J
Logged
« Reply #31 on: September 06, 2008, 07:50:10 AM »
| Paul Frindle wrote on Fri, 05 September 2008 13:04 |
| zmix wrote on Mon, 25 August 2008 18:54 |
Hi Paul,
I have listened to all of the audio files linked above and I found that in each case the 'processed' tracks seem to have a oddness in the HF, a slightly 'zingy' quality. Could this be an artifact of the high signal levels causing overs in those frequency bins of the mp3 conversion?
|
Back from hols now and had a listen to the mp3's from the site - and I can hear what you mean immediately.
The odd thing about this is that running up the original PT session with the same settings (as stored) does not exhibit this effect. And the stored wav files prior to mp3 coding also do not have it (obviously).
This therefore seems to be an artifact of the MP3 code/decode cycle - as you suggest. In some ways the original unprocessed file has this also when compared with the original wav file, but to a lesser extent.
Hmm... I don't think the MP3 coding process is likely to 'saturate' as such, but there may be a case for preferential driving of similar bands causing the coder to behave in a particular way? Obviously the denser the harmonic content becomes with compression the more stressed any lossy coding method will inevitably be.
I will think about that some more. In particular, I will see if using conventional compression causes a similar effect.
Thanks for alerting me to it 
|
Thanks for the open minded reply, Paul..! I still wonder about the mechanism at play here... is it possible that the filters used in the mpeg encoding process could be generating 'illegal sample values' and producing internal 'overs' (for lack of a more suitable nomenclature), which cause these audible ringing artifacts?
Logged
« Reply #32 on: September 06, 2008, 09:52:39 AM »
[quote title=zmix wrote on Sat, 06 September 2008 12:50] | Paul Frindle wrote on Fri, 05 September 2008 13:04 |
| zmix wrote on Mon, 25 August 2008 18:54 |
Hi Paul,
I will think about that some more. In particular, I will see if using conventional compression causes a similar effect.
Thanks for alerting me to it
|
Thanks for the open minded reply, Paul..!
I still wonder about the mechanism at play here... is it possible that the filters used in the mpeg encoding process could be generating 'illegal sample values' and producing internal 'overs' (for lack of a more suitable nomenclature), which cause these audible ringing artifacts?
|
No I don't think so to be honest. Last night I recorded and coded pink and white noise both with and without the DSM and these do not exhibit the effect at all. So steady state signal does not produce the effect - and so it seems like 'overloading' can be discounted and the issue appears to be one of dynamic action? The more I listen to this stuff the more I am aware of it on all coded material - whether DSM processed or not. What is more interesting is that polling my teenage girls and their friends today - making sure to carefully equalise perceived levels - they all immediately and unanimously prefer the MP3 coded versions over the wav versions and originals! Of course these are people who these days only listen to MP3s and WMAs - so maybe they are attuned to like these compression artefacts? I will do some more research
Logged
« Reply #33 on: September 06, 2008, 11:01:07 AM »
I should add that all this raises the sticky question of mastering deliberately for MP3 WMA destined listening, which is increasingly becoming the most likely final output people will actually hear - especially the all-important younger audience? There is scope here for 'pre-processing' program that is going to be coded, to ameliorate the coding artefacts we hear afterwards Fascinating stuff....
Logged
« Reply #34 on: September 06, 2008, 08:28:55 PM »
| Paul Frindle wrote on Sat, 06 September 2008 17:01 |
I should add that all this raises the sticky question of mastering deliberately for MP3 WMA destined listening, which is increasingly becoming the most likely final output people will actually hear - especially the all-important younger audience?
|
Better make sure that a well done wav is hitting those encoders so there is something remaining of how sounds sound like, not only something that slightly reminds of something that was there generations before. Best Regards Patrik
Logged
« Reply #35 on: September 07, 2008, 01:47:13 PM »
| Paul Frindle wrote on Sat, 06 September 2008 11:01 |
I should add that all this raises the sticky question of mastering deliberately for MP3 WMA destined listening, which is increasingly becoming the most likely final output people will actually hear - especially the all-important younger audience?
There is scope here for 'pre-processing' program that is going to be coded, to ameliorate the coding artefacts we hear afterwards
Fascinating stuff....
|
hi paul; i think that perhaps simply choosing a different codec to encode mp3s can make a tremendous difference. i've observed that each codec makes completely different distortions. i posted a series of fft analyses here about a year ago. i prefer the original fraunhofer codec; i think perhaps most of the others (many called "fraunhofer" too) are reverse engineered with less awareness of the psychoacoustic theories that make it work. | Paul Frindle wrote on Sat, 06 September 2008 07:13 |
We have never planned to release a PC version on the grounds of security. It seems that even iLok protection gets defeated on PC 
|
i don't use pc platform, but in my experience: that security system isn't completely fail-safe on mac either. however, on mac: it is the user who gets defeated. jeff dinces
Logged
« Reply #36 on: October 27, 2008, 11:44:33 AM »
For info - the Audio Units version of the Dynamic Spectrum Mapper has been released today.
Pro Audio DSP - Professional Audio Plugins
Please note that iLok authorisations apply to both the AU and PT versions, so existing and future customers are able to run either version.
Logged
« Reply #37 on: October 27, 2008, 12:35:03 PM »
Frank Foti has written that Omina needed to completely rethink how the mid-range bands of their broadcast processors work in order to minimize artifacts in both streaming and digital broadcasting.
Logged
« Reply #38 on: October 27, 2008, 12:39:42 PM »
| Paul Frindle wrote on Sat, 06 September 2008 07:13 |
J
Logged
« Reply #39 on: October 27, 2008, 08:59:01 PM »
| Quote: |
Also for ProTools the great majority of our potential clients are running on Mac..
|
| Quote: |
With all due respect, I don't think platform exclusivity is a good form of piracy protection. Granted if you don't think you'll have very many PC customers (given you are not even making a VST version) I don't think it's an issue, however, I think you'll find for every 10 people who steal the software 9.9 of them would have never bought it anyway. Making a PC VST version opens you up to a very large client base.
Noah
|
Yes - this is something I am aware of.
Logged
« Reply #40 on: October 28, 2008, 06:31:00 AM »
| Paul Frindle wrote on Tue, 28 October 2008 02:59 |
Yes - this is something I am aware of.
|
Why not then opt for Syncrosoft? It has never been cracked as far as I know and it is a very solid system for the end user as well. Cheers! Niklas
Logged
Niklas Silen @ Good Will Studios, Helsinki, Finland
« Reply #41 on: November 29, 2008, 04:10:31 PM »
Just for info, I have done an introduction to the Dynamic Spectrum mapper: Apologies for the sound quality of the upload video - but hopefully it still convey a message of what it's all about..
Logged
« Reply #42 on: November 29, 2008, 04:49:33 PM »
hi paul;
imo, that is an informative video demo.
dsm seems like a superb product.
jeff
Logged
« Reply #43 on: December 17, 2008, 12:19:26 PM »
Just for info I have done another video explaining the basics of de-essing and vocal control with the DSM.. It suffers a bit from video coding sound quality loss - but hopefully still gets the message across..
Logged
« Reply #44 on: December 17, 2008, 09:58:35 PM »
| Noah Mintz wrote on Mon, 27 October 2008 12:39 |
I think you'll find for every 10 people who steal the software 9.9 of them would have never bought it anyway. Making a PC VST version opens you up to a very large client base.
|
Somehow, I fail to find comfort in that...
Logged
|
|
|
