Hi,
is it just my setup or does Pro7 actually have a far lower volume than all other channels? I'm using Werner's 261d firmware with AC3overDVB from a FF 1.3 card over an optical SPDIF link into a home cinema system.
While we're at that, has anyone been working on a limiter to keep the advertising blocks' volumes down? As most of you will know, the ad blocks have an average level 3 dB above the average level in the rest of the material. This is very annoying particularly after a low volume scene in a film. ... :-( Noad's algos could be helpful here.
Harald Milz wrote:
Hi,
is it just my setup or does Pro7 actually have a far lower volume than all other channels? I'm using Werner's 261d firmware with AC3overDVB from a FF 1.3 card over an optical SPDIF link into a home cinema system.
ProSieben, Sat.1, ZDF and mybe others which also broadcast and AC3 audio have lower volume on AC3, even on domestic receivers. As it is AC3 over SPDIF, I don't think its volume can be adjusted elswhere than in your decoder amp. Maybe all other channels broadcasting MP2 should be adjusted lower.
While we're at that, has anyone been working on a limiter to keep the advertising blocks' volumes down? As most of you will know, the ad blocks have an average level 3 dB above the average level in the rest of the material. This is very annoying particularly after a low volume scene in a film. ... :-( Noad's algos could be helpful here.
On AC3 I think it's not possible, see above...
Just my 2 cents, Lucian
On Thu, Jun 09, 2005 at 08:25:36AM +0200, Harald Milz wrote:
Hi,
is it just my setup or does Pro7 actually have a far lower volume than all other channels? I'm using Werner's 261d firmware with AC3overDVB from a FF 1.3 card over an optical SPDIF link into a home cinema system.
Most AC3 channels do have a lower volume due to the fact that the dynanic range of AC3 is much more than 100dB. This is much more than you can do with Mpeg Audio and raw PCM at 16bit[1].
OK, in most cases at the most time for DD 2.0 this is not relevant, but for DD 5.1 you could lose the friendship of your neighbours on an explosion within a movie. Nevertheless, even for DD 2.0 the possible upper loudness is much higher than for Mpeg Audio or raw 16 bit PCM.
Werner
[1] PCM 44.1kHz/16bit includes a dynamic range of 96dB, but most mass markt audio CDs do not use them and any enhancement for a larger dynamic range and super bits is only marketing, which violates the shannon theorem.
Dr. Werner Fink schrieb:
Most AC3 channels do have a lower volume due to the fact that the dynanic range of AC3 is much more than 100dB. This is much more than you can do with Mpeg Audio and raw PCM at 16bit[1].
You know that you can feed floating-points into mpa and get the same out and there is also an extension for mc, so in theory you can get nearly the same dynamics with mpa as with ac3, but usually tracks are not mixed this way and as nearly no receiver handles mpa, it gets decoded to PCM anyway, so practically you are right.
On Thu, Jun 09, 2005 at 11:57:19AM +0200, Prakash Punnoor wrote:
Dr. Werner Fink schrieb:
Most AC3 channels do have a lower volume due to the fact that the dynanic range of AC3 is much more than 100dB. This is much more than you can do with Mpeg Audio and raw PCM at 16bit[1].
You know that you can feed floating-points into mpa and get the same out and there is also an extension for mc, so in theory you can get nearly the same dynamics with mpa as with ac3, but usually tracks are not mixed this way and as nearly no receiver handles mpa, it gets decoded to PCM anyway, so practically you are right.
OK, usually I do not feed my AV amplifier with mpa even if it can decode this. This because it can not detect this automatically as ti does it with PCM/AC3/DTS .
Do you have a mpa sample which demostrate the possible dynamic range? With the help of the bitsteamout plugin it is possible to forward the mpa to an AV amplifier/receiver.
Werner
Dr. Werner Fink schrieb:
On Thu, Jun 09, 2005 at 11:57:19AM +0200, Prakash Punnoor wrote: Do you have a mpa sample which demostrate the possible dynamic range? With the help of the bitsteamout plugin it is possible to forward the mpa to an AV amplifier/receiver.
Unfortunately no. But you could try hacking toolame or alike to eat floating point data (or can it already do it?) and transcode eg front channels from a 5.1 ac3 to it. Just make sure the ac3 is decoded to floating points, as well.
Should be a quick task for you. ;-)
Dr. Werner Fink werner@suse.de wrote:
Most AC3 channels do have a lower volume due to the fact that the dynanic range of AC3 is much more than 100dB. This is much more than you can do with Mpeg Audio and raw PCM at 16bit[1].
Hmmm - I'm missing the logic here. If the average level of an AC3 channel is below the other (PCM) channels, I turn the receiver up anyway. The next explosion - well, you get the idea.
hm@seneca.muc.de(Harald Milz) 10.06.05 22:24
Once upon a time "Harald Milz " shaped the electrons to say...
Dr. Werner Fink werner@suse.de wrote:
Most AC3 channels do have a lower volume due to the fact that the dynanic range of AC3 is much more than 100dB. This is much more than you can do with Mpeg Audio and raw PCM at 16bit[1].
Hmmm - I'm missing the logic here. If the average level of an AC3 channel is below the other (PCM) channels, I turn the receiver up anyway. The next explosion - well, you get the idea.
The trick must be not to increase the volume of ac3 but to reduce the volume of other channels.
