the output.

The best informed sources on this point tend to be rather open-minded about the possibilities.
 
What do you think the "System Idle Process" is doing when nothing else requires those CPU cycles?

Ed: some modern processors can reduce the clock speed and send cores "to sleep" when not doing useful work, older processors don't do that.
 
Dear god, from a computer science perspective there is very, very very little load on a modern day CPU decoding FLAC content. The system is practically asleep 99% of the time playing FLACs, and unless its a really really noisy system, is unlikely to make a spot of difference.

I'm officially keeping out of this thread as of now... Agree with spacey - here comes the ****.
 
Most music is recorded on computers, and processor activity changes all over the pace while doing so. With the use of external converters it should make no difference what the power rails in the PC do.
 
I'd be interested to see measured jitter for both outputs and also a variety of noise measurements, thd, eye pattern etc, stuff beyond bit-perfect. But frankly if your dac renders them different, it's not fit for purpose, it's PSRR isn't up to snuff and neither is its signal isolation.
 
item said:
It's common to encounter this viewpoint in serious discussion of computer transports, and it can't lightly be dismissed: any transport is sensitive to fluctuations in rail noise.

The objections raised against FLAC v WAV are framed in terms of what the PC is 'capable' of in percentage terms - a less than helpful mindset based on IT practice rather than audio system design.

In general terms, more processor activity is bad; more disk access is bad; more heat is bad; more vibration is bad; more current variation is bad; more interrupts are bad - all induce jitter and add to the interplay of galvanic noise. Quiet is good.

The bottom line is this: it's probably not possible that WAV can sound worse than FLAC, but on-the-fly decoding of a compressed lossless format like Apple Lossless or FLAC could: particularly in low-powered, clean-running computers. In our tests, more acute listeners were consistently able to detect subtle differences between the two formats, without knowing in advance what they were listening to. Similar results can be found in widely diverse circumstances.

Naim weighed into the debate recently after making some detailed measurements of their streamers. They concluded that there were measurable differences in terms of processor activity and rail noise, but that these were not significant enough to translate to audible effects on the output.

The best informed sources on this point tend to be rather open-minded about the possibilities.
Click to expand...

Hey I can take this lightly! But then again I have nothing to sell. Don't sweat it lads, your PC is more than up to the task, just don't run to many other processes at the same time or you may (heaven forbid) hear something akin to a click on a record.

Your Albertan Pal
Louballoo
 
Lou,

No need to be rude, so I have blasted your impolite post.

Joe
 
vln said:
This guy

http://www.audiostream.com/content/cut-flac

claims that WAV sounds better than FLAC due to the fact that the latter is compressed, and thus first needs to be decompressed before playing. The decompressing puts a load on the CPU and this supposedly affects sound quality in a negative way.

Sounds a bit far fetched to me - what do you think?


Samuel.
Click to expand...

I think that is nonsense.
itunes on a 2 year old mac playing apple lossless, which achieves much the same compression as FLAC, gives about 1% CPU usage.
 
Why do people making these wild claims of lossless formats sounding different seldom present proof to back them up?

Personally, I think the deliberate seeding of such misinformation is rude at best, and frankly highly irresponsible, and borders on a malignancy when such concepts, born of undisclosed, possibly commercial, motives, are aired on the internet, where their inevitable spread, and repetition on other fora, could be seen to lend credence where absolutely non is due.

You've probably all seen it, someone asks a question, poses a conundrum, and the answer is forthcoming. It's the audiophile catechism, put there for people googling and browsing the internet to see. People passively get answers without ever having asked the question, ideas planted. heh.

TL;DR - don't worry about what lossless format to use, though wav is pretty poor, it wastes space, very limited tagging support, no embedded album art...
 
Andrew.B said:
Why do people making these wild claims of lossless formats sounding different seldom present proof to back them up?

Personally, I think the deliberate seeding of such misinformation is rude at best, and frankly highly irresponsible, and borders on a malignancy when such concepts, born of undisclosed, possibly commercial, motives, are aired on the internet, where their inevitable spread, and repetition on other fora, could be seen to lend credence where absolutely non is due.

You've probably all seen it, someone asks a question, poses a conundrum, and the answer is forthcoming. It's the audiophile catechism, put there for people googling and browsing the internet to see. People passively get answers without ever having asked the question, ideas planted. heh.

TL;DR - don't worry about what lossless format to use, though wav is pretty poor, it wastes space, very limited tagging support, no embedded album art...
Click to expand...

Joe - he said what I said. Maybe a little nicer though.

Louballoo
 
Lou,

It's simple: Attack the argument, not the person.

Your first post was fine, the second wasn't.

Joe
 
Actually FLAC sounds better than WAV for the opposite reason.

WAV files are bigger than FLAC files, so there is more data that needs to be copied from disk/through the memory subsystem than FLAC when replaying. FLAC files trade off this extra noisy movement of data around within components of your computer with additional processing steps within the CPU against L2 cache.

Since the decompression of FLAC is so light on CPU cycles (less than 2%) the reduced memory load on the disk/memory subsystem reduces noise inside your computer, reducing voltage fluctuations/jitter in the output.

Cesare
:)
 
You must log in or register to reply here.


advertisement


Back
Top