96kHz test .wavs.

Forum for discussion of narrow-bandwidth mechanical television

Moderators: Dave Moll, Andrew Davie, Steve Anderson

Postby Steve Anderson » Mon Feb 11, 2008 6:47 pm

Telehor wrote:My Octagon replica , working on the 1928 US system with 48 lines and 20 fields is now working perfectly on an Aurora converter with a DVD player at his input. Denis


Denis, sorry for the slow response, it's been a hectic couple of weeks..

A standard consumer device with a playback (record too?) response might exist, as you say sampling at 96kHz or greater. But I'm unaware of one, but I'm sure there's the odd one that does.

I know the soundcard in my main desktop PC will play 96kHz sampled .wav files, as will my cheap laptop, but does it actually have an output response to some 40kHz? As yet I don't know. I just need some time to play around with it/them.

A quick calculation with your parameters shows you need a response to some 35kHz, plus any axtra for 'sync-time'. Much the same as Steve O's (Panrock) 80-line system, but his is at a lower frame/field rate.

..more on this in the main NBTV forum in due course.

Steve A.

P.S. I've just done some quick tests using the above .wav files and indeed my cheap on-board audio device is substantially flat, it plays the 32kHz .wav file at the same amplitude as the 1kHz one. The others also.

However, it could do with a filter on the output as there's quite a few artifacts of the sample rate mixed in. A better 'proper' soundcard one would hope would be an improvement on this. Some of the Creative products look very tempting, with sample rates up to 192kHz.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby AncientBrit » Tue Feb 12, 2008 11:46 pm

Steve, Denis,

I know you are primarily concerned with playback-only frequency response but out of interest I carried out a Record/Playback frequency run with my soundcard running at 96kHz.

The response is not very different from running with a sample rate of 44.1kHz, with a signal cutoff of about 21kHz.

Which is rather disappointing.

Regards,

GL
AncientBrit
Green padded cells are quite homely.
 
Posts: 858
Joined: Mon Mar 26, 2007 10:15 pm
Location: Billericay, UK

Postby Steve Anderson » Wed Feb 13, 2008 1:04 pm

AncientBrit wrote:.....The response is not very different from running with a sample rate of 44.1kHz, with a signal cutoff of about 21kHz. Which is rather disappointing. Regards, GL


...yes it is. I haven't performed that test, something I need to do PDQ. Without a record channel that's up to it, it all becomes somewhat academic.

The .wav files I posted here were generated in software, not actual 'recordings'.

Steve A.
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby AncientBrit » Wed Feb 13, 2008 6:28 pm

>Steve A.

Yes I was hoping that the card would have a native sample rate of 96kHz with everything else being derived from it.

If that were the case then the input filter should roll off around 35-40kHz, but it's not so.

Perhaps the native sample rate is 44.1 or 48kHz.

I seem to remember someone mentioning that before. Was it Gary?

If the analog inputs are not up to it how about getting a signal in through the digital input?

That should bypasss any LPF should it not?

It would mean building a hardware ADC and serialiser though.


Regards,


Graham
AncientBrit
Green padded cells are quite homely.
 
Posts: 858
Joined: Mon Mar 26, 2007 10:15 pm
Location: Billericay, UK

Postby AncientBrit » Thu Feb 14, 2008 12:11 am

Digging around a bit I came across the Texas PCM1802 stereo ADC chip which can cope with 96kbps sampling.

Audio LPF appears to be set at 0.4 of sample rate.

Only problem is the very small package design with pin spacings of 0.65mm

Regards,

Graham
AncientBrit
Green padded cells are quite homely.
 
Posts: 858
Joined: Mon Mar 26, 2007 10:15 pm
Location: Billericay, UK

Postby gary » Thu Feb 14, 2008 1:37 pm

AncientBrit wrote:Steve, Denis,

I know you are primarily concerned with playback-only frequency response but out of interest I carried out a Record/Playback frequency run with my soundcard running at 96kHz.

The response is not very different from running with a sample rate of 44.1kHz, with a signal cutoff of about 21kHz.

Which is rather disappointing.

Regards,

GL


From the "Overview of Audio Codec ‘97"

· fixes the DAC and ADC sample rates and all filtering at 48Kss, hence easier to optimize
(high quality sample rate conversions to/from 48Kss are performed by the digital controller)


However subsequent revisions to the AC'97 specification allow for variable sample rate (44.1, 88.2, and 96 kHz) as an option so you may find a soundcard to do this, however, since these sample rates are register enabled, the driver must also support this.

Every sound card I have tested have exhibited time base errors for sample rates other than 48kHz, so I don't think I have come across such a combination.

In addition, the clock rate of codec is given as 12.288 Mhz, and the 48kHz is derived from this (/ 256). I am not sure how 44100 would be accurately derived (practically) from this which may account for time base error.

I am not sure, from the spec, whether the anti-aliasing filter's cut-off frequency is increased substantially for higher rates as the additional frequencies would be ultrasonic (fc = fs/2). It is more likely that the higher rates are used for the same reason as oversampling, i.e. to improve the effectiveness of the anti-aliasing filter (of course you can get into hot water here from the "golden ears" brigade). :wink:
gary
 

Postby Steve Anderson » Thu Feb 14, 2008 1:48 pm

AncientBrit wrote:Only problem is the very small package design with pin spacings of 0.65mm. Regards, Graham


This is a continual problem nowadays, chip packaging is not 'home construction friendly', but designed for automated pick-and-place machinery as well as reducing board space in commercial designs.

There are adaptors available which convert an SMD device into the more familiar DIL packaging, but usually they cost more than the chip!

My 'real' work involves a lot of prototyping and this is a real headache, a chip manufacturer might say such-and-such chip is available in many different packages including DIL, but when you try to get one....sorry, none in stock.

Steve A.

Anyway, this should really be continued in the main NBTV section, I would move the last few posts there (including mine) if I could work out how to....some Mod. eh?.....worked out how to do it!
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand

Postby AncientBrit » Fri Feb 15, 2008 6:46 pm

>Gary, thanks. That's very informative. I suspected as much as regards the filtering on board.

That's the advantage in using an external chip where the filtering is linked to the sample rate.

I can even envisage a serial version of our CNNC standard where the RGB signal is analogue multiplexed on the way to the left channel of an outboard ADC chip. Multiplexing rate would be linked to the master ADC clock.

Not sure how the sound card would handle this digital signal presented to its electrical/optical SPDIF input. Would it be transparent and assume all filtering had occured in the ADC process?

>Steve. Yes it would be a good ideat to move this topic to the main section

Regards,

Graham
AncientBrit
Green padded cells are quite homely.
 
Posts: 858
Joined: Mon Mar 26, 2007 10:15 pm
Location: Billericay, UK

Postby Steve Anderson » Fri Feb 15, 2008 6:54 pm

AncientBrit wrote:Steve. Yes it would be a good ideat to move this topic to the main section. Regards, Graham


Done already, I'll delete this 'shadow' version (the one in "Images and Sounds") in a few days...

Steve A.

done...
User avatar
Steve Anderson
"Fester! Don't do that to 'Thing'"
 
Posts: 5360
Joined: Fri Mar 30, 2007 10:54 pm
Location: Bangkok, Thailand


Return to Mechanical NBTV

Who is online

Users browsing this forum: No registered users and 38 guests