Or -better- to insert a dynamic volume control (AVL audio/automatic volume limiter) because very few people will like to have 140dB(A) comimg from the TV... (if minimum environmetal "base" noise is assumed to be 40dB(A)), that would avoid the distortion you will of course get at the next explosion.
What's the sense of 100dB dyn. range? 1. Marketing, bigger number looks better (see PPMO) 2. Marketing, you have to buy better Amplifiers (I assume 4000W RMS output would be sufficient) 3. Marketing, bigger amp requires bigger loudspeaker 4. Marketing, bigger amp reuires bigger rooms 5. Marketing, you have to buy preamp that would be able to limit the peeks to protect your ears and increase the minimum above the environmental noises.
We already had this disussion with "CD" and the "giant" 90dB they claimed to give. But you will hardly find any popular music which has not a reduced dynamic range. If it is 65dB it'll be good and better for your ears and equipment.
Rainer
Rainer Zocholl schrieb:
hm@seneca.muc.de(Harald Milz) 10.06.05 22:24
Once upon a time "Harald Milz " shaped the electrons to say...
Dr. Werner Fink werner@suse.de wrote:
Most AC3 channels do have a lower volume due to the fact that the dynanic range of AC3 is much more than 100dB. This is much more than you can do with Mpeg Audio and raw PCM at 16bit[1].
Hmmm - I'm missing the logic here. If the average level of an AC3 channel is below the other (PCM) channels, I turn the receiver up anyway. The next explosion - well, you get the idea.
The trick must be not to increase the volume of ac3 but to reduce the volume of other channels.
Nope, in fact the correct way to do it, would be if a *correct* value for dialog normalization in the ac3 header would be sent and then the receiver amplifies the whole track relative to this and the volume setting you chose so that it matches "non-ac3" streams. Unfortunately no broadcaster really does this. Otherwise eg commercial breaks in movies with 5.1 wouldn't be this horrible load. So you have to complain to the broadcasting station to learn to transmit correctly filled ac3 headers...
If dynamics are too much for you, simply enable DRC, which every receiver should be able to handle. Then your ac3 5.1 sounds like every other compressed source, if you select heavy DRC. (Yes, every standard 2ch downmix is heavily compressed.)
I like to have dynamics. I agree that even 100db would probably enough for giving that, but I don't mind that at least 5.1 ac3 does actually provide dynamics in sound.
- Marketing, bigger number looks better (see PPMO)
I guess you are refering to PMPO? ;-)
We already had this disussion with "CD" and the "giant" 90dB they claimed to give.
IIRC, CD gives something like 114db with dithering...and 96db without.
(http://homerecording.de/modules/news/article.php?item_id=389) (http://www.digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page...)
On Sat, Jun 11, 2005 at 12:21:26PM +0200, Prakash Punnoor wrote:
We already had this disussion with "CD" and the "giant" 90dB they claimed to give.
IIRC, CD gives something like 114db with dithering...and 96db without.
Informations which are not recorded can not be recreated without guesssing. Therefore 16 bits are 96db and not more.
(http://homerecording.de/modules/news/article.php?item_id=389) (http://www.digido.com/portal/pmodule_id=11/pmdmode=fullscreen/pageadder_page...)
Werner
Dr. Werner Fink schrieb:
On Sat, Jun 11, 2005 at 12:21:26PM +0200, Prakash Punnoor wrote:
We already had this disussion with "CD" and the "giant" 90dB they claimed to give.
IIRC, CD gives something like 114db with dithering...and 96db without.
Informations which are not recorded can not be recreated without guesssing. Therefore 16 bits are 96db and not more.
Nope, you obviously don't know how dithering works. Every heard of DSD? So you think Sony's SACD doesn't work? DSD is pure dithering...
On Donnerstag 09 Juni 2005 08:25, Harald Milz wrote:
is it just my setup or does Pro7 actually have a far lower volume than all other channels? I'm using Werner's 261d firmware with AC3overDVB from a FF 1.3 card over an optical SPDIF link into a home cinema system.
my A/V receiver (Denon AVR 2805) lets me adjust the volume for each speaker separately for every type of input signal. If I increase the level for AC3 input by 6dB, it has about the same level as normal channels.
prakashp@arcor.de(Prakash Punnoor) 11.06.05 12:21
Rainer Zocholl schrieb:
hm@seneca.muc.de(Harald Milz) 10.06.05 22:24
Dr. Werner Fink werner@suse.de wrote:
Most AC3 channels do have a lower volume due to the fact that the dynanic range of AC3 is much more than 100dB. This is much more than you can do with Mpeg Audio and raw PCM at 16bit[1].
Hmmm - I'm missing the logic here. If the average level of an AC3 channel is below the other (PCM) channels, I turn the receiver up anyway. The next explosion - well, you get the idea.
The trick must be not to increase the volume of ac3 but to reduce the volume of other channels.
Nope, in fact the correct way to do it, would be if a *correct* value for dialog normalization in the ac3 header would be sent and then the receiver amplifies the whole track relative to this and the volume setting you chose so that it matches "non-ac3" streams. Unfortunately no broadcaster really does this. Otherwise eg commercial breaks in movies with 5.1 wouldn't be this horrible load. So you have to complain to the broadcasting station to learn to transmit correctly filled ac3 headers...
Or is it done by (stupid) intention to "emphesis" the commerical break, to wakeup the already sleeping viewers?
