<!-- Some styling for better description lists --><style type='text/css'>dt { font-weight: bold;float: left;display:inline;margin-right: 1em} dd { display:block; margin-left: 2em}</style> paulk-leonov: anyone familiar with omnivison sensors here? <br> trying to calculate exposition time, which is given in line size units <br> so the calculation involves hts (total number of photosites in a line) and clock <br> but I find conflicting info about what the clock should b <br> be <br> I strongly suspect it's pclk (pixel clock) <br> but some references mention sysclk or sclk jmondi: paulk-leonov: what sensor, I see you had a patch for ov5640 paulk-leonov: <u>jmondi</u>: I'm interested in ov5640, ov5648 and ov8865 jmondi: ah well, your question is pretty general after all paulk-leonov: yeah :) <br> <u>jmondi</u>: I've been fighting with omnivision clock trees for some days now jmondi: paulk-leonov: welcome to the party paulk-leonov: hehe jmondi: for ov5640 too ? maybe with CSI-2 bus ? paulk-leonov: well I'm not actually working on ov5640 but taking it as a reference <br> since it's the only driver that seems to make some calculation instead of reg arrays <br> I do have an ov5640 around, but in parallel jmondi: back to your question, hts and vst are usually expressed in pixel units, so I guess you should use the pixel clock if I'm not mistaken... paulk-leonov: <u>jmondi</u>: right, that's what I find to make the most sense <br> AFAIK sysclk is for the internal ISP hverkuil: <u>jmondi</u>: regarding your selection API documentation series: I get the feeling that I am missing something (see my latest reply). Perhaps we can discuss this more tomorrow morning in irc? paulk-leonov: but <br> ov5640_get_sysclk is used for debanding <br> which should also be pclk jmondi: paulk-leonov: where does that happen ? paulk-leonov: (and ov5640_get_sysclk is also used in ov5640_set_mode_exposure_calc) <br> <u>jmondi</u>: in mainline driver <br> ov5640_set_bandingfilter <br> for banding jmondi: also be aware that the ov5640 clock tree documentation and calculation procedure for CSI-2 bus is not "the possible best one" <br> paulk-leonov: yeah, I meant in which function :) paulk-leonov: right jmondi: paulk-leonov: there's an RFC where I tried to clarify documentation and the procedure a bit on the list: [RFC] media: i2c: ov5640: Rework CSI-2 clock tree <br> in case it might help paulk-leonov: ah nice <br> I'll take a look jmondi: <u>hverkuil</u>: sure, I'll be around tomorrrow morning <br> paulk-leonov: on the exposure calculation part I'm sorry I'm not familiar with that functions :) paulk-leonov: ah too bad :) <br> looks like there's no one on Earth who's familiar with it ;) jmondi: paulk-leonov: maybe it means nobody wants to touch it ? :) <br> <u>hverkuil</u>: seen your reply, let's talk tomorrow if you're available paulk-leonov: pretty sad considering what a role exposition plays in photography jmondi: indeed! the documentation we usually have from ov is of awful quality, that doesn't help much paulk-leonov: definitely yeah jmondi: very sad, hope that RFC helps, the current clock tree calculation is far from being perfect, or even 'good' <br> if you happen to have comments there to resurect that patch, feel free to send them out :) paulk-leonov: will do! ***: sszy has quit IRC (Quit: http://quassel-irc.org - Chat comfortably. Anywhere.) <br> _abbenormal has quit IRC (Read error: Connection reset by peer) <br> buzzmarshall has quit IRC (Ping timeout: 260 seconds) <br> RzR has quit IRC (Ping timeout: 240 seconds) <br> mszyprow|home has quit IRC (Ping timeout: 256 seconds) <br> b-rad has quit IRC (Ping timeout: 256 seconds) <br> headless has quit IRC (Quit: Konversation terminated!) <br> Robert_Zenz has quit IRC (Quit: Robert_Zenz) <br> ebarretto has quit IRC (Ping timeout: 256 seconds) <br> bingbu has quit IRC (Ping timeout: 256 seconds) <br> iive has quit IRC (Quit: They came for me...) <br> svarbanov has quit IRC (Ping timeout: 256 seconds)