Hearing loss buy Naim!

RDS, what are your subjects, that you try to impart, into to the uncultured younger memebers of soceity?, I would like to know sir. Wadia MOnster

 

IMO:

There is very little similarity between the two. an amplified voice, a voice with no amplification sound totally different. an acoustic guitar, an amplified acoustic guitar. very different.

Personally, Im not interested in reproducing a studio sound - because thats chasing your tail... after all, weve already said the amplified sound bares little resemblance to an unamplified sound. What is a 'studio sound' anyway? Does it actually sound good? enjoyable to listen to ? which studio in particular? did the engineers intend you to listen to your recording with the same amplifier and speakers as them, or did they create it with peaks/troughs that would sound better on shop bought kit? If so, which shop bought kit?... who says they were right anyway?

The whole 'studio sound' thing, is bulloney IMO - getting a sound that you think sounds nice, and right to your ears is all there is to it. I suspect the thread wasnt massively replied to - because most people already know this!!!!

All the best
Chris

 

I'm certain my hearing doesn't extend beyond 17KHz, as I can't say for sure that I can hear a 18KKhz tone unless it's fairly loud, and definitely not a 20Khz one at all. Save to say, that the Wad 861 is still the best ingrated player (in stock form) that I've year heard so that Digimaster filter must be doing something right.

Speaking of frequency response, why do companies such as LC Audio with their Patriot amp, and Spectral just to name 2 examples build such an insanely wide bandwith frequency repsonse into their amps i.e. upto and beyond 1Ghz!! Upto 4Ghz for the LC Audio, why? The complete polar opposite philosophy to earlier Naim kit as Julian pointed out in an earlier post.

That explains why MP3 is such a popular music carrier these days with the masses.

I was having a conversation with a (younger than me) colleague at work who can't believe that I can hear hear the difference between compressed and non-compressed audio formats and that in his opinon very soon comparessed audio will be indistiguishable from CD (he is non-audiophile of course). The colleague I was talking to yesterday is also a postgraduate who's head is still very much lost in academia. Believing that theory and theorems are the final arbitors rather believing in his own sensory feedback mechanisms.

Nice post Chris, and totally agreed. I always find electroinc and a lot of studio recordings of electronic music to be a subjective thing, and find as well that the best approach is to ask yourself if you are enjoying the sound (of the music). Purely acoustic based music is different, and perhaps the only type of recorded music which you can place objective measurements and responses against.

 

You will not be able to tell the difference between the original and something like a FLAC or Monkey's Audio file, as these are lossless. So your friend is in fact correct.

 

Lossless in the sense that they do not compress the *.wav files or lossless in the sense that they are, in fact, wav format but with a funny name?
If they do compress and/or alter the format then they are not the same, and do leave open the possibility that the contents of the file has in some way been altered (whether or not this is audible/detectable may be another matter).

 

lossless compression is the technical term for a compression scheme that reduces the ammount of space a 'file' needs when compressed. uncompressing the 'file' results in something that is EXACTLY the same as the origional. pkzip is a good example of this which everyone here has probably encountered at some point or other - where data must be reproduced exactly or the program or data will be gibberish. the main problem with this type of compression is that it's very difficult to keep a constantly streaming bit rate as data is usually compressed via a table where common bytes are encoded as a few bits and less common data as more bits. therefore the bit to byte rate will vary according to how common that data is. the way to deal with this is to either have a large ram buffer and decompress large chunks of the file at once and then clock it out at the playback rate or to develop an algorythm that does not rely on bit table mapping so heavily. there are lossless compression systems out there that should be indistinguishable from wav's however i would think that jitter might be a large issue with some unless they are heavily buffered.
lossy systems in contrast reduce the bitrate according to a psycoaccoustic model by removing frequency ranges supposedly masked by other sounds at that point in time.

correctly implimented lossless compression i have no problem with. incorrectly implimented lossless and lossy compression should only be used for portables and phone ring tones imo.

btw. did you know that both dvda and sacd use 'lossless' compression systems for multichannel music?
cheers

julian

 

Yeah, I'm somewhat familiar with various forms of compression. MPEG - of which MP3 is a flavour - is not usually lossless (I've not yet come across a genuinely lossless flavour), and the truth about codecs and compression is really pretty subtle.
One other thing to remember is that even with lossless compression, you still have to decompress the file - this is normally done "on the fly", but of course that cute phrase also hides a multitude of sins...