IIRC that are simple SCART-Plugs which reduces the volume during commericial breaks. OTOH: As the commecial are not always ac3(are they at all?), the "dialog normalization value" must be changed on evey break. Add-dropping software would be lucky to have such an exact signal to detect "commericals" ends and beginns!
If dynamics are too much for you, simply enable DRC, which every receiver should be able to handle. Then your ac3 5.1 sounds like every other compressed source, if you select heavy DRC. (Yes, every standard 2ch downmix is heavily compressed.)
A, i see thanks.
I like to have dynamics. I agree that even 100db would probably enough for giving that, but I don't mind that at least 5.1 ac3 does actually provide dynamics in sound.
- Marketing, bigger number looks better (see PPMO)
I guess you are refering to PMPO? ;-)
Yes, what else? ;)
We already had this disussion with "CD" and the "giant" 90dB they claimed to give.
IIRC, CD gives something like 114db with dithering...and 96db without.
The absoulte Value does not matter. The point was that even 90dB would hard to use in normal home environment, or are you able to turn your heart beat (reversibel) off to reduce the ground noise? (At aprox. 1kHz a healthy ear (in the youth) is almost able to hear the noise the blood flow causes and the brown's movements of air molecules..) I don't think that any real hifi enthuiast will kill his ears with sound levels above 100dB(A), but if you already have 20dB in your -very- silent listning room, so you have to gerate 120DB(A) in the peaks to get 100dB range! I know that most musicians(incl. classics "unpluged"!) to uses "In-Ear-Monitors" Not only because that works MUCH better (allowing Stage InterCom etc.), too their ears will be protected. Sorry, i don't think that you will have fun with such a dynamic range in the long term. Be cautious! (Have a look for "Tinitus" etc. in the Web).
Rainer---<=====> Vertraulich // // <=====>--------------ocholl, Kiel, Germany ------------
Rainer Zocholl schrieb:
prakashp@arcor.de(Prakash Punnoor) 11.06.05 12:21 Or is it done by (stupid) intention to "emphesis" the commerical break, to wakeup the already sleeping viewers?
Hmm, well that they already do in 2ch formats (compress more to get it louder). In ac3 case I rather think it is ignorance.
As the commecial are not always ac3(are they at all?), the
In case of Pro7, Sat.1, ZDF you have one seperate mpa and ac3 track and on the ac3 track you only have ac3 transmitted - either in 5.1 or 2.0.
"dialog normalization value" must be changed on evey break.
In a perfect worls, yes. :-) But usually ads are eual loud, so that is not the problem.
Add-dropping software would be lucky to have such an exact signal to detect "commericals" ends and beginns!
You already have if you watch a 5.1 movie: Commercials on ac3 track are to 99% 2.0 (except one Yamaha ad for av receivers, which is 5.1 - guess why).
in your -very- silent listning room, so you have to gerate 120DB(A) in the peaks to get 100dB range!
Yes, true. but I think loud low frequs are not that bad (except for the neighbors ;-) than eg 1kHz at 120db... I never measured the loudness I listen too, but I am not listening movies at soft volumes. ;-)
dynamic range in the long term. Be cautious! (Have a look for "Tinitus" etc. in the Web).
Oh, well, I already have but because of stress...(it comes and goes...)
Wolfgang Rohdewald wolfgang@rohdewald.de wrote:
my A/V receiver (Denon AVR 2805) lets me adjust the volume for each speaker separately for every type of input signal. If I increase the level for AC3 input by 6dB, it has about the same level as normal channels.
Not bad - unfortunately, mine doesn't.
Rainer Zocholl UseNet-Posting-Nospam-74308-@zocki.toppoint.de wrote:
Or is it done by (stupid) intention to "emphesis" the commerical break, to wakeup the already sleeping viewers?
Honni soit ... but I am afraid this is exactly the point.
As the commecial are not always ac3(are they at all?), the
Well thtey could start to use AC3 if they wanted so this could easily be circumvented.
Add-dropping software would be lucky to have such an exact signal to detect "commericals" ends and beginns!
Do we have any chance to evaluate this value within VDR, or is the AC3 stream sent to the SPDIF out directly within the AV7110 chip? Werner, _you_ are the firmware guru, what do you think?
The absoulte Value does not matter. The point was that even 90dB would hard to use in normal home environment, or are you able to turn your heart beat (reversibel) off to reduce the ground noise?
Or stop breathing (~ 10 dBA) ;-)
Sorry, i don't think that you will have fun with such a dynamic range in the long term. Be cautious! (Have a look for "Tinitus" etc. in the Web).
Nothing to add here. (good grief).
Rainer Zocholl UseNet-Posting-Nospam-74308-@zocki.toppoint.de wrote:
Or -better- to insert a dynamic volume control (AVL audio/automatic volume limiter) because very few people will like to have 140dB(A) comimg from the TV... (if minimum environmetal "base" noise is assumed to be 40dB(A)),
<offtopic>
Hehe, I'd like to see a TV set that is _capable_ of delivering 140 dB (at 1m or so). The loudest bands in the 70s (Deep Purple and Grateful Dead) had like 120 dB on stage, with umpteen 100W guitar amps running in parallel. The surviving Deads are nearly deaf today. As for Ritchie, I don't know. Today, the musicians are smarter, and use in-ear monitors or dampeners.
What's the sense of 100dB dyn. range?
