Hi-res formats, impressions?

It is awesome....

As good as 1080P versus standard def for television.

I have some 24bit/96kHz material from Linn Records - Claire Martin, Fiona MacKenzie and a little Mozart. Fantastic - if only all music was available like this!

You can find two samples here, 96k/24 bit.

The recording was made a month ago in the workshop where a Steinway piano restoration (originally made in 1904) is taking place. It is in the basement, the side wall was quite close, so the acoustics is not first class. We used one of the very best stereo ribbon microphone ever made, the Royer SF-24, feeding a Korg MR-1000 hard diskrecorder in 128x DSD mode (double bandwidth then the standard SACD) directly, then converted to 96k/24 bit. As pure as it can be, no processing, level adjustment, compression, limiting were used and the recorder was running from a battery.

Sample 1
Sample 2

Right click on the file and "save as"



You'll find it's a .wav file.

quote:

Originally posted by ferenc:
We used one of the very best stereo ribbon microphone ever made, the Royer SF-24,

Its fabulous, isnt it? I adore mine. Most natural sounding microphone I have ever used.

I've listened to some 24/96 files ripped from the mastertape at home recently. ( via HDX )

They were good recordings but since I was not familiar with the material at all, had nothing to compare against them.

My impression was the same as my past 24/96 dealings where they have a great ambient and detail comparatively.

The problem remains the availability of wider choice in music.

quote:

Originally posted by kuma:
The problem remains the availability of wider choice in music.

As ever. That is why I like mp3 so much, you can get absolutely every thing in that format, except music. C'est la vie.

But the classical pieces at the end are pretty great, no? I suppose that is Iona Brown conducting. I have really enjoyed listening to that one a few times. there is no metadata on the files save song title.

Well...

Some disco would have been nice.

quote:

Originally posted by jon honeyball:

quote:

Originally posted by ferenc:
We used one of the very best stereo ribbon microphone ever made, the Royer SF-24,

Its fabulous, isnt it? I adore mine. Most natural sounding microphone I have ever used.

Yes it is. We have a tabla recording as well, which is very good, there is a chance that ECM will make a recording with the artist sometime soon using our Royer microphones. I hope I can upload to our ftp site a short sample of this tabla recording soon too.

If anybody is interested in the comparison between the 96k and 44.1k files there are 44.1k files on the ftp site too.

Sample 1 - 44.1k

Sample 2 - 44.1k

Converted from the same 128x DSD original recording

These should be like chalk versus cheese on a good DAC -- those with Lavry etc should try them and report back.

quote:

Originally posted by ferenc:
You can find two samples here, 96k/24 bit.

hello ferenc,

could you please let me know what music this is (composer, title,...)?

thanks a lot,
michael

I have been fortunate to have been able to hear a number of tracks/albums with 24 bit playback from master recordings and compared the same to 16 bit ripped and 16 bit playback via the HDX.

The 24 bit has much more detail, especially with subtile nuances which surprise you after hearing the original 16 bit. I think Kuma summed it up nicely.

Wav files: 24>>16 ripped>16

The extra information was also readily apparent with the DA-10 for Lavry users

quote:

Originally posted by michael1702:

quote:

Originally posted by ferenc:
You can find two samples here, 96k/24 bit.

hello ferenc,

could you please let me know what music this is (composer, title,...)?

thanks a lot,
michael

Hi Michael,

it is an improvization by the gentleman, who is doing the restoration. He was a keyboardist in a famous underground Hungarian progressive rock band from the late 60's. Their name was "Kex".

ferenc,

Thanks for posting those samples. The 24/96 clearly sounds "less like CD" than the 16/44.

I am currently using a macbook, and I have to manually switch the output sample rate with the Audio/Midi utility. Does the Konnect8 switch automatically to match the file/source? Or does it also require manual change of sample rate?

Nice piano, by the way, and a nice job of recording.

Thanks!

Charlie

It will switch but with a short lag. You'll have to start the file so it senses the format. It will switch and then you'll need to restart the file for sound. Takes a few seconds when it's set up correctly. You may still have to reset the format in Itunes first so it will not up or down sample. I don't know for certain as I prefer not to use it or the WM player.

js,

Thanks. iTunes lists the bit rate at 4608, and I made no change to iTunes settings for the 24/96 files. The DAC indicates the sample rate as 96 or 44 respectively, according to how the Audio/Midi utility is set - NOT according to the data in the file.

What player software do you prefer?


Charlie

quote:

Originally posted by CharlieP:
js,

Thanks. iTunes lists the bit rate at 4608, and I made no change to iTunes settings for the 24/96 files. The DAC indicates the sample rate as 96 or 44 respectively, according to how the Audio/Midi utility is set - NOT according to the data in the file.

What player software do you prefer?


Charlie

For PC, Wavelab which isn't a media player/library and Media Jukebox which is a free ASIO capable standard type music player. I'm not sure for a MAC as these programs aren't available. While stock MACs sound better than stock PCs, there doesn't seem to be an after market media player like Media Jukebox or Foobar (with an ASIO plugin) that is able to bypass the internal mixer or ITunes player which is what's needed to be equivalent. At least not that I'm aware of. Media Jukebox is a scaled down player of what can be purchased for more versatility in play lists, video decoding etc. There are pro sound programs that can be used in MACs with ASIO but less convenient.