 

Paul - I reckoned she was a babe from the start - and for a blonde to do that to me is pretty rare

Hmm - how much hearing damage do you get from an office A/C system and 40 computers running full tilt - it's easily as noisy as a car with the windows down in here?

My ears often block due to my eustachion tubes gumming up - when they pop I can hear really well, when they don't (ie most of the time) my bass response is well curtailed and everything is quieter. However, at least in my left ear, I can hear if a camera flash is on (warmed up, not just warming up) or not, and usually if a monitor or TV is on - that high frequency noise drives me mad. FOr some reason my right ear isn't as good, but then neither is my right eye, so maybe the fact my brain is working harder to make sense of the right eye is limiting the ammount of attention my right ear gets??

Oh yeah - we'll ignore that tinnitus I've had since birth - going to gigs doesn't seem to make it any worse, just more obvious for a day or so. Come on - we need DI for our systems to our brains - speakers, amps and ears get right in the way!

 

joel,
i was speaking in general terms about the 2 types of compression lossy and lossless. (or compression - lossy and compaction - lossless as some would have it these days).

i agree that anything which interferes with the directness of the data is usually a bad thing which is why i think sacd is such a monumental joke. well, that and the huge wadges of ultrasonic noise above 22khz, why not just make a cdp that adds this tat on in the upper frequencies and then voila... sacd from 16 bit. but i digress.

mp3, mpeg, dolby digital, dts, atrac, wma, etc.. are all in the lossy camp.
mlp, huffman, pkzip, rle, sacd are all lossless. i.e. they recover the data exactly from a smaller than origional dataset. what i'm not convinced they recover is the timing. i.e. i think they add jitter, unless they are decomplessed into a buffer and then clocked out at the required rate.

cheers

julian

 

I wasn't thinking about lossless compression algorithms, but good point, my argument was mainly against lossy compression.

The colleague I was having the discussion with also specifically mentioned AC3 (i.e. the compression algorithm used in Dolby Digital) as being a good example of a compressed audio format that stands good comparison with CD.

 

ahh, but what do we mean by "lossless". I am on thin ice ice here, but some of what we generally think of as lossless is not exactly that... For instance Tiff format quite often covers dramaticall alterations to an image - it depends on how the format is implemented in camera. Something that we have learned over the years with digital cameras in particular. miniDV is supposedly lossless and works on the fly - but it takes a lot of resources to do this. Heck, even the format for CD needs to be uncompressed...
It's enough to make me want to go back to vinyl and super16mm film

 

lossless compression should reproduce exactly what was compressed this must work otherwise any program file that was compressed and then uncompressed would crash immediately if even a few bits were wrong. applied to audio though there are issues arround recovering the stream too quickly or slowly hence the new algorythm or buffering necessity.

what do you mean that even the format for cd needs to be uncompressed? at it's simplest red book audio recovers 16 bits off a disk and sends this word straight to a dac. if this is a 16 bit 0x oversampling dac it could even be, i believe, 16 resistors and some transistors but i'm not that electronics savvy so there may be some extra bits.
when you get into deta-sigma and 1 bit dacs with noise shaping or hybrid designs then there are some pretty strange things going on with the data read off the cd. however as far as i know the data itself is represented as 1's and 0's in pits and flats on the disk and this data is raw.

tiff if i remember correctly is a header/ packet based sytem where the data can be compressed or uncompressed at different levels or with different schemes that are controlled by the header information for each packet.

.bmp, .png, .rle etc. are better (quality wise) than jpg, tiff or gif as they are losslessly compressed. they have to be as i used them in my job for texture maps and sprites which were very low resolution (8x8 - 256x256 pixels) so any lossy compression would have been very bad indeed and would have p*ssed off some very militant artists.

minidv i know nothing about so i bow to your wisdom on that one.

cheers

julian

 

Cheers for explaining that better than I probably would Jules. Also some other interesting points raised here which I admit I never really gave much thought before.