Well why are soundcard makers advertising 24/96 or even 24/192? Same thing I guess.
- Marketing, you have to buy better Amplifiers (I assume 4000W RMS output would be sufficient)
Depending on the loudspeaker efficiency. Let's assume an average speaker for home use that delivers about 85 dB @ 1W @ 1m. To get 140 dB from this loudspeaker (within the microsecond before it turns into a cloud of grey smoke), you would have to feed it 10^((140-85)/10) W ~ 316 kW. A loudspeaker delivering 140 dB from 4.4kW would have to have an efficiency of 140 dB - 10*log 4400 ~ 103 dB @ 1W @ 1m, which is in the range of the best performing instrument speakers (Electro Voice or Celestion). Their power handling is more in the 100..200W range at more than 5-10% distortion, so you would want to have an array of at least 40 of them, powered by 44 100 W Marshall or Boogie heads - in the living room - well ... I do have an old Acoustic G100T guitar amplifier (100W tubes feeding an EVM 12L speaker) in my living room, and if I turned it up half-way I would be arrested by the police within 5 minutes (especially considering my guitar playing skills but this is a different story).
You get the idea... ;-)
Alas, the point is not the high amplifier power. The dynamic range is defined as 10 * log (highest level at a certain distortion level / lowest level above the inherent noise of the all circuitries together) (http://en.wikipedia.org/wiki/Dynamic_range, http://www.absoluteastronomy.com/encyclopedia/s/si/signal-to-noise_ratio.htm). For average room level music, 100 mW is enough (75 dB for the above example speaker @ 1m), 1W is plenty, 10W wakes up your neighbour and affects your hearing, and 100W starts to hurt. 4000W allows you to run rock concerts in mid-size halls for a couple of hundred people, and requires a 10 kW 3-phase feed which you don't usually have at home.
So for all practical purposes, we're talking about a usable range of 10W and below. Very few analog power amplifiers have a residual noise in the µW range, so ... it's all very theoretical. No hi-fi op amp I know of has a usable dynamic range of 140 dB (but I didn't check the data sheets lately), and the PCB layout for such a circuitry is not far from rocket science.
We already had this disussion with "CD" and the "giant" 90dB they claimed to give.
Which is perfectly fine for most non-pro listeners. Well, back then the record companies were about as flexible as they are today, and would rather have kept selling vinyl LPs with their max. 65 dB dynamic range.
But you will hardly find any popular music which has not a reduced dynamic range. If it is 65dB it'll be good and better for your ears and equipment.
Most pop music has been mixed and compressed at far less dynamic range. In many cases they are in the 30 dB ballpark, to keep things loud enough when listening to car stereos or your average MP3 player in a subway train. And radio stations tend to use additional limiters before the power stages...
But we're talking about AC3 audio, anyway, which is a different thing... 24 bits * 6 db/bit sounds too good, doesn't it... IMHO 24 bits are fine for the recording and mixing stage to keep rounding errors and quantization noises low but for listening, 16 bits is fine, for all practical purposes.
(And while we're at it, most teenagers who go to the disco frequently do not have an actual dynamic range of 140 dB any more when they turn 20, especially not above 5 or 6 kHz. For them, the standard 44.1 kHz CD sample rate is an overkill, and for the kind of music they often listen to, 16 bits is overkill as well, which is why very many people can't tell the difference between a clean CD recording and its 128 kbit/s MP3 equivalent ... ).
So yes, it's marketing to a "certain" extent.
</offtopic>
Back to VDR - I suspect there's no easy way to volume adjust the AC3 stream coming from Pro7 and others -- sh*t, this would require decoding AC3 which isn't allowed without a Dolby (R) license, right? BTW ZDF does not have this low volume problem compared to other (non-AC3) channels, only Pro7. But then, ZDF broadcasts less explosions than Pro7, and this is unlikely to change if the Springer publishers (Bild-Zeitung) actually take over Pro7/Sat1.
Hi,
I've got a problem with the auto sender "find/update" functionality of VDR. Though a sender is found, I am not able to watch it because the Audio/Video PIDs don't get recognized.
Example of channel entry in channels.conf: Multisignal:362000:C0M64:C:6900:0:0:0:0:208:133:4:0
even leaving it tuned to the channel for some time doesn't fix the problem. Currently I'm using VDR 1.3.25 with Kernel 2.6.10-rc2 and also tested it on Kernel 2.6.12-rc6. My DVB-Card is a Siemens DVB-C. Any idea on how to solve this ?!?
Thanks
Tom
Prakash Punnoor prakashp@arcor.de wrote:
dynamic range in the long term. Be cautious! (Have a look for "Tinitus" etc. in the Web).
Oh, well, I already have but because of stress...(it comes and goes...)
AFAIK more than 15% of the population have a more or less severe Tinnitus. Mine is a very high-pitched permanent noise which is definitely well above 8 kHz (the highest frequency my doctor's listening test equipment would check, and no notch here - I would have suspected a certain masking effect but fortunately, nothing - fingers crossed). Does anyone know about a good ALSA compatible tone generator which I could use to find the peaks? My sound card (well, a 24/96 one, ummm ...) and my Sennheiser 580 phones should be able to provide a near-flat frequency response up to about 20 kHz...
Tomahawk wrote:
Hi,
I've got a problem with the auto sender "find/update" functionality of VDR. Though a sender is found, I am not able to watch it because the Audio/Video PIDs don't get recognized.