I uploadad a digitized LP track to an ftp site in 44k, 48k, 88k, 96k, 176k, 192k /16 and 24 bit (converted from the 128x DSD file) versions, plus the original 128x DSD file as well. The digitized track is a good sounding typical early stereo jazz recording from 1957/1958.

There was only one step of processing: the original recording level was decreased by 1.2 dB in the original DSD file before the conversion to PCM just to avoid clipping, being on the safe side. Surprisingly some of the peaks of the 176k and 196k files are 0.1 dB less than the rest of the PCM files, probably the nature of the conversion algorithm, but I think it will not have any influence on the listening impression.

The original recording was made using a Korg MR-1000 pro hard disk recorder and some DIY cable. I do not want to give details of the LP playing system, but it is well known and quite expensive system: English record player and arm (not Linn!), Japanese cartridge and a custom made battery powered solid state MC phono amp.

You can find the files here:

http://www.penna-media.hu/hires_jazz_lp_sampler/

Thanks Ferenc!

It is great to have these stuff to compare,
I have demoed 96/24 against standard resolution and I agree there is a significative difference.

Now I have just bought a Lynx AES16 to replace my M-audio sound card, I hope this will be another step ahead.

ferenc

Thank you. Downloading now. Will try and listen today but it may have to wait until the weekend. Will report back

In my experience of listening to 24/96 for a couple years, as well as listening to these samples graciously supplied by ferenc, bit depth makes all the difference.

24 bit outperforms every time, although the sample rate seems to not matter to my ears.

I have some 24/48 LP transfers (some of the very first I did), which I later replaced with 24/96s. I could not hear a difference between the two. Same goes for some 24/192s I have, albeit I dont have anything to A/B them with.

I have always thought the law of diminishing returns applies here.

Along that line of thought, I am also wondered about the benefits of 32 bit and so on... Is there a limit to how good it can be? Is there a point at which the bit depth is too great and is no longer an improvement?

-2cent p

AFAIK bit depth determines dynamic range, 24 bits goes up to 144db vs. 96db in standard 16 bits. Going further seems too much for human ear and besides systems' noise floor wouldn't let you hear these differences but who knows, after all, recordings usually compress dynamics so deeper bit depth could still result in an audible benefit.

Respect to sample rate, I agree bit depth is more noticeable and there is a logical reason for that. In theory, higher sample rate determines the bandwidth frequency range. For instance, standard 44,1 results in a maximum of 22khz, which actually surpasses the human treshold. Higher sample rates mean your ears are receiving higher frequency sounds which you are not able to perceive. Nonetheless, it is said that bigger frequency bandwidth make digital music sound more analog.

Goldfinch, I'm not sure I quite understand your logic regarding bit-depth, but it does raise an idea. You rightly state that higher bit-depth allows greater dynamic range, but that this is perhaps not necessary due to the system, environmental and biological limits already being well-served by a 96db range of 16-bit systems. Compressed recordings use even less of the full 96db, so higher bit-rates on these doesn't seem useful at first glance.

Nevertheless, a recording compressed, but still recorded in 24-bit to preserve the "resolution" could perhaps be de-compressed to provide increased dynamic range at a quality that is no worse than a 16-bit system. It's maybe a way of combating the "Loudness wars". It's kind of like mucking around in "levels" or "curves" in photoshop (or just using the contrast slider). If you first convert your image to 16-bit (rather than the default 8-bit of jpeg) you can compress and more importantly decompress the contrast (dynamic range) but create far fewer quality degrading artifacts like banding than if you were working at a lower bit depth.

Software (and even hardware) can be configured to decompress sound (make soft sounds softer and loud sounds louder). The trick is to do it without degrading the sound and without making it sound "fake". Starting with high bit-rate data presumably makes it easier.

The RIAA EQ standard on vinyl is actually sort of an analogue analogue of this idea.

Recording engineers might object to hi-fi manufacturers and users having this sort of control, but I suspect that many components don't have strictly linear gain curves anyway, so the horse may have already bolted one way or the other. In any case, if the quality of a lot of modern rock/pop recordings is anything to go by, they have no-one but themselves to blame!

It's all probably moot, as those who would release 24-bit music are not those who would compress the life out of it in the first place.

ferenc

Thanks again for making all those files available

I find the differences more obvious on the live piano recording which sounds clanky from the 44.1 files in comparison to the 96k

In some ways the same sort of differences that someone demonstrated to me many years ago when he removed a good proportion from my then Class A power amps. (Something only the very brave or very foolish would do unless they have access to free parts and service.)

I found the difference between 44.1 and 88,2 both 16bit to be greater than the difference between 88.2 16bit to 88.2 24Bit.

Finally and I know I'm being very cheeky. Any chance of doing the same think but with a vinyl rip of a classic rock album?

Somehow I just find it so much easier to spend all the time listening to changes when I like the music.

I won't be in the slightest bit offended btw if you reply saying that you are too busy.

Thanks again a really useful resource to have.

Finally, do you mind if I use them to demo the differences to my industry colleagues?