Dom, glad to see we have another bloke with taste on this forum other than myself and Neil

 

Your taste is not nearly as good as some of the recent avatar competition entries on Talk audio.

 

Hey I entered that as well. Didnt get any votes. Some people on here would probably be offended if I put pictures of beanflickers on my avatar

 

Yup....

Best thing to do - is forward all proposals to my PM account - and i will check them out personally:MILD: .

I didn't notice your entry - there was some stiff competition though. Worthy winner.

Sorry, forgot what this thread was about now, it seems to have gone off on all tangents - i will stop interfering

 

Ahem! I didn't make any negative comments either. Only thing I added to the thread was my observation that it seemed to be those who were already spoken for who were making the criticisms.

Anyway, back on topic...

 

Not disagreeing with that at all - just pointing out for the record that it's not ruler flat (although the less good sounding 'B' filter is).

Minor correction - you're only a factor of 1000 out! Power bandwidth (-3dB points) 0-4 MegaHz. They also only quote the distortion (<0.01%) over the conventional 2-20kHz. Maybe this bandwidth (and the associated clean impulse response) is just what's required to avoid any convolution (and hence distortion) of the signal coming from the source. The audible spectrum should also be such a small part of the passband that any phase dispersion near the cutoff frequency will be irrelevant? Interesting that the two amps you've heard that have that clarity also have insane bandwidths.

Concurring with Julian here - lossless compression allows bit-perfect decompression, such that a comparison of source and reconstructed data streams are identical. MPEG is not lossless, TIFF is just a means of packaging data, which may be compressed or uncompressed. Introduction of timing errors/jitter is purely a matter of implementation of the decompressor - buffers/FIFOs are always a good idea! Not at all sure I believe that CD data needs to be uncompressed - it's certainly de-interleaved and error corrected, i.e. not read off the disc in a linear fashion like the groove on an LP; maybe that's what you're thinking of. IIRC, since the error correction bits end up being thrown away after they've been used you could actually argue that there is some data-reduction going on (although of ideally redundant information before anyone gets excited about information loss).

I guess it depends who she shared that heavy night with......I wonder what she's like at summarising Proust (Cannot be worse than me BTW)

 

Even though I have that dubious Ford KA model?

That unfortunate incident aside - :beer: We should invent a genetic photocopier so we can make clones of this babe (and others - can someone get me a hair of both Tori Amos' and Alison Hannigan's heads please please please...

She's alive - allllivvvveee!

 

Ultra wide bandwidth, for bats or for appreciable linearity within the audio band?

I wasn't disgreeing either. Just goes to proove that ruler flat frequency responses and ultra clean spectral plots aren't the final arbitor for musically engaging hi-fi equipment.

On the other hand, maybe there is something to having a "gentle" roll off filter rather than a brick wall e.g. DigiMaster, very high fs upsampling, SACD/DSD, etc. More akin to analog response i.e. gradual tailing off rather than blunt cut off, not just talking about frequency response here but also transient responses too.

Lifted from www.spectralaudio.com about their DMA-100S power amp:-

Frequency Response:

±0.1 dB, DC-150 KHz
±1 dB, DC-1 MHz
±3 dB, DC-1.8 MHz

Distortion:
Static:

Less than 0.015% from DC
to 100 KHz,
typically 0.009% @ 100
Watts RMS/8 ohms

Dynamic:
8 Tone Cluster Test 20 KHz
@ 500 Hz separation
0.01% 8 ohms
0.015% 4 ohms

Speed:
Rise Time:

Less than 300 nanoseconds

Settling:
1.5 microseconds to -40dB

Slew Rate:
1,000 volts/microsecond

Strange, I had exactly the same thoughts in my head re the issue on ultra wide bandwidth in amps. Any experts in the field care to offer an explanation? Also, both amps are heavily biased towards class A too. Going on the reviews at Audioreview.com, a Passlabs X series power amp maybe very much up my street too, but have yet to hear one, might be the next best thing to a Patriot V100 but at cheper than Spectral prices maybe?