Example of channel entry in channels.conf: Multisignal:362000:C0M64:C:6900:0:0:0:0:208:133:4:0
even leaving it tuned to the channel for some time doesn't fix the problem. Currently I'm using VDR 1.3.25 with Kernel 2.6.10-rc2 and also tested it on Kernel 2.6.12-rc6. My DVB-Card is a Siemens DVB-C. Any idea on how to solve this ?!?
Are you sure that this channel is broadcasting at the time you test it?
Klaus
Yes when using the dbox2 as recveiver I can see the channel is broadcasting.
The problem appeared to me in connection with recent changes in transponder Layout by German "Premiere" Broadcaster. Unfortuantely I'm not able to copy the channel data to VDR because all problem channels are "optional" Channels and not saved in the Bouqet.
Greets
Tom
Klaus Schmidinger schrieb:
Tomahawk wrote:
Hi,
I've got a problem with the auto sender "find/update" functionality of VDR. Though a sender is found, I am not able to watch it because the Audio/Video PIDs don't get recognized.
Example of channel entry in channels.conf: Multisignal:362000:C0M64:C:6900:0:0:0:0:208:133:4:0
even leaving it tuned to the channel for some time doesn't fix the problem. Currently I'm using VDR 1.3.25 with Kernel 2.6.10-rc2 and also tested it on Kernel 2.6.12-rc6. My DVB-Card is a Siemens DVB-C. Any idea on how to solve this ?!?
Are you sure that this channel is broadcasting at the time you test it?
Klaus
vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
Tomahawk wrote:
Yes when using the dbox2 as recveiver I can see the channel is broadcasting.
The problem appeared to me in connection with recent changes in transponder Layout by German "Premiere" Broadcaster. Unfortuantely I'm not able to copy the channel data to VDR because all problem channels are "optional" Channels and not saved in the Bouqet.
Greets
Tom
Klaus Schmidinger schrieb:
Tomahawk wrote:
Hi,
I've got a problem with the auto sender "find/update" functionality of VDR. Though a sender is found, I am not able to watch it because the Audio/Video PIDs don't get recognized.
Example of channel entry in channels.conf: Multisignal:362000:C0M64:C:6900:0:0:0:0:208:133:4:0
even leaving it tuned to the channel for some time doesn't fix the problem. Currently I'm using VDR 1.3.25 with Kernel 2.6.10-rc2 and also tested it on Kernel 2.6.12-rc6. My DVB-Card is a Siemens DVB-C. Any idea on how to solve this ?!?
Are you sure that this channel is broadcasting at the time you test it?
The problem is that the channel has the wrong frequency. IMHO it has the same frequency as the "Master" channel which is not necessarily the case. Editing the channel manually and trying all freqencies (there aren't many for Premiere and cable) is the only help I found. Once you have the correct frequency, VDR fills in the PID data correctly.
Klaus: Is there any debug switch for libsi, or is the only way to add debug prints ? Additionally there is one question: Why are the link channels in the channel list. Basically, there is no need to add them there as there is always a portal which collects them. The dbox or WinDVBLive don't have them in the channel list.
dvbscan also doesn't detect the channels correctly. There they don't even have a name.
Regards
Dominik
I demand that Tomahawk may or may not have written...
I've got a problem with the auto sender "find/update" functionality of VDR. Though a sender is found, I am not able to watch it because the Audio/Video PIDs don't get recognized.
<thwap>
Next time, start a new thread...
[snip]
Dominik Strasser wrote:
Tomahawk wrote:
Yes when using the dbox2 as recveiver I can see the channel is broadcasting.
The problem appeared to me in connection with recent changes in transponder Layout by German "Premiere" Broadcaster. Unfortuantely I'm not able to copy the channel data to VDR because all problem channels are "optional" Channels and not saved in the Bouqet.
Greets
Tom
Klaus Schmidinger schrieb:
Tomahawk wrote:
Hi,
I've got a problem with the auto sender "find/update" functionality of VDR. Though a sender is found, I am not able to watch it because the Audio/Video PIDs don't get recognized.
Example of channel entry in channels.conf: Multisignal:362000:C0M64:C:6900:0:0:0:0:208:133:4:0
even leaving it tuned to the channel for some time doesn't fix the problem. Currently I'm using VDR 1.3.25 with Kernel 2.6.10-rc2 and also tested it on Kernel 2.6.12-rc6. My DVB-Card is a Siemens DVB-C. Any idea on how to solve this ?!?
Are you sure that this channel is broadcasting at the time you test it?
The problem is that the channel has the wrong frequency. IMHO it has the same frequency as the "Master" channel which is not necessarily the case. Editing the channel manually and trying all freqencies (there aren't many for Premiere and cable) is the only help I found. Once you have the correct frequency, VDR fills in the PID data correctly.
Klaus: Is there any debug switch for libsi, or is the only way to add debug prints ?
I'm afraid the only way is to add debug prints.
Additionally there is one question: Why are the link channels in the channel list. Basically, there is no need to add them there as there is always a portal which collects them. The dbox or WinDVBLive don't have them in the channel list.
Currently VDR's channel list is the list of _all_ known channels, so the link channels have to be in there just as well.
Once I have implemented user definable "favorite channel lists" you will be able to have only those channels in your list that you really want.
Klaus
don't know what you mean exactly, but I thought additional Information like what driver/kernel etc. could be useful.
Darren Salt schrieb:
I demand that Tomahawk may or may not have written...
I've got a problem with the auto sender "find/update" functionality of VDR. Though a sender is found, I am not able to watch it because the Audio/Video PIDs don't get recognized.
<thwap>
Next time, start a new thread...
[snip]
Klaus Schmidinger schrieb:
Dominik Strasser wrote:
Tomahawk wrote:
Yes when using the dbox2 as recveiver I can see the channel is broadcasting.
The problem appeared to me in connection with recent changes in transponder Layout by German "Premiere" Broadcaster. Unfortuantely I'm not able to copy the channel data to VDR because all problem channels are "optional" Channels and not saved in the Bouqet.
Greets
Tom
Klaus Schmidinger schrieb:
Tomahawk wrote:
Hi,
I've got a problem with the auto sender "find/update" functionality of VDR. Though a sender is found, I am not able to watch it because the Audio/Video PIDs don't get recognized.
Example of channel entry in channels.conf: Multisignal:362000:C0M64:C:6900:0:0:0:0:208:133:4:0
even leaving it tuned to the channel for some time doesn't fix the problem. Currently I'm using VDR 1.3.25 with Kernel 2.6.10-rc2 and also tested it on Kernel 2.6.12-rc6. My DVB-Card is a Siemens DVB-C. Any idea on how to solve this ?!?
Are you sure that this channel is broadcasting at the time you test it?
The problem is that the channel has the wrong frequency. IMHO it has the same frequency as the "Master" channel which is not necessarily the case. Editing the channel manually and trying all freqencies (there aren't many for Premiere and cable) is the only help I found. Once you have the correct frequency, VDR fills in the PID data correctly.
Klaus: Is there any debug switch for libsi, or is the only way to add debug prints ?
I'm afraid the only way is to add debug prints.
Additionally there is one question: Why are the link channels in the channel list. Basically, there is no need to add them there as there is always a portal which collects them. The dbox or WinDVBLive don't have them in the channel list.
Currently VDR's channel list is the list of _all_ known channels, so the link channels have to be in there just as well.
Once I have implemented user definable "favorite channel lists" you will be able to have only those channels in your list that you really want.
Klaus
vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
Getting back to the core problem, why is it that VDR detects the channel (even gives it the right name), but takes the Frequency of the base channel for this one instead of the correct frequency ? E.g. he takes the the 362 MHz of th base channel instead of 378 MHz where the channel physically resides.
Tomahawk wrote:
... Getting back to the core problem, why is it that VDR detects the channel (even gives it the right name), but takes the Frequency of the base channel for this one instead of the correct frequency ? E.g. he takes the the 362 MHz of th base channel instead of 378 MHz where the channel physically resides.
I guess that's because in the SDT filer (where new channels are detected) the only known frequency is that of the channel that is currently delivering the data. But maybe there is still something missing in VDR's way of handling this, so please feel free to debug into this and let me know if you find something.
Klaus
I demand that Tomahawk may or may not have top-posted...
Darren Salt schrieb:
I demand that Tomahawk may or may not have written...
I've got a problem with the auto sender "find/update" functionality of VDR. Though a sender is found, I am not able to watch it because the Audio/Video PIDs don't get recognized.
<thwap>
Next time, start a new thread...
don't know what you mean exactly, [...]
You followed up to an existing message when you should not have done so: because your message's headers contain references to other messages, it appeared as part of what should have been an entirely separate thread.
On Sat, Jun 11, 2005 at 01:17:11PM +0200, Prakash Punnoor wrote:
Dr. Werner Fink schrieb:
On Sat, Jun 11, 2005 at 12:21:26PM +0200, Prakash Punnoor wrote:
We already had this disussion with "CD" and the "giant" 90dB they claimed to give.
IIRC, CD gives something like 114db with dithering...and 96db without.
Informations which are not recorded can not be recreated without guesssing. Therefore 16 bits are 96db and not more.
Nope, you obviously don't know how dithering works. Every heard of DSD? So you think Sony's SACD doesn't work? DSD is pure dithering...
-> http://www.hfm-detmold.de/eti/projekte/diplomarbeiten/2004/dsdpcm/23.htm
On Sat, Jun 11, 2005 at 03:54:21PM +0200, Harald Milz wrote:
Rainer Zocholl UseNet-Posting-Nospam-74308-@zocki.toppoint.de wrote:
Add-dropping software would be lucky to have such an exact signal to detect "commericals" ends and beginns!
Do we have any chance to evaluate this value within VDR, or is the AC3 stream sent to the SPDIF out directly within the AV7110 chip? Werner, _you_ are the firmware guru, what do you think?
VDR sends the PS1 PES frames with its AC3 payload to the DVB card. The firmware only forward PCM, AC3, and DTS to the S/P-DIF out. No way and no resources (memory and CPU power) to filter or modify bigger part of the streams. Currently only the first few bytes of the DTS/AC3 data frames are checked to get the correct timings, data frame size and sampling rates.
Werner
Dr. Werner Fink schrieb:
On Sat, Jun 11, 2005 at 01:17:11PM +0200, Prakash Punnoor wrote:
Dr. Werner Fink schrieb:
On Sat, Jun 11, 2005 at 12:21:26PM +0200, Prakash Punnoor wrote:
We already had this disussion with "CD" and the "giant" 90dB they claimed to give.
IIRC, CD gives something like 114db with dithering...and 96db without.
Informations which are not recorded can not be recreated without guesssing. Therefore 16 bits are 96db and not more.
Nope, you obviously don't know how dithering works. Every heard of DSD? So you think Sony's SACD doesn't work? DSD is pure dithering...
-> http://www.hfm-detmold.de/eti/projekte/diplomarbeiten/2004/dsdpcm/23.htm
But what do you want to tell me? You are right or me? I am not an audio expert, but if I may quote following, I think it rather backs me up, as AFAIk noise shaping is a form of dithering, too. I understand:
96kHz, 20 bit pcm ~ SACD ~ 96kHz, 16 bit pcm w/ noise shaping
So dithering does help to reduce noise levels. Or did I completely misunderstand anything?
I am not suggesting to add dither later to a signal, but if you have a 20bit source and dither down to 16bit, you will have a better signal than 16 bit truncated/rounded - which was my original claim.
" Von den DVD-A-Befürwortern wird die SACD mit ihrer Bandbreite und Dynamik oftmals mit einer DVD-A von 20 Bit und 96 kHz Samplingfrequenz verglichen, welche jedoch den geringen Noise-Floor über die gesamte Bandbreite bis 48 kHz beibehalten kann. Gerne wird darüber hinaus betont, dass auch bei der DVD-A aufnahmeseitig Noise Shaping verwendet werden und somit bei einer Quantisierung von nur 16 Bit der Rauschpegel der SACD (bei einem deutlich geringeren Gesamtrauschpegel) mühelos unterboten werden kann. "
On Mon, Jun 13, 2005 at 04:08:47PM +0200, Prakash Punnoor wrote:
But what do you want to tell me? You are right or me? I am not an audio expert, but if I may quote following, I think it rather backs me up, as AFAIk noise shaping is a form of dithering, too. I understand:
96kHz, 20 bit pcm ~ SACD ~ 96kHz, 16 bit pcm w/ noise shaping
IMHO you do not get bak the exact original shape/wave form regardless how good the dithering is. What you can do is IMHO a mostly perfect guessing how it may look like at recording time and before quantization. This is somewhat of a higher art of fitting.
So dithering does help to reduce noise levels. Or did I completely misunderstand anything?
It helps to get the signal noise ratio as low as possible to avoid the requantization noise.
I am not suggesting to add dither later to a signal, but if you have a 20bit source and dither down to 16bit, you will have a better signal than 16 bit truncated/rounded - which was my original claim.
Hmmm ... IMHO the problem is: You may restore at most the original wave form, nevertheless you are never sure that you find the correct one without the ``forgotten'' information.
Werner
Prakash Punnoor wrote:
But what do you want to tell me? You are right or me? I am not an audio expert, but if I may quote following, I think it rather backs me up, as AFAIk noise shaping is a form of dithering, too. I understand:
96kHz, 20 bit pcm ~ SACD ~ 96kHz, 16 bit pcm w/ noise shaping
So dithering does help to reduce noise levels. Or did I completely misunderstand anything?
I think, yes.
Dithering does add noise to get a smoother sound at low volumes. I have read that the reason for dithering has something to do with psychoacoustics: Human ear does not like distortion like it is produced with quantisation noise, so "analogue" noise is added which makes the sound smoother while at the same time the noise is ignored or filtered by our mind. So in result, it sounds better even if we have added noise.
I found this explanation at http://homerecording.de/modules/news/article.php?storyid=389, so if you can read german or use a translator, you maybe might understand what I wanted to say above :-)
With kind regards
Joerg
Joerg Knitter schrieb:
Prakash Punnoor wrote:
But what do you want to tell me? You are right or me? I am not an audio expert, but if I may quote following, I think it rather backs me up, as AFAIk noise shaping is a form of dithering, too. I understand:
96kHz, 20 bit pcm ~ SACD ~ 96kHz, 16 bit pcm w/ noise shaping
So dithering does help to reduce noise levels. Or did I completely misunderstand anything?
I think, yes.
Dithering does add noise to get a smoother sound at low volumes. I have read that the reason for dithering has something to do with psychoacoustics: Human ear does not like distortion like it is produced with quantisation noise, so "analogue" noise is added which makes the sound smoother while at the same time the noise is ignored or filtered by our mind. So in result, it sounds better even if we have added noise.
Oh yes, I know about this. I didn't want to say "dithering is all great". It is a tradeoff. Dithering adds unwanted frequencies, right. For that you get better snr. Question is just, which is better. I case of 1bit (SACD, PC-Speaker, C64 digitized sounds...) you definately want better snr. ;-)
Just tried to insulate the available data using dvbsnoop
As far as I can see the needed data is not in SDT but because you get the Program number it seems you would have to scan all PATs (perhaps by caching them everytime you are tuned to their transponder) for the right program number. But at least you would know it is not on the current transponder because its not in the actual PAT.
Program Association Table (PAT) for 378 MHz in my case holds the correct information for the channel Program_number: 221 (0x00dd) reserved: 7 (0x07) Program_map_PID: 97 (0x0061)
I will dig into the tuxbox code where this works without problems to determine their way of working on this.
Greetz
Tom
Klaus Schmidinger schrieb:
Tomahawk wrote:
... Getting back to the core problem, why is it that VDR detects the channel (even gives it the right name), but takes the Frequency of the base channel for this one instead of the correct frequency ? E.g. he takes the the 362 MHz of th base channel instead of 378 MHz where the channel physically resides.
I guess that's because in the SDT filer (where new channels are detected) the only known frequency is that of the channel that is currently delivering the data. But maybe there is still something missing in VDR's way of handling this, so please feel free to debug into this and let me know if you find something.
Klaus
hi,
are you talking 'bout premiere subchannels on "direkt" and "sportportal"?
regards mws
On Monday 13 June 2005 20:23, Tomahawk wrote:
Just tried to insulate the available data using dvbsnoop
As far as I can see the needed data is not in SDT but because you get the Program number it seems you would have to scan all PATs (perhaps by caching them everytime you are tuned to their transponder) for the right program number. But at least you would know it is not on the current transponder because its not in the actual PAT.
Program Association Table (PAT) for 378 MHz in my case holds the correct information for the channel Program_number: 221 (0x00dd) reserved: 7 (0x07) Program_map_PID: 97 (0x0061)
I will dig into the tuxbox code where this works without problems to determine their way of working on this.
Greetz
Tom
Klaus Schmidinger schrieb:
Tomahawk wrote:
... Getting back to the core problem, why is it that VDR detects the channel (even gives it the right name), but takes the Frequency of the base channel for this one instead of the correct frequency ? E.g. he takes the the 362 MHz of th base channel instead of 378 MHz where the channel physically resides.
I guess that's because in the SDT filer (where new channels are detected) the only known frequency is that of the channel that is currently delivering the data. But maybe there is still something missing in VDR's way of handling this, so please feel free to debug into this and let me know if you find something.
Klaus
vdr mailing list vdr@linuxtv.org http://www.linuxtv.org/cgi-bin/mailman/listinfo/vdr
You're right thats what I'm talking about.
Mws schrieb:
hi,
are you talking 'bout premiere subchannels on "direkt" and "sportportal"?
regards mws
On Monday 13 June 2005 20:23, Tomahawk wrote:
Just tried to insulate the available data using dvbsnoop
As far as I can see the needed data is not in SDT but because you get the Program number it seems you would have to scan all PATs (perhaps by caching them everytime you are tuned to their transponder) for the right program number. But at least you would know it is not on the current transponder because its not in the actual PAT.
Program Association Table (PAT) for 378 MHz in my case holds the correct information for the channel Program_number: 221 (0x00dd) reserved: 7 (0x07) Program_map_PID: 97 (0x0061)
I will dig into the tuxbox code where this works without problems to determine their way of working on this.
Greetz
Tom
Klaus Schmidinger schrieb:
Tomahawk wrote:
... Getting back to the core problem, why is it that VDR detects the channel (even gives it the right name), but takes the Frequency of the base channel for this one instead of the correct frequency ? E.g. he takes the the 362 MHz of th base channel instead of 378 MHz where the channel physically resides.
I guess that's because in the SDT filer (where new channels are detected) the only known frequency is that of the channel that is currently delivering the data. But maybe there is still something missing in VDR's way of handling this, so please feel free to debug into this and let me know if you find something.
Klaus
Does anyone know about a good ALSA compatible tone generator which I could use to find the peaks? My sound card (well, a 24/96 one, ummm ...) and my Sennheiser 580 phones should be able to provide a near-flat frequency response up to about 20 kHz...
You could use XMMS and add a url tone://12000 to the playlist (you need the tone generator plugin, of course). Or try Audacity, even though that's kind of oversized ;)
Robert
Harald Milz wrote:
Does anyone know about a good ALSA compatible tone generator which I could use to find the peaks? My sound card (well, a 24/96 one, ummm ...) and my Sennheiser 580 phones should be able to provide a near-flat frequency response up to about 20 kHz...
Try speaker-test, in MDK that's part of alsa-utils package.
# speaker-test --help
speaker-test 1.0.8
Usage: speaker-test [OPTION]... -h,--help help -D,--device playback device -r,--rate stream rate in Hz -c,--channels count of channels in stream -f,--frequency sine wave frequency in Hz -b,--buffer ring buffer size in us -p,--period period size in us -s,--speaker single speaker test. Values 1=Left or 2=right
Anssi Hannula anssi.hannula@gmail.com wrote:
speaker-test 1.0.8
Ummm - everything above -f 5000 gives the same frequency...
Robert Huitl vdr@huitl.de wrote:
You could use XMMS and add a url tone://12000 to the playlist (you need the tone generator plugin, of course). Or try Audacity, even though that's kind of oversized ;)
aaaah - I forgot that. OK I can hear 17 kHz just fine. 18 is very low but that may be the plugin, soundcard, or headphone. Or my ears. Dunno. Will have to x-check.
Harald Milz wrote:
Anssi Hannula anssi.hannula@gmail.com wrote:
speaker-test 1.0.8
Ummm - everything above -f 5000 gives the same frequency...
Now that is strange, the speaker-test limits the frequency to 50-5000. I'm going to ask the alsa devs about that...
-
Anssi Hannula -
Darren Salt -
Dominik Strasser -
Dr. Werner Fink -
Harald Milz -
Joerg Knitter -
Klaus Schmidinger -
Lucian Muresan -
Mws -
Prakash Punnoor -
Rainer Zocholl -
Robert Huitl -
Tomahawk -
Wolfgang Rohdewald